1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 This filter supports the all above options as @ref{commands}.
649 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
653 Remove impulsive noise from input audio.
655 Samples detected as impulsive noise are replaced by interpolated samples using
656 autoregressive modelling.
660 Set window size, in milliseconds. Allowed range is from @code{10} to
661 @code{100}. Default value is @code{55} milliseconds.
662 This sets size of window which will be processed at once.
665 Set window overlap, in percentage of window size. Allowed range is from
666 @code{50} to @code{95}. Default value is @code{75} percent.
667 Setting this to a very high value increases impulsive noise removal but makes
668 whole process much slower.
671 Set autoregression order, in percentage of window size. Allowed range is from
672 @code{0} to @code{25}. Default value is @code{2} percent. This option also
673 controls quality of interpolated samples using neighbour good samples.
676 Set threshold value. Allowed range is from @code{1} to @code{100}.
677 Default value is @code{2}.
678 This controls the strength of impulsive noise which is going to be removed.
679 The lower value, the more samples will be detected as impulsive noise.
682 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
683 @code{10}. Default value is @code{2}.
684 If any two samples detected as noise are spaced less than this value then any
685 sample between those two samples will be also detected as noise.
690 It accepts the following values:
693 Select overlap-add method. Even not interpolated samples are slightly
694 changed with this method.
697 Select overlap-save method. Not interpolated samples remain unchanged.
700 Default value is @code{a}.
704 Remove clipped samples from input audio.
706 Samples detected as clipped are replaced by interpolated samples using
707 autoregressive modelling.
711 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
712 Default value is @code{55} milliseconds.
713 This sets size of window which will be processed at once.
716 Set window overlap, in percentage of window size. Allowed range is from @code{50}
717 to @code{95}. Default value is @code{75} percent.
720 Set autoregression order, in percentage of window size. Allowed range is from
721 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
722 quality of interpolated samples using neighbour good samples.
725 Set threshold value. Allowed range is from @code{1} to @code{100}.
726 Default value is @code{10}. Higher values make clip detection less aggressive.
729 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
730 Default value is @code{1000}. Higher values make clip detection less aggressive.
735 It accepts the following values:
738 Select overlap-add method. Even not interpolated samples are slightly changed
742 Select overlap-save method. Not interpolated samples remain unchanged.
745 Default value is @code{a}.
750 Delay one or more audio channels.
752 Samples in delayed channel are filled with silence.
754 The filter accepts the following option:
758 Set list of delays in milliseconds for each channel separated by '|'.
759 Unused delays will be silently ignored. If number of given delays is
760 smaller than number of channels all remaining channels will not be delayed.
761 If you want to delay exact number of samples, append 'S' to number.
762 If you want instead to delay in seconds, append 's' to number.
765 Use last set delay for all remaining channels. By default is disabled.
766 This option if enabled changes how option @code{delays} is interpreted.
773 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
774 the second channel (and any other channels that may be present) unchanged.
780 Delay second channel by 500 samples, the third channel by 700 samples and leave
781 the first channel (and any other channels that may be present) unchanged.
787 Delay all channels by same number of samples:
789 adelay=delays=64S:all=1
794 Remedy denormals in audio by adding extremely low-level noise.
796 This filter shall be placed before any filter that can produce denormals.
798 A description of the accepted parameters follows.
802 Set level of added noise in dB. Default is @code{-351}.
803 Allowed range is from -451 to -90.
806 Set type of added noise.
819 Default is @code{dc}.
824 This filter supports the all above options as @ref{commands}.
826 @section aderivative, aintegral
828 Compute derivative/integral of audio stream.
830 Applying both filters one after another produces original audio.
834 Apply echoing to the input audio.
836 Echoes are reflected sound and can occur naturally amongst mountains
837 (and sometimes large buildings) when talking or shouting; digital echo
838 effects emulate this behaviour and are often used to help fill out the
839 sound of a single instrument or vocal. The time difference between the
840 original signal and the reflection is the @code{delay}, and the
841 loudness of the reflected signal is the @code{decay}.
842 Multiple echoes can have different delays and decays.
844 A description of the accepted parameters follows.
848 Set input gain of reflected signal. Default is @code{0.6}.
851 Set output gain of reflected signal. Default is @code{0.3}.
854 Set list of time intervals in milliseconds between original signal and reflections
855 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
856 Default is @code{1000}.
859 Set list of loudness of reflected signals separated by '|'.
860 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
861 Default is @code{0.5}.
868 Make it sound as if there are twice as many instruments as are actually playing:
870 aecho=0.8:0.88:60:0.4
874 If delay is very short, then it sounds like a (metallic) robot playing music:
880 A longer delay will sound like an open air concert in the mountains:
882 aecho=0.8:0.9:1000:0.3
886 Same as above but with one more mountain:
888 aecho=0.8:0.9:1000|1800:0.3|0.25
893 Audio emphasis filter creates or restores material directly taken from LPs or
894 emphased CDs with different filter curves. E.g. to store music on vinyl the
895 signal has to be altered by a filter first to even out the disadvantages of
896 this recording medium.
897 Once the material is played back the inverse filter has to be applied to
898 restore the distortion of the frequency response.
900 The filter accepts the following options:
910 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
911 use @code{production} mode. Default is @code{reproduction} mode.
914 Set filter type. Selects medium. Can be one of the following:
926 select Compact Disc (CD).
932 select 50µs (FM-KF).
934 select 75µs (FM-KF).
940 This filter supports the all above options as @ref{commands}.
944 Modify an audio signal according to the specified expressions.
946 This filter accepts one or more expressions (one for each channel),
947 which are evaluated and used to modify a corresponding audio signal.
949 It accepts the following parameters:
953 Set the '|'-separated expressions list for each separate channel. If
954 the number of input channels is greater than the number of
955 expressions, the last specified expression is used for the remaining
958 @item channel_layout, c
959 Set output channel layout. If not specified, the channel layout is
960 specified by the number of expressions. If set to @samp{same}, it will
961 use by default the same input channel layout.
964 Each expression in @var{exprs} can contain the following constants and functions:
968 channel number of the current expression
971 number of the evaluated sample, starting from 0
977 time of the evaluated sample expressed in seconds
980 @item nb_out_channels
981 input and output number of channels
984 the value of input channel with number @var{CH}
987 Note: this filter is slow. For faster processing you should use a
996 aeval=val(ch)/2:c=same
1000 Invert phase of the second channel:
1002 aeval=val(0)|-val(1)
1008 An exciter is used to produce high sound that is not present in the
1009 original signal. This is done by creating harmonic distortions of the
1010 signal which are restricted in range and added to the original signal.
1011 An Exciter raises the upper end of an audio signal without simply raising
1012 the higher frequencies like an equalizer would do to create a more
1013 "crisp" or "brilliant" sound.
1015 The filter accepts the following options:
1019 Set input level prior processing of signal.
1020 Allowed range is from 0 to 64.
1024 Set output level after processing of signal.
1025 Allowed range is from 0 to 64.
1029 Set the amount of harmonics added to original signal.
1030 Allowed range is from 0 to 64.
1034 Set the amount of newly created harmonics.
1035 Allowed range is from 0.1 to 10.
1036 Default value is 8.5.
1039 Set the octave of newly created harmonics.
1040 Allowed range is from -10 to 10.
1044 Set the lower frequency limit of producing harmonics in Hz.
1045 Allowed range is from 2000 to 12000 Hz.
1049 Set the upper frequency limit of producing harmonics.
1050 Allowed range is from 9999 to 20000 Hz.
1051 If value is lower than 10000 Hz no limit is applied.
1054 Mute the original signal and output only added harmonics.
1055 By default is disabled.
1058 @subsection Commands
1060 This filter supports the all above options as @ref{commands}.
1065 Apply fade-in/out effect to input audio.
1067 A description of the accepted parameters follows.
1071 Specify the effect type, can be either @code{in} for fade-in, or
1072 @code{out} for a fade-out effect. Default is @code{in}.
1074 @item start_sample, ss
1075 Specify the number of the start sample for starting to apply the fade
1076 effect. Default is 0.
1078 @item nb_samples, ns
1079 Specify the number of samples for which the fade effect has to last. At
1080 the end of the fade-in effect the output audio will have the same
1081 volume as the input audio, at the end of the fade-out transition
1082 the output audio will be silence. Default is 44100.
1084 @item start_time, st
1085 Specify the start time of the fade effect. Default is 0.
1086 The value must be specified as a time duration; see
1087 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1088 for the accepted syntax.
1089 If set this option is used instead of @var{start_sample}.
1092 Specify the duration of the fade effect. See
1093 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1094 for the accepted syntax.
1095 At the end of the fade-in effect the output audio will have the same
1096 volume as the input audio, at the end of the fade-out transition
1097 the output audio will be silence.
1098 By default the duration is determined by @var{nb_samples}.
1099 If set this option is used instead of @var{nb_samples}.
1102 Set curve for fade transition.
1104 It accepts the following values:
1107 select triangular, linear slope (default)
1109 select quarter of sine wave
1111 select half of sine wave
1113 select exponential sine wave
1117 select inverted parabola
1131 select inverted quarter of sine wave
1133 select inverted half of sine wave
1135 select double-exponential seat
1137 select double-exponential sigmoid
1139 select logistic sigmoid
1141 select sine cardinal function
1143 select inverted sine cardinal function
1149 @subsection Commands
1151 This filter supports the all above options as @ref{commands}.
1153 @subsection Examples
1157 Fade in first 15 seconds of audio:
1159 afade=t=in:ss=0:d=15
1163 Fade out last 25 seconds of a 900 seconds audio:
1165 afade=t=out:st=875:d=25
1170 Denoise audio samples with FFT.
1172 A description of the accepted parameters follows.
1176 Set the noise reduction in dB, allowed range is 0.01 to 97.
1177 Default value is 12 dB.
1180 Set the noise floor in dB, allowed range is -80 to -20.
1181 Default value is -50 dB.
1186 It accepts the following values:
1195 Select shellac noise.
1198 Select custom noise, defined in @code{bn} option.
1200 Default value is white noise.
1204 Set custom band noise for every one of 15 bands.
1205 Bands are separated by ' ' or '|'.
1208 Set the residual floor in dB, allowed range is -80 to -20.
1209 Default value is -38 dB.
1212 Enable noise tracking. By default is disabled.
1213 With this enabled, noise floor is automatically adjusted.
1216 Enable residual tracking. By default is disabled.
1219 Set the output mode.
1221 It accepts the following values:
1224 Pass input unchanged.
1227 Pass noise filtered out.
1232 Default value is @var{o}.
1236 @subsection Commands
1238 This filter supports the following commands:
1240 @item sample_noise, sn
1241 Start or stop measuring noise profile.
1242 Syntax for the command is : "start" or "stop" string.
1243 After measuring noise profile is stopped it will be
1244 automatically applied in filtering.
1246 @item noise_reduction, nr
1247 Change noise reduction. Argument is single float number.
1248 Syntax for the command is : "@var{noise_reduction}"
1250 @item noise_floor, nf
1251 Change noise floor. Argument is single float number.
1252 Syntax for the command is : "@var{noise_floor}"
1254 @item output_mode, om
1255 Change output mode operation.
1256 Syntax for the command is : "i", "o" or "n" string.
1260 Apply arbitrary expressions to samples in frequency domain.
1264 Set frequency domain real expression for each separate channel separated
1265 by '|'. Default is "re".
1266 If the number of input channels is greater than the number of
1267 expressions, the last specified expression is used for the remaining
1271 Set frequency domain imaginary expression for each separate channel
1272 separated by '|'. Default is "im".
1274 Each expression in @var{real} and @var{imag} can contain the following
1275 constants and functions:
1282 current frequency bin number
1285 number of available bins
1288 channel number of the current expression
1297 current real part of frequency bin of current channel
1300 current imaginary part of frequency bin of current channel
1303 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1306 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1310 Set window size. Allowed range is from 16 to 131072.
1311 Default is @code{4096}
1314 Set window function. Default is @code{hann}.
1317 Set window overlap. If set to 1, the recommended overlap for selected
1318 window function will be picked. Default is @code{0.75}.
1321 @subsection Examples
1325 Leave almost only low frequencies in audio:
1327 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1331 Apply robotize effect:
1333 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1337 Apply whisper effect:
1339 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1346 Apply an arbitrary Finite Impulse Response filter.
1348 This filter is designed for applying long FIR filters,
1349 up to 60 seconds long.
1351 It can be used as component for digital crossover filters,
1352 room equalization, cross talk cancellation, wavefield synthesis,
1353 auralization, ambiophonics, ambisonics and spatialization.
1355 This filter uses the streams higher than first one as FIR coefficients.
1356 If the non-first stream holds a single channel, it will be used
1357 for all input channels in the first stream, otherwise
1358 the number of channels in the non-first stream must be same as
1359 the number of channels in the first stream.
1361 It accepts the following parameters:
1365 Set dry gain. This sets input gain.
1368 Set wet gain. This sets final output gain.
1371 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1374 Enable applying gain measured from power of IR.
1376 Set which approach to use for auto gain measurement.
1380 Do not apply any gain.
1383 select peak gain, very conservative approach. This is default value.
1386 select DC gain, limited application.
1389 select gain to noise approach, this is most popular one.
1393 Set gain to be applied to IR coefficients before filtering.
1394 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1397 Set format of IR stream. Can be @code{mono} or @code{input}.
1398 Default is @code{input}.
1401 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1402 Allowed range is 0.1 to 60 seconds.
1405 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1406 By default it is disabled.
1409 Set for which IR channel to display frequency response. By default is first channel
1410 displayed. This option is used only when @var{response} is enabled.
1413 Set video stream size. This option is used only when @var{response} is enabled.
1416 Set video stream frame rate. This option is used only when @var{response} is enabled.
1419 Set minimal partition size used for convolution. Default is @var{8192}.
1420 Allowed range is from @var{1} to @var{32768}.
1421 Lower values decreases latency at cost of higher CPU usage.
1424 Set maximal partition size used for convolution. Default is @var{8192}.
1425 Allowed range is from @var{8} to @var{32768}.
1426 Lower values may increase CPU usage.
1429 Set number of input impulse responses streams which will be switchable at runtime.
1430 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1433 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1434 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1435 This option can be changed at runtime via @ref{commands}.
1438 @subsection Examples
1442 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1444 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1451 Set output format constraints for the input audio. The framework will
1452 negotiate the most appropriate format to minimize conversions.
1454 It accepts the following parameters:
1457 @item sample_fmts, f
1458 A '|'-separated list of requested sample formats.
1460 @item sample_rates, r
1461 A '|'-separated list of requested sample rates.
1463 @item channel_layouts, cl
1464 A '|'-separated list of requested channel layouts.
1466 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1467 for the required syntax.
1470 If a parameter is omitted, all values are allowed.
1472 Force the output to either unsigned 8-bit or signed 16-bit stereo
1474 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1478 Apply frequency shift to input audio samples.
1480 The filter accepts the following options:
1484 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1485 Default value is 0.0.
1488 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1489 Default value is 1.0.
1492 @subsection Commands
1494 This filter supports the all above options as @ref{commands}.
1498 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1499 processing reduces disturbing noise between useful signals.
1501 Gating is done by detecting the volume below a chosen level @var{threshold}
1502 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1503 floor is set via @var{range}. Because an exact manipulation of the signal
1504 would cause distortion of the waveform the reduction can be levelled over
1505 time. This is done by setting @var{attack} and @var{release}.
1507 @var{attack} determines how long the signal has to fall below the threshold
1508 before any reduction will occur and @var{release} sets the time the signal
1509 has to rise above the threshold to reduce the reduction again.
1510 Shorter signals than the chosen attack time will be left untouched.
1514 Set input level before filtering.
1515 Default is 1. Allowed range is from 0.015625 to 64.
1518 Set the mode of operation. Can be @code{upward} or @code{downward}.
1519 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1520 will be amplified, expanding dynamic range in upward direction.
1521 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1524 Set the level of gain reduction when the signal is below the threshold.
1525 Default is 0.06125. Allowed range is from 0 to 1.
1526 Setting this to 0 disables reduction and then filter behaves like expander.
1529 If a signal rises above this level the gain reduction is released.
1530 Default is 0.125. Allowed range is from 0 to 1.
1533 Set a ratio by which the signal is reduced.
1534 Default is 2. Allowed range is from 1 to 9000.
1537 Amount of milliseconds the signal has to rise above the threshold before gain
1539 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1542 Amount of milliseconds the signal has to fall below the threshold before the
1543 reduction is increased again. Default is 250 milliseconds.
1544 Allowed range is from 0.01 to 9000.
1547 Set amount of amplification of signal after processing.
1548 Default is 1. Allowed range is from 1 to 64.
1551 Curve the sharp knee around the threshold to enter gain reduction more softly.
1552 Default is 2.828427125. Allowed range is from 1 to 8.
1555 Choose if exact signal should be taken for detection or an RMS like one.
1556 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1559 Choose if the average level between all channels or the louder channel affects
1561 Default is @code{average}. Can be @code{average} or @code{maximum}.
1564 @subsection Commands
1566 This filter supports the all above options as @ref{commands}.
1570 Apply an arbitrary Infinite Impulse Response filter.
1572 It accepts the following parameters:
1576 Set B/numerator/zeros/reflection coefficients.
1579 Set A/denominator/poles/ladder coefficients.
1591 Set coefficients format.
1595 lattice-ladder function
1597 analog transfer function
1599 digital transfer function
1601 Z-plane zeros/poles, cartesian (default)
1603 Z-plane zeros/poles, polar radians
1605 Z-plane zeros/poles, polar degrees
1611 Set type of processing.
1623 Set filtering precision.
1627 double-precision floating-point (default)
1629 single-precision floating-point
1637 Normalize filter coefficients, by default is enabled.
1638 Enabling it will normalize magnitude response at DC to 0dB.
1641 How much to use filtered signal in output. Default is 1.
1642 Range is between 0 and 1.
1645 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1646 By default it is disabled.
1649 Set for which IR channel to display frequency response. By default is first channel
1650 displayed. This option is used only when @var{response} is enabled.
1653 Set video stream size. This option is used only when @var{response} is enabled.
1656 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1659 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1660 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1663 Different coefficients and gains can be provided for every channel, in such case
1664 use '|' to separate coefficients or gains. Last provided coefficients will be
1665 used for all remaining channels.
1667 @subsection Examples
1671 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1673 aiir=k=1:z=7.957584807809675810E-1 -2.575128568908332300 3.674839853930788710 -2.57512875289799137 7.957586296317130880E-1:p=1 -2.86950072432325953 3.63022088054647218 -2.28075678147272232 6.361362326477423500E-1:f=tf:r=d
1677 Same as above but in @code{zp} format:
1679 aiir=k=0.79575848078096756:z=0.80918701+0.58773007i 0.80918701-0.58773007i 0.80884700+0.58784055i 0.80884700-0.58784055i:p=0.63892345+0.59951235i 0.63892345-0.59951235i 0.79582691+0.44198673i 0.79582691-0.44198673i:f=zp:r=s
1683 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1685 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1691 The limiter prevents an input signal from rising over a desired threshold.
1692 This limiter uses lookahead technology to prevent your signal from distorting.
1693 It means that there is a small delay after the signal is processed. Keep in mind
1694 that the delay it produces is the attack time you set.
1696 The filter accepts the following options:
1700 Set input gain. Default is 1.
1703 Set output gain. Default is 1.
1706 Don't let signals above this level pass the limiter. Default is 1.
1709 The limiter will reach its attenuation level in this amount of time in
1710 milliseconds. Default is 5 milliseconds.
1713 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1714 Default is 50 milliseconds.
1717 When gain reduction is always needed ASC takes care of releasing to an
1718 average reduction level rather than reaching a reduction of 0 in the release
1722 Select how much the release time is affected by ASC, 0 means nearly no changes
1723 in release time while 1 produces higher release times.
1726 Auto level output signal. Default is enabled.
1727 This normalizes audio back to 0dB if enabled.
1730 Depending on picked setting it is recommended to upsample input 2x or 4x times
1731 with @ref{aresample} before applying this filter.
1735 Apply a two-pole all-pass filter with central frequency (in Hz)
1736 @var{frequency}, and filter-width @var{width}.
1737 An all-pass filter changes the audio's frequency to phase relationship
1738 without changing its frequency to amplitude relationship.
1740 The filter accepts the following options:
1744 Set frequency in Hz.
1747 Set method to specify band-width of filter.
1762 Specify the band-width of a filter in width_type units.
1765 How much to use filtered signal in output. Default is 1.
1766 Range is between 0 and 1.
1769 Specify which channels to filter, by default all available are filtered.
1772 Normalize biquad coefficients, by default is disabled.
1773 Enabling it will normalize magnitude response at DC to 0dB.
1776 Set the filter order, can be 1 or 2. Default is 2.
1779 Set transform type of IIR filter.
1788 Set precison of filtering.
1791 Pick automatic sample format depending on surround filters.
1793 Always use signed 16-bit.
1795 Always use signed 32-bit.
1797 Always use float 32-bit.
1799 Always use float 64-bit.
1803 @subsection Commands
1805 This filter supports the following commands:
1808 Change allpass frequency.
1809 Syntax for the command is : "@var{frequency}"
1812 Change allpass width_type.
1813 Syntax for the command is : "@var{width_type}"
1816 Change allpass width.
1817 Syntax for the command is : "@var{width}"
1821 Syntax for the command is : "@var{mix}"
1828 The filter accepts the following options:
1832 Set the number of loops. Setting this value to -1 will result in infinite loops.
1836 Set maximal number of samples. Default is 0.
1839 Set first sample of loop. Default is 0.
1845 Merge two or more audio streams into a single multi-channel stream.
1847 The filter accepts the following options:
1852 Set the number of inputs. Default is 2.
1856 If the channel layouts of the inputs are disjoint, and therefore compatible,
1857 the channel layout of the output will be set accordingly and the channels
1858 will be reordered as necessary. If the channel layouts of the inputs are not
1859 disjoint, the output will have all the channels of the first input then all
1860 the channels of the second input, in that order, and the channel layout of
1861 the output will be the default value corresponding to the total number of
1864 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1865 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1866 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1867 first input, b1 is the first channel of the second input).
1869 On the other hand, if both input are in stereo, the output channels will be
1870 in the default order: a1, a2, b1, b2, and the channel layout will be
1871 arbitrarily set to 4.0, which may or may not be the expected value.
1873 All inputs must have the same sample rate, and format.
1875 If inputs do not have the same duration, the output will stop with the
1878 @subsection Examples
1882 Merge two mono files into a stereo stream:
1884 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1888 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1890 ffmpeg -i input.mkv -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6] amerge=inputs=6" -c:a pcm_s16le output.mkv
1896 Mixes multiple audio inputs into a single output.
1898 Note that this filter only supports float samples (the @var{amerge}
1899 and @var{pan} audio filters support many formats). If the @var{amix}
1900 input has integer samples then @ref{aresample} will be automatically
1901 inserted to perform the conversion to float samples.
1905 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1907 will mix 3 input audio streams to a single output with the same duration as the
1908 first input and a dropout transition time of 3 seconds.
1910 It accepts the following parameters:
1914 The number of inputs. If unspecified, it defaults to 2.
1917 How to determine the end-of-stream.
1921 The duration of the longest input. (default)
1924 The duration of the shortest input.
1927 The duration of the first input.
1931 @item dropout_transition
1932 The transition time, in seconds, for volume renormalization when an input
1933 stream ends. The default value is 2 seconds.
1936 Specify weight of each input audio stream as sequence.
1937 Each weight is separated by space. By default all inputs have same weight.
1940 Always scale inputs instead of only doing summation of samples.
1941 Beware of heavy clipping if inputs are not normalized prior or after filtering
1942 by this filter if this option is disabled. By default is enabled.
1945 @subsection Commands
1947 This filter supports the following commands:
1951 Syntax is same as option with same name.
1956 Multiply first audio stream with second audio stream and store result
1957 in output audio stream. Multiplication is done by multiplying each
1958 sample from first stream with sample at same position from second stream.
1960 With this element-wise multiplication one can create amplitude fades and
1961 amplitude modulations.
1963 @section anequalizer
1965 High-order parametric multiband equalizer for each channel.
1967 It accepts the following parameters:
1971 This option string is in format:
1972 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1973 Each equalizer band is separated by '|'.
1977 Set channel number to which equalization will be applied.
1978 If input doesn't have that channel the entry is ignored.
1981 Set central frequency for band.
1982 If input doesn't have that frequency the entry is ignored.
1985 Set band width in Hertz.
1988 Set band gain in dB.
1991 Set filter type for band, optional, can be:
1995 Butterworth, this is default.
2006 With this option activated frequency response of anequalizer is displayed
2010 Set video stream size. Only useful if curves option is activated.
2013 Set max gain that will be displayed. Only useful if curves option is activated.
2014 Setting this to a reasonable value makes it possible to display gain which is derived from
2015 neighbour bands which are too close to each other and thus produce higher gain
2016 when both are activated.
2019 Set frequency scale used to draw frequency response in video output.
2020 Can be linear or logarithmic. Default is logarithmic.
2023 Set color for each channel curve which is going to be displayed in video stream.
2024 This is list of color names separated by space or by '|'.
2025 Unrecognised or missing colors will be replaced by white color.
2028 @subsection Examples
2032 Lower gain by 10 of central frequency 200Hz and width 100 Hz
2033 for first 2 channels using Chebyshev type 1 filter:
2035 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
2039 @subsection Commands
2041 This filter supports the following commands:
2044 Alter existing filter parameters.
2045 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
2047 @var{fN} is existing filter number, starting from 0, if no such filter is available
2049 @var{freq} set new frequency parameter.
2050 @var{width} set new width parameter in Hertz.
2051 @var{gain} set new gain parameter in dB.
2053 Full filter invocation with asendcmd may look like this:
2054 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
2059 Reduce broadband noise in audio samples using Non-Local Means algorithm.
2061 Each sample is adjusted by looking for other samples with similar contexts. This
2062 context similarity is defined by comparing their surrounding patches of size
2063 @option{p}. Patches are searched in an area of @option{r} around the sample.
2065 The filter accepts the following options:
2069 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2072 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2073 Default value is 2 milliseconds.
2076 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2077 Default value is 6 milliseconds.
2080 Set the output mode.
2082 It accepts the following values:
2085 Pass input unchanged.
2088 Pass noise filtered out.
2093 Default value is @var{o}.
2097 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2100 @subsection Commands
2102 This filter supports the all above options as @ref{commands}.
2105 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2107 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2108 relate to producing the least mean square of the error signal (difference between the desired,
2109 2nd input audio stream and the actual signal, the 1st input audio stream).
2111 A description of the accepted options follows.
2124 Set the filter leakage.
2127 It accepts the following values:
2136 Pass filtered samples.
2139 Pass difference between desired and filtered samples.
2141 Default value is @var{o}.
2145 @subsection Examples
2149 One of many usages of this filter is noise reduction, input audio is filtered
2150 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2152 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2156 @subsection Commands
2158 This filter supports the same commands as options, excluding option @code{order}.
2162 Pass the audio source unchanged to the output.
2166 Pad the end of an audio stream with silence.
2168 This can be used together with @command{ffmpeg} @option{-shortest} to
2169 extend audio streams to the same length as the video stream.
2171 A description of the accepted options follows.
2175 Set silence packet size. Default value is 4096.
2178 Set the number of samples of silence to add to the end. After the
2179 value is reached, the stream is terminated. This option is mutually
2180 exclusive with @option{whole_len}.
2183 Set the minimum total number of samples in the output audio stream. If
2184 the value is longer than the input audio length, silence is added to
2185 the end, until the value is reached. This option is mutually exclusive
2186 with @option{pad_len}.
2189 Specify the duration of samples of silence to add. See
2190 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2191 for the accepted syntax. Used only if set to non-zero value.
2194 Specify the minimum total duration in the output audio stream. See
2195 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2196 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2197 the input audio length, silence is added to the end, until the value is reached.
2198 This option is mutually exclusive with @option{pad_dur}
2201 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2202 nor @option{whole_dur} option is set, the filter will add silence to the end of
2203 the input stream indefinitely.
2205 @subsection Examples
2209 Add 1024 samples of silence to the end of the input:
2215 Make sure the audio output will contain at least 10000 samples, pad
2216 the input with silence if required:
2218 apad=whole_len=10000
2222 Use @command{ffmpeg} to pad the audio input with silence, so that the
2223 video stream will always result the shortest and will be converted
2224 until the end in the output file when using the @option{shortest}
2227 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2232 Add a phasing effect to the input audio.
2234 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2235 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2237 A description of the accepted parameters follows.
2241 Set input gain. Default is 0.4.
2244 Set output gain. Default is 0.74
2247 Set delay in milliseconds. Default is 3.0.
2250 Set decay. Default is 0.4.
2253 Set modulation speed in Hz. Default is 0.5.
2256 Set modulation type. Default is triangular.
2258 It accepts the following values:
2265 @section aphaseshift
2266 Apply phase shift to input audio samples.
2268 The filter accepts the following options:
2272 Specify phase shift. Allowed range is from -1.0 to 1.0.
2273 Default value is 0.0.
2276 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2277 Default value is 1.0.
2280 @subsection Commands
2282 This filter supports the all above options as @ref{commands}.
2286 Audio pulsator is something between an autopanner and a tremolo.
2287 But it can produce funny stereo effects as well. Pulsator changes the volume
2288 of the left and right channel based on a LFO (low frequency oscillator) with
2289 different waveforms and shifted phases.
2290 This filter have the ability to define an offset between left and right
2291 channel. An offset of 0 means that both LFO shapes match each other.
2292 The left and right channel are altered equally - a conventional tremolo.
2293 An offset of 50% means that the shape of the right channel is exactly shifted
2294 in phase (or moved backwards about half of the frequency) - pulsator acts as
2295 an autopanner. At 1 both curves match again. Every setting in between moves the
2296 phase shift gapless between all stages and produces some "bypassing" sounds with
2297 sine and triangle waveforms. The more you set the offset near 1 (starting from
2298 the 0.5) the faster the signal passes from the left to the right speaker.
2300 The filter accepts the following options:
2304 Set input gain. By default it is 1. Range is [0.015625 - 64].
2307 Set output gain. By default it is 1. Range is [0.015625 - 64].
2310 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2311 sawup or sawdown. Default is sine.
2314 Set modulation. Define how much of original signal is affected by the LFO.
2317 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2320 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2323 Set pulse width. Default is 1. Allowed range is [0 - 2].
2326 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2329 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2333 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2337 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2338 if timing is set to hz.
2344 Resample the input audio to the specified parameters, using the
2345 libswresample library. If none are specified then the filter will
2346 automatically convert between its input and output.
2348 This filter is also able to stretch/squeeze the audio data to make it match
2349 the timestamps or to inject silence / cut out audio to make it match the
2350 timestamps, do a combination of both or do neither.
2352 The filter accepts the syntax
2353 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2354 expresses a sample rate and @var{resampler_options} is a list of
2355 @var{key}=@var{value} pairs, separated by ":". See the
2356 @ref{Resampler Options,,"Resampler Options" section in the
2357 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2358 for the complete list of supported options.
2360 @subsection Examples
2364 Resample the input audio to 44100Hz:
2370 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2371 samples per second compensation:
2373 aresample=async=1000
2379 Reverse an audio clip.
2381 Warning: This filter requires memory to buffer the entire clip, so trimming
2384 @subsection Examples
2388 Take the first 5 seconds of a clip, and reverse it.
2390 atrim=end=5,areverse
2396 Reduce noise from speech using Recurrent Neural Networks.
2398 This filter accepts the following options:
2402 Set train model file to load. This option is always required.
2405 Set how much to mix filtered samples into final output.
2406 Allowed range is from -1 to 1. Default value is 1.
2407 Negative values are special, they set how much to keep filtered noise
2408 in the final filter output. Set this option to -1 to hear actual
2409 noise removed from input signal.
2412 @subsection Commands
2414 This filter supports the all above options as @ref{commands}.
2416 @section asetnsamples
2418 Set the number of samples per each output audio frame.
2420 The last output packet may contain a different number of samples, as
2421 the filter will flush all the remaining samples when the input audio
2424 The filter accepts the following options:
2428 @item nb_out_samples, n
2429 Set the number of frames per each output audio frame. The number is
2430 intended as the number of samples @emph{per each channel}.
2431 Default value is 1024.
2434 If set to 1, the filter will pad the last audio frame with zeroes, so
2435 that the last frame will contain the same number of samples as the
2436 previous ones. Default value is 1.
2439 For example, to set the number of per-frame samples to 1234 and
2440 disable padding for the last frame, use:
2442 asetnsamples=n=1234:p=0
2447 Set the sample rate without altering the PCM data.
2448 This will result in a change of speed and pitch.
2450 The filter accepts the following options:
2453 @item sample_rate, r
2454 Set the output sample rate. Default is 44100 Hz.
2459 Show a line containing various information for each input audio frame.
2460 The input audio is not modified.
2462 The shown line contains a sequence of key/value pairs of the form
2463 @var{key}:@var{value}.
2465 The following values are shown in the output:
2469 The (sequential) number of the input frame, starting from 0.
2472 The presentation timestamp of the input frame, in time base units; the time base
2473 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2476 The presentation timestamp of the input frame in seconds.
2479 position of the frame in the input stream, -1 if this information in
2480 unavailable and/or meaningless (for example in case of synthetic audio)
2489 The sample rate for the audio frame.
2492 The number of samples (per channel) in the frame.
2495 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2496 audio, the data is treated as if all the planes were concatenated.
2498 @item plane_checksums
2499 A list of Adler-32 checksums for each data plane.
2503 Apply audio soft clipping.
2505 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2506 along a smooth curve, rather than the abrupt shape of hard-clipping.
2508 This filter accepts the following options:
2512 Set type of soft-clipping.
2514 It accepts the following values:
2528 Set threshold from where to start clipping. Default value is 0dB or 1.
2531 Set gain applied to output. Default value is 0dB or 1.
2534 Set additional parameter which controls sigmoid function.
2537 Set oversampling factor.
2540 @subsection Commands
2542 This filter supports the all above options as @ref{commands}.
2545 Automatic Speech Recognition
2547 This filter uses PocketSphinx for speech recognition. To enable
2548 compilation of this filter, you need to configure FFmpeg with
2549 @code{--enable-pocketsphinx}.
2551 It accepts the following options:
2555 Set sampling rate of input audio. Defaults is @code{16000}.
2556 This need to match speech models, otherwise one will get poor results.
2559 Set dictionary containing acoustic model files.
2562 Set pronunciation dictionary.
2565 Set language model file.
2568 Set language model set.
2571 Set which language model to use.
2574 Set output for log messages.
2577 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2582 Display time domain statistical information about the audio channels.
2583 Statistics are calculated and displayed for each audio channel and,
2584 where applicable, an overall figure is also given.
2586 It accepts the following option:
2589 Short window length in seconds, used for peak and trough RMS measurement.
2590 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2594 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2595 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2598 Available keys for each channel are:
2644 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2645 this @code{lavfi.astats.Overall.Peak_count}.
2647 For description what each key means read below.
2650 Set number of frame after which stats are going to be recalculated.
2651 Default is disabled.
2653 @item measure_perchannel
2654 Select the entries which need to be measured per channel. The metadata keys can
2655 be used as flags, default is @option{all} which measures everything.
2656 @option{none} disables all per channel measurement.
2658 @item measure_overall
2659 Select the entries which need to be measured overall. The metadata keys can
2660 be used as flags, default is @option{all} which measures everything.
2661 @option{none} disables all overall measurement.
2665 A description of each shown parameter follows:
2669 Mean amplitude displacement from zero.
2672 Minimal sample level.
2675 Maximal sample level.
2677 @item Min difference
2678 Minimal difference between two consecutive samples.
2680 @item Max difference
2681 Maximal difference between two consecutive samples.
2683 @item Mean difference
2684 Mean difference between two consecutive samples.
2685 The average of each difference between two consecutive samples.
2687 @item RMS difference
2688 Root Mean Square difference between two consecutive samples.
2692 Standard peak and RMS level measured in dBFS.
2696 Peak and trough values for RMS level measured over a short window.
2699 Standard ratio of peak to RMS level (note: not in dB).
2702 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2703 (i.e. either @var{Min level} or @var{Max level}).
2706 Number of occasions (not the number of samples) that the signal attained either
2707 @var{Min level} or @var{Max level}.
2709 @item Noise floor dB
2710 Minimum local peak measured in dBFS over a short window.
2712 @item Noise floor count
2713 Number of occasions (not the number of samples) that the signal attained
2717 Overall bit depth of audio. Number of bits used for each sample.
2720 Measured dynamic range of audio in dB.
2722 @item Zero crossings
2723 Number of points where the waveform crosses the zero level axis.
2725 @item Zero crossings rate
2726 Rate of Zero crossings and number of audio samples.
2730 Boost subwoofer frequencies.
2732 The filter accepts the following options:
2736 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2737 Default value is 0.7.
2740 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2741 Default value is 0.7.
2744 Set delay line decay gain value. Allowed range is from 0 to 1.
2745 Default value is 0.7.
2748 Set delay line feedback gain value. Allowed range is from 0 to 1.
2749 Default value is 0.9.
2752 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2753 Default value is 100.
2756 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2757 Default value is 0.5.
2760 Set delay. Allowed range is from 1 to 100.
2761 Default value is 20.
2764 @subsection Commands
2766 This filter supports the all above options as @ref{commands}.
2769 Cut subwoofer frequencies.
2771 This filter allows to set custom, steeper
2772 roll off than highpass filter, and thus is able to more attenuate
2773 frequency content in stop-band.
2775 The filter accepts the following options:
2779 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2780 Default value is 20.
2783 Set filter order. Available values are from 3 to 20.
2784 Default value is 10.
2787 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2790 @subsection Commands
2792 This filter supports the all above options as @ref{commands}.
2795 Cut super frequencies.
2797 The filter accepts the following options:
2801 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2802 Default value is 20000.
2805 Set filter order. Available values are from 3 to 20.
2806 Default value is 10.
2809 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2812 @subsection Commands
2814 This filter supports the all above options as @ref{commands}.
2817 Apply high order Butterworth band-pass filter.
2819 The filter accepts the following options:
2823 Set center frequency in Hertz. Allowed range is 2 to 999999.
2824 Default value is 1000.
2827 Set filter order. Available values are from 4 to 20.
2831 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2834 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2837 @subsection Commands
2839 This filter supports the all above options as @ref{commands}.
2842 Apply high order Butterworth band-stop filter.
2844 The filter accepts the following options:
2848 Set center frequency in Hertz. Allowed range is 2 to 999999.
2849 Default value is 1000.
2852 Set filter order. Available values are from 4 to 20.
2856 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2859 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2862 @subsection Commands
2864 This filter supports the all above options as @ref{commands}.
2870 The filter accepts exactly one parameter, the audio tempo. If not
2871 specified then the filter will assume nominal 1.0 tempo. Tempo must
2872 be in the [0.5, 100.0] range.
2874 Note that tempo greater than 2 will skip some samples rather than
2875 blend them in. If for any reason this is a concern it is always
2876 possible to daisy-chain several instances of atempo to achieve the
2877 desired product tempo.
2879 @subsection Examples
2883 Slow down audio to 80% tempo:
2889 To speed up audio to 300% tempo:
2895 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2897 atempo=sqrt(3),atempo=sqrt(3)
2901 @subsection Commands
2903 This filter supports the following commands:
2906 Change filter tempo scale factor.
2907 Syntax for the command is : "@var{tempo}"
2912 Trim the input so that the output contains one continuous subpart of the input.
2914 It accepts the following parameters:
2917 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2918 sample with the timestamp @var{start} will be the first sample in the output.
2921 Specify time of the first audio sample that will be dropped, i.e. the
2922 audio sample immediately preceding the one with the timestamp @var{end} will be
2923 the last sample in the output.
2926 Same as @var{start}, except this option sets the start timestamp in samples
2930 Same as @var{end}, except this option sets the end timestamp in samples instead
2934 The maximum duration of the output in seconds.
2937 The number of the first sample that should be output.
2940 The number of the first sample that should be dropped.
2943 @option{start}, @option{end}, and @option{duration} are expressed as time
2944 duration specifications; see
2945 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2947 Note that the first two sets of the start/end options and the @option{duration}
2948 option look at the frame timestamp, while the _sample options simply count the
2949 samples that pass through the filter. So start/end_pts and start/end_sample will
2950 give different results when the timestamps are wrong, inexact or do not start at
2951 zero. Also note that this filter does not modify the timestamps. If you wish
2952 to have the output timestamps start at zero, insert the asetpts filter after the
2955 If multiple start or end options are set, this filter tries to be greedy and
2956 keep all samples that match at least one of the specified constraints. To keep
2957 only the part that matches all the constraints at once, chain multiple atrim
2960 The defaults are such that all the input is kept. So it is possible to set e.g.
2961 just the end values to keep everything before the specified time.
2966 Drop everything except the second minute of input:
2968 ffmpeg -i INPUT -af atrim=60:120
2972 Keep only the first 1000 samples:
2974 ffmpeg -i INPUT -af atrim=end_sample=1000
2979 @section axcorrelate
2980 Calculate normalized cross-correlation between two input audio streams.
2982 Resulted samples are always between -1 and 1 inclusive.
2983 If result is 1 it means two input samples are highly correlated in that selected segment.
2984 Result 0 means they are not correlated at all.
2985 If result is -1 it means two input samples are out of phase, which means they cancel each
2988 The filter accepts the following options:
2992 Set size of segment over which cross-correlation is calculated.
2993 Default is 256. Allowed range is from 2 to 131072.
2996 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2997 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2998 are always zero and thus need much less calculations to make.
2999 This is generally not true, but is valid for typical audio streams.
3002 @subsection Examples
3006 Calculate correlation between channels in stereo audio stream:
3008 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
3014 Apply a two-pole Butterworth band-pass filter with central
3015 frequency @var{frequency}, and (3dB-point) band-width width.
3016 The @var{csg} option selects a constant skirt gain (peak gain = Q)
3017 instead of the default: constant 0dB peak gain.
3018 The filter roll off at 6dB per octave (20dB per decade).
3020 The filter accepts the following options:
3024 Set the filter's central frequency. Default is @code{3000}.
3027 Constant skirt gain if set to 1. Defaults to 0.
3030 Set method to specify band-width of filter.
3045 Specify the band-width of a filter in width_type units.
3048 How much to use filtered signal in output. Default is 1.
3049 Range is between 0 and 1.
3052 Specify which channels to filter, by default all available are filtered.
3055 Normalize biquad coefficients, by default is disabled.
3056 Enabling it will normalize magnitude response at DC to 0dB.
3059 Set transform type of IIR filter.
3068 Set precison of filtering.
3071 Pick automatic sample format depending on surround filters.
3073 Always use signed 16-bit.
3075 Always use signed 32-bit.
3077 Always use float 32-bit.
3079 Always use float 64-bit.
3083 @subsection Commands
3085 This filter supports the following commands:
3088 Change bandpass frequency.
3089 Syntax for the command is : "@var{frequency}"
3092 Change bandpass width_type.
3093 Syntax for the command is : "@var{width_type}"
3096 Change bandpass width.
3097 Syntax for the command is : "@var{width}"
3100 Change bandpass mix.
3101 Syntax for the command is : "@var{mix}"
3106 Apply a two-pole Butterworth band-reject filter with central
3107 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3108 The filter roll off at 6dB per octave (20dB per decade).
3110 The filter accepts the following options:
3114 Set the filter's central frequency. Default is @code{3000}.
3117 Set method to specify band-width of filter.
3132 Specify the band-width of a filter in width_type units.
3135 How much to use filtered signal in output. Default is 1.
3136 Range is between 0 and 1.
3139 Specify which channels to filter, by default all available are filtered.
3142 Normalize biquad coefficients, by default is disabled.
3143 Enabling it will normalize magnitude response at DC to 0dB.
3146 Set transform type of IIR filter.
3155 Set precison of filtering.
3158 Pick automatic sample format depending on surround filters.
3160 Always use signed 16-bit.
3162 Always use signed 32-bit.
3164 Always use float 32-bit.
3166 Always use float 64-bit.
3170 @subsection Commands
3172 This filter supports the following commands:
3175 Change bandreject frequency.
3176 Syntax for the command is : "@var{frequency}"
3179 Change bandreject width_type.
3180 Syntax for the command is : "@var{width_type}"
3183 Change bandreject width.
3184 Syntax for the command is : "@var{width}"
3187 Change bandreject mix.
3188 Syntax for the command is : "@var{mix}"
3191 @section bass, lowshelf
3193 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3194 shelving filter with a response similar to that of a standard
3195 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3197 The filter accepts the following options:
3201 Give the gain at 0 Hz. Its useful range is about -20
3202 (for a large cut) to +20 (for a large boost).
3203 Beware of clipping when using a positive gain.
3206 Set the filter's central frequency and so can be used
3207 to extend or reduce the frequency range to be boosted or cut.
3208 The default value is @code{100} Hz.
3211 Set method to specify band-width of filter.
3226 Determine how steep is the filter's shelf transition.
3229 Set number of poles. Default is 2.
3232 How much to use filtered signal in output. Default is 1.
3233 Range is between 0 and 1.
3236 Specify which channels to filter, by default all available are filtered.
3239 Normalize biquad coefficients, by default is disabled.
3240 Enabling it will normalize magnitude response at DC to 0dB.
3243 Set transform type of IIR filter.
3252 Set precison of filtering.
3255 Pick automatic sample format depending on surround filters.
3257 Always use signed 16-bit.
3259 Always use signed 32-bit.
3261 Always use float 32-bit.
3263 Always use float 64-bit.
3267 @subsection Commands
3269 This filter supports the following commands:
3272 Change bass frequency.
3273 Syntax for the command is : "@var{frequency}"
3276 Change bass width_type.
3277 Syntax for the command is : "@var{width_type}"
3281 Syntax for the command is : "@var{width}"
3285 Syntax for the command is : "@var{gain}"
3289 Syntax for the command is : "@var{mix}"
3294 Apply a biquad IIR filter with the given coefficients.
3295 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3296 are the numerator and denominator coefficients respectively.
3297 and @var{channels}, @var{c} specify which channels to filter, by default all
3298 available are filtered.
3300 @subsection Commands
3302 This filter supports the following commands:
3310 Change biquad parameter.
3311 Syntax for the command is : "@var{value}"
3314 How much to use filtered signal in output. Default is 1.
3315 Range is between 0 and 1.
3318 Specify which channels to filter, by default all available are filtered.
3321 Normalize biquad coefficients, by default is disabled.
3322 Enabling it will normalize magnitude response at DC to 0dB.
3325 Set transform type of IIR filter.
3334 Set precison of filtering.
3337 Pick automatic sample format depending on surround filters.
3339 Always use signed 16-bit.
3341 Always use signed 32-bit.
3343 Always use float 32-bit.
3345 Always use float 64-bit.
3350 Bauer stereo to binaural transformation, which improves headphone listening of
3351 stereo audio records.
3353 To enable compilation of this filter you need to configure FFmpeg with
3354 @code{--enable-libbs2b}.
3356 It accepts the following parameters:
3360 Pre-defined crossfeed level.
3364 Default level (fcut=700, feed=50).
3367 Chu Moy circuit (fcut=700, feed=60).
3370 Jan Meier circuit (fcut=650, feed=95).
3375 Cut frequency (in Hz).
3384 Remap input channels to new locations.
3386 It accepts the following parameters:
3389 Map channels from input to output. The argument is a '|'-separated list of
3390 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3391 @var{in_channel} form. @var{in_channel} can be either the name of the input
3392 channel (e.g. FL for front left) or its index in the input channel layout.
3393 @var{out_channel} is the name of the output channel or its index in the output
3394 channel layout. If @var{out_channel} is not given then it is implicitly an
3395 index, starting with zero and increasing by one for each mapping.
3397 @item channel_layout
3398 The channel layout of the output stream.
3401 If no mapping is present, the filter will implicitly map input channels to
3402 output channels, preserving indices.
3404 @subsection Examples
3408 For example, assuming a 5.1+downmix input MOV file,
3410 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3412 will create an output WAV file tagged as stereo from the downmix channels of
3416 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3418 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3422 @section channelsplit
3424 Split each channel from an input audio stream into a separate output stream.
3426 It accepts the following parameters:
3428 @item channel_layout
3429 The channel layout of the input stream. The default is "stereo".
3431 A channel layout describing the channels to be extracted as separate output streams
3432 or "all" to extract each input channel as a separate stream. The default is "all".
3434 Choosing channels not present in channel layout in the input will result in an error.
3437 @subsection Examples
3441 For example, assuming a stereo input MP3 file,
3443 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3445 will create an output Matroska file with two audio streams, one containing only
3446 the left channel and the other the right channel.
3449 Split a 5.1 WAV file into per-channel files:
3451 ffmpeg -i in.wav -filter_complex
3452 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3453 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3454 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3459 Extract only LFE from a 5.1 WAV file:
3461 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3462 -map '[LFE]' lfe.wav
3467 Add a chorus effect to the audio.
3469 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3471 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3472 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3473 The modulation depth defines the range the modulated delay is played before or after
3474 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3475 sound tuned around the original one, like in a chorus where some vocals are slightly
3478 It accepts the following parameters:
3481 Set input gain. Default is 0.4.
3484 Set output gain. Default is 0.4.
3487 Set delays. A typical delay is around 40ms to 60ms.
3499 @subsection Examples
3505 chorus=0.7:0.9:55:0.4:0.25:2
3511 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3515 Fuller sounding chorus with three delays:
3517 chorus=0.5:0.9:50|60|40:0.4|0.32|0.3:0.25|0.4|0.3:2|2.3|1.3
3522 Compress or expand the audio's dynamic range.
3524 It accepts the following parameters:
3530 A list of times in seconds for each channel over which the instantaneous level
3531 of the input signal is averaged to determine its volume. @var{attacks} refers to
3532 increase of volume and @var{decays} refers to decrease of volume. For most
3533 situations, the attack time (response to the audio getting louder) should be
3534 shorter than the decay time, because the human ear is more sensitive to sudden
3535 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3536 a typical value for decay is 0.8 seconds.
3537 If specified number of attacks & decays is lower than number of channels, the last
3538 set attack/decay will be used for all remaining channels.
3541 A list of points for the transfer function, specified in dB relative to the
3542 maximum possible signal amplitude. Each key points list must be defined using
3543 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3544 @code{x0/y0 x1/y1 x2/y2 ....}
3546 The input values must be in strictly increasing order but the transfer function
3547 does not have to be monotonically rising. The point @code{0/0} is assumed but
3548 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3549 function are @code{-70/-70|-60/-20|1/0}.
3552 Set the curve radius in dB for all joints. It defaults to 0.01.
3555 Set the additional gain in dB to be applied at all points on the transfer
3556 function. This allows for easy adjustment of the overall gain.
3560 Set an initial volume, in dB, to be assumed for each channel when filtering
3561 starts. This permits the user to supply a nominal level initially, so that, for
3562 example, a very large gain is not applied to initial signal levels before the
3563 companding has begun to operate. A typical value for audio which is initially
3564 quiet is -90 dB. It defaults to 0.
3567 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3568 delayed before being fed to the volume adjuster. Specifying a delay
3569 approximately equal to the attack/decay times allows the filter to effectively
3570 operate in predictive rather than reactive mode. It defaults to 0.
3574 @subsection Examples
3578 Make music with both quiet and loud passages suitable for listening to in a
3581 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3584 Another example for audio with whisper and explosion parts:
3586 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3590 A noise gate for when the noise is at a lower level than the signal:
3592 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3596 Here is another noise gate, this time for when the noise is at a higher level
3597 than the signal (making it, in some ways, similar to squelch):
3599 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3603 2:1 compression starting at -6dB:
3605 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3609 2:1 compression starting at -9dB:
3611 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3615 2:1 compression starting at -12dB:
3617 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3621 2:1 compression starting at -18dB:
3623 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3627 3:1 compression starting at -15dB:
3629 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3635 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3641 compand=attacks=0:points=-80/-169|-54/-80|-49.5/-64.6|-41.1/-41.1|-25.8/-15|-10.8/-4.5|0/0|20/8.3
3645 Hard limiter at -6dB:
3647 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3651 Hard limiter at -12dB:
3653 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3657 Hard noise gate at -35 dB:
3659 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3665 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3669 @section compensationdelay
3671 Compensation Delay Line is a metric based delay to compensate differing
3672 positions of microphones or speakers.
3674 For example, you have recorded guitar with two microphones placed in
3675 different locations. Because the front of sound wave has fixed speed in
3676 normal conditions, the phasing of microphones can vary and depends on
3677 their location and interposition. The best sound mix can be achieved when
3678 these microphones are in phase (synchronized). Note that a distance of
3679 ~30 cm between microphones makes one microphone capture the signal in
3680 antiphase to the other microphone. That makes the final mix sound moody.
3681 This filter helps to solve phasing problems by adding different delays
3682 to each microphone track and make them synchronized.
3684 The best result can be reached when you take one track as base and
3685 synchronize other tracks one by one with it.
3686 Remember that synchronization/delay tolerance depends on sample rate, too.
3687 Higher sample rates will give more tolerance.
3689 The filter accepts the following parameters:
3693 Set millimeters distance. This is compensation distance for fine tuning.
3697 Set cm distance. This is compensation distance for tightening distance setup.
3701 Set meters distance. This is compensation distance for hard distance setup.
3705 Set dry amount. Amount of unprocessed (dry) signal.
3709 Set wet amount. Amount of processed (wet) signal.
3713 Set temperature in degrees Celsius. This is the temperature of the environment.
3718 Apply headphone crossfeed filter.
3720 Crossfeed is the process of blending the left and right channels of stereo
3722 It is mainly used to reduce extreme stereo separation of low frequencies.
3724 The intent is to produce more speaker like sound to the listener.
3726 The filter accepts the following options:
3730 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3731 This sets gain of low shelf filter for side part of stereo image.
3732 Default is -6dB. Max allowed is -30db when strength is set to 1.
3735 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3736 This sets cut off frequency of low shelf filter. Default is cut off near
3737 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3740 Set curve slope of low shelf filter. Default is 0.5.
3741 Allowed range is from 0.01 to 1.
3744 Set input gain. Default is 0.9.
3747 Set output gain. Default is 1.
3750 @subsection Commands
3752 This filter supports the all above options as @ref{commands}.
3754 @section crystalizer
3755 Simple algorithm for audio noise sharpening.
3757 This filter linearly increases differences betweeen each audio sample.
3759 The filter accepts the following options:
3763 Sets the intensity of effect (default: 2.0). Must be in range between -10.0 to 0
3764 (unchanged sound) to 10.0 (maximum effect).
3765 To inverse filtering use negative value.
3768 Enable clipping. By default is enabled.
3771 @subsection Commands
3773 This filter supports the all above options as @ref{commands}.
3776 Apply a DC shift to the audio.
3778 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3779 in the recording chain) from the audio. The effect of a DC offset is reduced
3780 headroom and hence volume. The @ref{astats} filter can be used to determine if
3781 a signal has a DC offset.
3785 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3789 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3790 used to prevent clipping.
3795 Apply de-essing to the audio samples.
3799 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3803 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3807 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3811 Set the output mode.
3813 It accepts the following values:
3816 Pass input unchanged.
3819 Pass ess filtered out.
3824 Default value is @var{o}.
3830 Measure audio dynamic range.
3832 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3833 is found in transition material. And anything less that 8 have very poor dynamics
3834 and is very compressed.
3836 The filter accepts the following options:
3840 Set window length in seconds used to split audio into segments of equal length.
3841 Default is 3 seconds.
3845 Dynamic Audio Normalizer.
3847 This filter applies a certain amount of gain to the input audio in order
3848 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3849 contrast to more "simple" normalization algorithms, the Dynamic Audio
3850 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3851 This allows for applying extra gain to the "quiet" sections of the audio
3852 while avoiding distortions or clipping the "loud" sections. In other words:
3853 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3854 sections, in the sense that the volume of each section is brought to the
3855 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3856 this goal *without* applying "dynamic range compressing". It will retain 100%
3857 of the dynamic range *within* each section of the audio file.
3861 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3862 Default is 500 milliseconds.
3863 The Dynamic Audio Normalizer processes the input audio in small chunks,
3864 referred to as frames. This is required, because a peak magnitude has no
3865 meaning for just a single sample value. Instead, we need to determine the
3866 peak magnitude for a contiguous sequence of sample values. While a "standard"
3867 normalizer would simply use the peak magnitude of the complete file, the
3868 Dynamic Audio Normalizer determines the peak magnitude individually for each
3869 frame. The length of a frame is specified in milliseconds. By default, the
3870 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3871 been found to give good results with most files.
3872 Note that the exact frame length, in number of samples, will be determined
3873 automatically, based on the sampling rate of the individual input audio file.
3876 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3877 number. Default is 31.
3878 Probably the most important parameter of the Dynamic Audio Normalizer is the
3879 @code{window size} of the Gaussian smoothing filter. The filter's window size
3880 is specified in frames, centered around the current frame. For the sake of
3881 simplicity, this must be an odd number. Consequently, the default value of 31
3882 takes into account the current frame, as well as the 15 preceding frames and
3883 the 15 subsequent frames. Using a larger window results in a stronger
3884 smoothing effect and thus in less gain variation, i.e. slower gain
3885 adaptation. Conversely, using a smaller window results in a weaker smoothing
3886 effect and thus in more gain variation, i.e. faster gain adaptation.
3887 In other words, the more you increase this value, the more the Dynamic Audio
3888 Normalizer will behave like a "traditional" normalization filter. On the
3889 contrary, the more you decrease this value, the more the Dynamic Audio
3890 Normalizer will behave like a dynamic range compressor.
3893 Set the target peak value. This specifies the highest permissible magnitude
3894 level for the normalized audio input. This filter will try to approach the
3895 target peak magnitude as closely as possible, but at the same time it also
3896 makes sure that the normalized signal will never exceed the peak magnitude.
3897 A frame's maximum local gain factor is imposed directly by the target peak
3898 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3899 It is not recommended to go above this value.
3902 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3903 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3904 factor for each input frame, i.e. the maximum gain factor that does not
3905 result in clipping or distortion. The maximum gain factor is determined by
3906 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3907 additionally bounds the frame's maximum gain factor by a predetermined
3908 (global) maximum gain factor. This is done in order to avoid excessive gain
3909 factors in "silent" or almost silent frames. By default, the maximum gain
3910 factor is 10.0, For most inputs the default value should be sufficient and
3911 it usually is not recommended to increase this value. Though, for input
3912 with an extremely low overall volume level, it may be necessary to allow even
3913 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3914 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3915 Instead, a "sigmoid" threshold function will be applied. This way, the
3916 gain factors will smoothly approach the threshold value, but never exceed that
3920 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3921 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3922 This means that the maximum local gain factor for each frame is defined
3923 (only) by the frame's highest magnitude sample. This way, the samples can
3924 be amplified as much as possible without exceeding the maximum signal
3925 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3926 Normalizer can also take into account the frame's root mean square,
3927 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3928 determine the power of a time-varying signal. It is therefore considered
3929 that the RMS is a better approximation of the "perceived loudness" than
3930 just looking at the signal's peak magnitude. Consequently, by adjusting all
3931 frames to a constant RMS value, a uniform "perceived loudness" can be
3932 established. If a target RMS value has been specified, a frame's local gain
3933 factor is defined as the factor that would result in exactly that RMS value.
3934 Note, however, that the maximum local gain factor is still restricted by the
3935 frame's highest magnitude sample, in order to prevent clipping.
3938 Enable channels coupling. By default is enabled.
3939 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3940 amount. This means the same gain factor will be applied to all channels, i.e.
3941 the maximum possible gain factor is determined by the "loudest" channel.
3942 However, in some recordings, it may happen that the volume of the different
3943 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3944 In this case, this option can be used to disable the channel coupling. This way,
3945 the gain factor will be determined independently for each channel, depending
3946 only on the individual channel's highest magnitude sample. This allows for
3947 harmonizing the volume of the different channels.
3950 Enable DC bias correction. By default is disabled.
3951 An audio signal (in the time domain) is a sequence of sample values.
3952 In the Dynamic Audio Normalizer these sample values are represented in the
3953 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3954 audio signal, or "waveform", should be centered around the zero point.
3955 That means if we calculate the mean value of all samples in a file, or in a
3956 single frame, then the result should be 0.0 or at least very close to that
3957 value. If, however, there is a significant deviation of the mean value from
3958 0.0, in either positive or negative direction, this is referred to as a
3959 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3960 Audio Normalizer provides optional DC bias correction.
3961 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3962 the mean value, or "DC correction" offset, of each input frame and subtract
3963 that value from all of the frame's sample values which ensures those samples
3964 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3965 boundaries, the DC correction offset values will be interpolated smoothly
3966 between neighbouring frames.
3968 @item altboundary, b
3969 Enable alternative boundary mode. By default is disabled.
3970 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3971 around each frame. This includes the preceding frames as well as the
3972 subsequent frames. However, for the "boundary" frames, located at the very
3973 beginning and at the very end of the audio file, not all neighbouring
3974 frames are available. In particular, for the first few frames in the audio
3975 file, the preceding frames are not known. And, similarly, for the last few
3976 frames in the audio file, the subsequent frames are not known. Thus, the
3977 question arises which gain factors should be assumed for the missing frames
3978 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3979 to deal with this situation. The default boundary mode assumes a gain factor
3980 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3981 "fade out" at the beginning and at the end of the input, respectively.
3984 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3985 By default, the Dynamic Audio Normalizer does not apply "traditional"
3986 compression. This means that signal peaks will not be pruned and thus the
3987 full dynamic range will be retained within each local neighbourhood. However,
3988 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3989 normalization algorithm with a more "traditional" compression.
3990 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3991 (thresholding) function. If (and only if) the compression feature is enabled,
3992 all input frames will be processed by a soft knee thresholding function prior
3993 to the actual normalization process. Put simply, the thresholding function is
3994 going to prune all samples whose magnitude exceeds a certain threshold value.
3995 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3996 value. Instead, the threshold value will be adjusted for each individual
3998 In general, smaller parameters result in stronger compression, and vice versa.
3999 Values below 3.0 are not recommended, because audible distortion may appear.
4002 Set the target threshold value. This specifies the lowest permissible
4003 magnitude level for the audio input which will be normalized.
4004 If input frame volume is above this value frame will be normalized.
4005 Otherwise frame may not be normalized at all. The default value is set
4006 to 0, which means all input frames will be normalized.
4007 This option is mostly useful if digital noise is not wanted to be amplified.
4010 @subsection Commands
4012 This filter supports the all above options as @ref{commands}.
4016 Make audio easier to listen to on headphones.
4018 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
4019 so that when listened to on headphones the stereo image is moved from
4020 inside your head (standard for headphones) to outside and in front of
4021 the listener (standard for speakers).
4027 Apply a two-pole peaking equalisation (EQ) filter. With this
4028 filter, the signal-level at and around a selected frequency can
4029 be increased or decreased, whilst (unlike bandpass and bandreject
4030 filters) that at all other frequencies is unchanged.
4032 In order to produce complex equalisation curves, this filter can
4033 be given several times, each with a different central frequency.
4035 The filter accepts the following options:
4039 Set the filter's central frequency in Hz.
4042 Set method to specify band-width of filter.
4057 Specify the band-width of a filter in width_type units.
4060 Set the required gain or attenuation in dB.
4061 Beware of clipping when using a positive gain.
4064 How much to use filtered signal in output. Default is 1.
4065 Range is between 0 and 1.
4068 Specify which channels to filter, by default all available are filtered.
4071 Normalize biquad coefficients, by default is disabled.
4072 Enabling it will normalize magnitude response at DC to 0dB.
4075 Set transform type of IIR filter.
4084 Set precison of filtering.
4087 Pick automatic sample format depending on surround filters.
4089 Always use signed 16-bit.
4091 Always use signed 32-bit.
4093 Always use float 32-bit.
4095 Always use float 64-bit.
4099 @subsection Examples
4102 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4104 equalizer=f=1000:t=h:width=200:g=-10
4108 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4110 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4114 @subsection Commands
4116 This filter supports the following commands:
4119 Change equalizer frequency.
4120 Syntax for the command is : "@var{frequency}"
4123 Change equalizer width_type.
4124 Syntax for the command is : "@var{width_type}"
4127 Change equalizer width.
4128 Syntax for the command is : "@var{width}"
4131 Change equalizer gain.
4132 Syntax for the command is : "@var{gain}"
4135 Change equalizer mix.
4136 Syntax for the command is : "@var{mix}"
4139 @section extrastereo
4141 Linearly increases the difference between left and right channels which
4142 adds some sort of "live" effect to playback.
4144 The filter accepts the following options:
4148 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4149 (average of both channels), with 1.0 sound will be unchanged, with
4150 -1.0 left and right channels will be swapped.
4153 Enable clipping. By default is enabled.
4156 @subsection Commands
4158 This filter supports the all above options as @ref{commands}.
4160 @section firequalizer
4161 Apply FIR Equalization using arbitrary frequency response.
4163 The filter accepts the following option:
4167 Set gain curve equation (in dB). The expression can contain variables:
4170 the evaluated frequency
4174 channel number, set to 0 when multichannels evaluation is disabled
4176 channel id, see libavutil/channel_layout.h, set to the first channel id when
4177 multichannels evaluation is disabled
4181 channel_layout, see libavutil/channel_layout.h
4186 @item gain_interpolate(f)
4187 interpolate gain on frequency f based on gain_entry
4188 @item cubic_interpolate(f)
4189 same as gain_interpolate, but smoother
4191 This option is also available as command. Default is @code{gain_interpolate(f)}.
4194 Set gain entry for gain_interpolate function. The expression can
4198 store gain entry at frequency f with value g
4200 This option is also available as command.
4203 Set filter delay in seconds. Higher value means more accurate.
4204 Default is @code{0.01}.
4207 Set filter accuracy in Hz. Lower value means more accurate.
4208 Default is @code{5}.
4211 Set window function. Acceptable values are:
4214 rectangular window, useful when gain curve is already smooth
4216 hann window (default)
4222 3-terms continuous 1st derivative nuttall window
4224 minimum 3-terms discontinuous nuttall window
4226 4-terms continuous 1st derivative nuttall window
4228 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4230 blackman-harris window
4236 If enabled, use fixed number of audio samples. This improves speed when
4237 filtering with large delay. Default is disabled.
4240 Enable multichannels evaluation on gain. Default is disabled.
4243 Enable zero phase mode by subtracting timestamp to compensate delay.
4244 Default is disabled.
4247 Set scale used by gain. Acceptable values are:
4250 linear frequency, linear gain
4252 linear frequency, logarithmic (in dB) gain (default)
4254 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4256 logarithmic frequency, logarithmic gain
4260 Set file for dumping, suitable for gnuplot.
4263 Set scale for dumpfile. Acceptable values are same with scale option.
4267 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4268 Default is disabled.
4271 Enable minimum phase impulse response. Default is disabled.
4274 @subsection Examples
4279 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4282 lowpass at 1000 Hz with gain_entry:
4284 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4287 custom equalization:
4289 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4292 higher delay with zero phase to compensate delay:
4294 firequalizer=delay=0.1:fixed=on:zero_phase=on
4297 lowpass on left channel, highpass on right channel:
4299 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4300 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4305 Apply a flanging effect to the audio.
4307 The filter accepts the following options:
4311 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4314 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4317 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4321 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4322 Default value is 71.
4325 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4328 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4329 Default value is @var{sinusoidal}.
4332 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4333 Default value is 25.
4336 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4337 Default is @var{linear}.
4341 Apply Haas effect to audio.
4343 Note that this makes most sense to apply on mono signals.
4344 With this filter applied to mono signals it give some directionality and
4345 stretches its stereo image.
4347 The filter accepts the following options:
4351 Set input level. By default is @var{1}, or 0dB
4354 Set output level. By default is @var{1}, or 0dB.
4357 Set gain applied to side part of signal. By default is @var{1}.
4360 Set kind of middle source. Can be one of the following:
4370 Pick middle part signal of stereo image.
4373 Pick side part signal of stereo image.
4377 Change middle phase. By default is disabled.
4380 Set left channel delay. By default is @var{2.05} milliseconds.
4383 Set left channel balance. By default is @var{-1}.
4386 Set left channel gain. By default is @var{1}.
4389 Change left phase. By default is disabled.
4392 Set right channel delay. By defaults is @var{2.12} milliseconds.
4395 Set right channel balance. By default is @var{1}.
4398 Set right channel gain. By default is @var{1}.
4401 Change right phase. By default is enabled.
4406 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4407 embedded HDCD codes is expanded into a 20-bit PCM stream.
4409 The filter supports the Peak Extend and Low-level Gain Adjustment features
4410 of HDCD, and detects the Transient Filter flag.
4413 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4416 When using the filter with wav, note the default encoding for wav is 16-bit,
4417 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4418 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4420 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4421 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4424 The filter accepts the following options:
4427 @item disable_autoconvert
4428 Disable any automatic format conversion or resampling in the filter graph.
4430 @item process_stereo
4431 Process the stereo channels together. If target_gain does not match between
4432 channels, consider it invalid and use the last valid target_gain.
4435 Set the code detect timer period in ms.
4438 Always extend peaks above -3dBFS even if PE isn't signaled.
4441 Replace audio with a solid tone and adjust the amplitude to signal some
4442 specific aspect of the decoding process. The output file can be loaded in
4443 an audio editor alongside the original to aid analysis.
4445 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4452 Gain adjustment level at each sample
4454 Samples where peak extend occurs
4456 Samples where the code detect timer is active
4458 Samples where the target gain does not match between channels
4464 Apply head-related transfer functions (HRTFs) to create virtual
4465 loudspeakers around the user for binaural listening via headphones.
4466 The HRIRs are provided via additional streams, for each channel
4467 one stereo input stream is needed.
4469 The filter accepts the following options:
4473 Set mapping of input streams for convolution.
4474 The argument is a '|'-separated list of channel names in order as they
4475 are given as additional stream inputs for filter.
4476 This also specify number of input streams. Number of input streams
4477 must be not less than number of channels in first stream plus one.
4480 Set gain applied to audio. Value is in dB. Default is 0.
4483 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4484 processing audio in time domain which is slow.
4485 @var{freq} is processing audio in frequency domain which is fast.
4486 Default is @var{freq}.
4489 Set custom gain for LFE channels. Value is in dB. Default is 0.
4492 Set size of frame in number of samples which will be processed at once.
4493 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4496 Set format of hrir stream.
4497 Default value is @var{stereo}. Alternative value is @var{multich}.
4498 If value is set to @var{stereo}, number of additional streams should
4499 be greater or equal to number of input channels in first input stream.
4500 Also each additional stream should have stereo number of channels.
4501 If value is set to @var{multich}, number of additional streams should
4502 be exactly one. Also number of input channels of additional stream
4503 should be equal or greater than twice number of channels of first input
4507 @subsection Examples
4511 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4512 each amovie filter use stereo file with IR coefficients as input.
4513 The files give coefficients for each position of virtual loudspeaker:
4516 -filter_complex "amovie=azi_270_ele_0_DFC.wav[sr];amovie=azi_90_ele_0_DFC.wav[sl];amovie=azi_225_ele_0_DFC.wav[br];amovie=azi_135_ele_0_DFC.wav[bl];amovie=azi_0_ele_0_DFC.wav,asplit[fc][lfe];amovie=azi_35_ele_0_DFC.wav[fl];amovie=azi_325_ele_0_DFC.wav[fr];[0:a][fl][fr][fc][lfe][bl][br][sl][sr]headphone=FL|FR|FC|LFE|BL|BR|SL|SR"
4521 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4522 but now in @var{multich} @var{hrir} format.
4524 ffmpeg -i input.wav -filter_complex "amovie=minp.wav[hrirs];[0:a][hrirs]headphone=map=FL|FR|FC|LFE|BL|BR|SL|SR:hrir=multich"
4531 Apply a high-pass filter with 3dB point frequency.
4532 The filter can be either single-pole, or double-pole (the default).
4533 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4535 The filter accepts the following options:
4539 Set frequency in Hz. Default is 3000.
4542 Set number of poles. Default is 2.
4545 Set method to specify band-width of filter.
4560 Specify the band-width of a filter in width_type units.
4561 Applies only to double-pole filter.
4562 The default is 0.707q and gives a Butterworth response.
4565 How much to use filtered signal in output. Default is 1.
4566 Range is between 0 and 1.
4569 Specify which channels to filter, by default all available are filtered.
4572 Normalize biquad coefficients, by default is disabled.
4573 Enabling it will normalize magnitude response at DC to 0dB.
4576 Set transform type of IIR filter.
4585 Set precison of filtering.
4588 Pick automatic sample format depending on surround filters.
4590 Always use signed 16-bit.
4592 Always use signed 32-bit.
4594 Always use float 32-bit.
4596 Always use float 64-bit.
4600 @subsection Commands
4602 This filter supports the following commands:
4605 Change highpass frequency.
4606 Syntax for the command is : "@var{frequency}"
4609 Change highpass width_type.
4610 Syntax for the command is : "@var{width_type}"
4613 Change highpass width.
4614 Syntax for the command is : "@var{width}"
4617 Change highpass mix.
4618 Syntax for the command is : "@var{mix}"
4623 Join multiple input streams into one multi-channel stream.
4625 It accepts the following parameters:
4629 The number of input streams. It defaults to 2.
4631 @item channel_layout
4632 The desired output channel layout. It defaults to stereo.
4635 Map channels from inputs to output. The argument is a '|'-separated list of
4636 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4637 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4638 can be either the name of the input channel (e.g. FL for front left) or its
4639 index in the specified input stream. @var{out_channel} is the name of the output
4643 The filter will attempt to guess the mappings when they are not specified
4644 explicitly. It does so by first trying to find an unused matching input channel
4645 and if that fails it picks the first unused input channel.
4647 Join 3 inputs (with properly set channel layouts):
4649 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4652 Build a 5.1 output from 6 single-channel streams:
4654 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4655 'join=inputs=6:channel_layout=5.1:map=0.0-FL|1.0-FR|2.0-FC|3.0-SL|4.0-SR|5.0-LFE'
4661 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4663 To enable compilation of this filter you need to configure FFmpeg with
4664 @code{--enable-ladspa}.
4668 Specifies the name of LADSPA plugin library to load. If the environment
4669 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4670 each one of the directories specified by the colon separated list in
4671 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4672 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4673 @file{/usr/lib/ladspa/}.
4676 Specifies the plugin within the library. Some libraries contain only
4677 one plugin, but others contain many of them. If this is not set filter
4678 will list all available plugins within the specified library.
4681 Set the '|' separated list of controls which are zero or more floating point
4682 values that determine the behavior of the loaded plugin (for example delay,
4684 Controls need to be defined using the following syntax:
4685 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4686 @var{valuei} is the value set on the @var{i}-th control.
4687 Alternatively they can be also defined using the following syntax:
4688 @var{value0}|@var{value1}|@var{value2}|..., where
4689 @var{valuei} is the value set on the @var{i}-th control.
4690 If @option{controls} is set to @code{help}, all available controls and
4691 their valid ranges are printed.
4693 @item sample_rate, s
4694 Specify the sample rate, default to 44100. Only used if plugin have
4698 Set the number of samples per channel per each output frame, default
4699 is 1024. Only used if plugin have zero inputs.
4702 Set the minimum duration of the sourced audio. See
4703 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4704 for the accepted syntax.
4705 Note that the resulting duration may be greater than the specified duration,
4706 as the generated audio is always cut at the end of a complete frame.
4707 If not specified, or the expressed duration is negative, the audio is
4708 supposed to be generated forever.
4709 Only used if plugin have zero inputs.
4712 Enable latency compensation, by default is disabled.
4713 Only used if plugin have inputs.
4716 @subsection Examples
4720 List all available plugins within amp (LADSPA example plugin) library:
4726 List all available controls and their valid ranges for @code{vcf_notch}
4727 plugin from @code{VCF} library:
4729 ladspa=f=vcf:p=vcf_notch:c=help
4733 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4736 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4740 Add reverberation to the audio using TAP-plugins
4741 (Tom's Audio Processing plugins):
4743 ladspa=file=tap_reverb:tap_reverb
4747 Generate white noise, with 0.2 amplitude:
4749 ladspa=file=cmt:noise_source_white:c=c0=.2
4753 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4754 @code{C* Audio Plugin Suite} (CAPS) library:
4756 ladspa=file=caps:Click:c=c1=20'
4760 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4762 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4766 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4767 @code{SWH Plugins} collection:
4769 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4773 Attenuate low frequencies using Multiband EQ from Steve Harris
4774 @code{SWH Plugins} collection:
4776 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4780 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4783 ladspa=caps:Narrower
4787 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4789 ladspa=caps:White:.2
4793 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4795 ladspa=caps:Fractal:c=c1=1
4799 Dynamic volume normalization using @code{VLevel} plugin:
4801 ladspa=vlevel-ladspa:vlevel_mono
4805 @subsection Commands
4807 This filter supports the following commands:
4810 Modify the @var{N}-th control value.
4812 If the specified value is not valid, it is ignored and prior one is kept.
4817 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4818 Support for both single pass (livestreams, files) and double pass (files) modes.
4819 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4820 detect true peaks, the audio stream will be upsampled to 192 kHz.
4821 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4823 The filter accepts the following options:
4827 Set integrated loudness target.
4828 Range is -70.0 - -5.0. Default value is -24.0.
4831 Set loudness range target.
4832 Range is 1.0 - 20.0. Default value is 7.0.
4835 Set maximum true peak.
4836 Range is -9.0 - +0.0. Default value is -2.0.
4838 @item measured_I, measured_i
4839 Measured IL of input file.
4840 Range is -99.0 - +0.0.
4842 @item measured_LRA, measured_lra
4843 Measured LRA of input file.
4844 Range is 0.0 - 99.0.
4846 @item measured_TP, measured_tp
4847 Measured true peak of input file.
4848 Range is -99.0 - +99.0.
4850 @item measured_thresh
4851 Measured threshold of input file.
4852 Range is -99.0 - +0.0.
4855 Set offset gain. Gain is applied before the true-peak limiter.
4856 Range is -99.0 - +99.0. Default is +0.0.
4859 Normalize by linearly scaling the source audio.
4860 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4861 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4862 be lower than source LRA and the change in integrated loudness shouldn't
4863 result in a true peak which exceeds the target TP. If any of these
4864 conditions aren't met, normalization mode will revert to @var{dynamic}.
4865 Options are @code{true} or @code{false}. Default is @code{true}.
4868 Treat mono input files as "dual-mono". If a mono file is intended for playback
4869 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4870 If set to @code{true}, this option will compensate for this effect.
4871 Multi-channel input files are not affected by this option.
4872 Options are true or false. Default is false.
4875 Set print format for stats. Options are summary, json, or none.
4876 Default value is none.
4881 Apply a low-pass filter with 3dB point frequency.
4882 The filter can be either single-pole or double-pole (the default).
4883 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4885 The filter accepts the following options:
4889 Set frequency in Hz. Default is 500.
4892 Set number of poles. Default is 2.
4895 Set method to specify band-width of filter.
4910 Specify the band-width of a filter in width_type units.
4911 Applies only to double-pole filter.
4912 The default is 0.707q and gives a Butterworth response.
4915 How much to use filtered signal in output. Default is 1.
4916 Range is between 0 and 1.
4919 Specify which channels to filter, by default all available are filtered.
4922 Normalize biquad coefficients, by default is disabled.
4923 Enabling it will normalize magnitude response at DC to 0dB.
4926 Set transform type of IIR filter.
4935 Set precison of filtering.
4938 Pick automatic sample format depending on surround filters.
4940 Always use signed 16-bit.
4942 Always use signed 32-bit.
4944 Always use float 32-bit.
4946 Always use float 64-bit.
4950 @subsection Examples
4953 Lowpass only LFE channel, it LFE is not present it does nothing:
4959 @subsection Commands
4961 This filter supports the following commands:
4964 Change lowpass frequency.
4965 Syntax for the command is : "@var{frequency}"
4968 Change lowpass width_type.
4969 Syntax for the command is : "@var{width_type}"
4972 Change lowpass width.
4973 Syntax for the command is : "@var{width}"
4977 Syntax for the command is : "@var{mix}"
4982 Load a LV2 (LADSPA Version 2) plugin.
4984 To enable compilation of this filter you need to configure FFmpeg with
4985 @code{--enable-lv2}.
4989 Specifies the plugin URI. You may need to escape ':'.
4992 Set the '|' separated list of controls which are zero or more floating point
4993 values that determine the behavior of the loaded plugin (for example delay,
4995 If @option{controls} is set to @code{help}, all available controls and
4996 their valid ranges are printed.
4998 @item sample_rate, s
4999 Specify the sample rate, default to 44100. Only used if plugin have
5003 Set the number of samples per channel per each output frame, default
5004 is 1024. Only used if plugin have zero inputs.
5007 Set the minimum duration of the sourced audio. See
5008 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5009 for the accepted syntax.
5010 Note that the resulting duration may be greater than the specified duration,
5011 as the generated audio is always cut at the end of a complete frame.
5012 If not specified, or the expressed duration is negative, the audio is
5013 supposed to be generated forever.
5014 Only used if plugin have zero inputs.
5017 @subsection Examples
5021 Apply bass enhancer plugin from Calf:
5023 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
5027 Apply vinyl plugin from Calf:
5029 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
5033 Apply bit crusher plugin from ArtyFX:
5035 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
5040 Multiband Compress or expand the audio's dynamic range.
5042 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
5043 This is akin to the crossover of a loudspeaker, and results in flat frequency
5044 response when absent compander action.
5046 It accepts the following parameters:
5050 This option syntax is:
5051 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
5052 For explanation of each item refer to compand filter documentation.
5058 Mix channels with specific gain levels. The filter accepts the output
5059 channel layout followed by a set of channels definitions.
5061 This filter is also designed to efficiently remap the channels of an audio
5064 The filter accepts parameters of the form:
5065 "@var{l}|@var{outdef}|@var{outdef}|..."
5069 output channel layout or number of channels
5072 output channel specification, of the form:
5073 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
5076 output channel to define, either a channel name (FL, FR, etc.) or a channel
5077 number (c0, c1, etc.)
5080 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5083 input channel to use, see out_name for details; it is not possible to mix
5084 named and numbered input channels
5087 If the `=' in a channel specification is replaced by `<', then the gains for
5088 that specification will be renormalized so that the total is 1, thus
5089 avoiding clipping noise.
5091 @subsection Mixing examples
5093 For example, if you want to down-mix from stereo to mono, but with a bigger
5094 factor for the left channel:
5096 pan=1c|c0=0.9*c0+0.1*c1
5099 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5100 7-channels surround:
5102 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5105 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5106 that should be preferred (see "-ac" option) unless you have very specific
5109 @subsection Remapping examples
5111 The channel remapping will be effective if, and only if:
5114 @item gain coefficients are zeroes or ones,
5115 @item only one input per channel output,
5118 If all these conditions are satisfied, the filter will notify the user ("Pure
5119 channel mapping detected"), and use an optimized and lossless method to do the
5122 For example, if you have a 5.1 source and want a stereo audio stream by
5123 dropping the extra channels:
5125 pan="stereo| c0=FL | c1=FR"
5128 Given the same source, you can also switch front left and front right channels
5129 and keep the input channel layout:
5131 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5134 If the input is a stereo audio stream, you can mute the front left channel (and
5135 still keep the stereo channel layout) with:
5140 Still with a stereo audio stream input, you can copy the right channel in both
5141 front left and right:
5143 pan="stereo| c0=FR | c1=FR"
5148 ReplayGain scanner filter. This filter takes an audio stream as an input and
5149 outputs it unchanged.
5150 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5154 Convert the audio sample format, sample rate and channel layout. It is
5155 not meant to be used directly.
5158 Apply time-stretching and pitch-shifting with librubberband.
5160 To enable compilation of this filter, you need to configure FFmpeg with
5161 @code{--enable-librubberband}.
5163 The filter accepts the following options:
5167 Set tempo scale factor.
5170 Set pitch scale factor.
5173 Set transients detector.
5174 Possible values are:
5183 Possible values are:
5192 Possible values are:
5199 Set processing window size.
5200 Possible values are:
5209 Possible values are:
5216 Enable formant preservation when shift pitching.
5217 Possible values are:
5225 Possible values are:
5234 Possible values are:
5241 @subsection Commands
5243 This filter supports the following commands:
5246 Change filter tempo scale factor.
5247 Syntax for the command is : "@var{tempo}"
5250 Change filter pitch scale factor.
5251 Syntax for the command is : "@var{pitch}"
5254 @section sidechaincompress
5256 This filter acts like normal compressor but has the ability to compress
5257 detected signal using second input signal.
5258 It needs two input streams and returns one output stream.
5259 First input stream will be processed depending on second stream signal.
5260 The filtered signal then can be filtered with other filters in later stages of
5261 processing. See @ref{pan} and @ref{amerge} filter.
5263 The filter accepts the following options:
5267 Set input gain. Default is 1. Range is between 0.015625 and 64.
5270 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5271 Default is @code{downward}.
5274 If a signal of second stream raises above this level it will affect the gain
5275 reduction of first stream.
5276 By default is 0.125. Range is between 0.00097563 and 1.
5279 Set a ratio about which the signal is reduced. 1:2 means that if the level
5280 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5281 Default is 2. Range is between 1 and 20.
5284 Amount of milliseconds the signal has to rise above the threshold before gain
5285 reduction starts. Default is 20. Range is between 0.01 and 2000.
5288 Amount of milliseconds the signal has to fall below the threshold before
5289 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5292 Set the amount by how much signal will be amplified after processing.
5293 Default is 1. Range is from 1 to 64.
5296 Curve the sharp knee around the threshold to enter gain reduction more softly.
5297 Default is 2.82843. Range is between 1 and 8.
5300 Choose if the @code{average} level between all channels of side-chain stream
5301 or the louder(@code{maximum}) channel of side-chain stream affects the
5302 reduction. Default is @code{average}.
5305 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5306 of @code{rms}. Default is @code{rms} which is mainly smoother.
5309 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5312 How much to use compressed signal in output. Default is 1.
5313 Range is between 0 and 1.
5316 @subsection Commands
5318 This filter supports the all above options as @ref{commands}.
5320 @subsection Examples
5324 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5325 depending on the signal of 2nd input and later compressed signal to be
5326 merged with 2nd input:
5328 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5332 @section sidechaingate
5334 A sidechain gate acts like a normal (wideband) gate but has the ability to
5335 filter the detected signal before sending it to the gain reduction stage.
5336 Normally a gate uses the full range signal to detect a level above the
5338 For example: If you cut all lower frequencies from your sidechain signal
5339 the gate will decrease the volume of your track only if not enough highs
5340 appear. With this technique you are able to reduce the resonation of a
5341 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5343 It needs two input streams and returns one output stream.
5344 First input stream will be processed depending on second stream signal.
5346 The filter accepts the following options:
5350 Set input level before filtering.
5351 Default is 1. Allowed range is from 0.015625 to 64.
5354 Set the mode of operation. Can be @code{upward} or @code{downward}.
5355 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5356 will be amplified, expanding dynamic range in upward direction.
5357 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5360 Set the level of gain reduction when the signal is below the threshold.
5361 Default is 0.06125. Allowed range is from 0 to 1.
5362 Setting this to 0 disables reduction and then filter behaves like expander.
5365 If a signal rises above this level the gain reduction is released.
5366 Default is 0.125. Allowed range is from 0 to 1.
5369 Set a ratio about which the signal is reduced.
5370 Default is 2. Allowed range is from 1 to 9000.
5373 Amount of milliseconds the signal has to rise above the threshold before gain
5375 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5378 Amount of milliseconds the signal has to fall below the threshold before the
5379 reduction is increased again. Default is 250 milliseconds.
5380 Allowed range is from 0.01 to 9000.
5383 Set amount of amplification of signal after processing.
5384 Default is 1. Allowed range is from 1 to 64.
5387 Curve the sharp knee around the threshold to enter gain reduction more softly.
5388 Default is 2.828427125. Allowed range is from 1 to 8.
5391 Choose if exact signal should be taken for detection or an RMS like one.
5392 Default is rms. Can be peak or rms.
5395 Choose if the average level between all channels or the louder channel affects
5397 Default is average. Can be average or maximum.
5400 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5403 @subsection Commands
5405 This filter supports the all above options as @ref{commands}.
5407 @section silencedetect
5409 Detect silence in an audio stream.
5411 This filter logs a message when it detects that the input audio volume is less
5412 or equal to a noise tolerance value for a duration greater or equal to the
5413 minimum detected noise duration.
5415 The printed times and duration are expressed in seconds. The
5416 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5417 is set on the first frame whose timestamp equals or exceeds the detection
5418 duration and it contains the timestamp of the first frame of the silence.
5420 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5421 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5422 keys are set on the first frame after the silence. If @option{mono} is
5423 enabled, and each channel is evaluated separately, the @code{.X}
5424 suffixed keys are used, and @code{X} corresponds to the channel number.
5426 The filter accepts the following options:
5430 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5431 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5434 Set silence duration until notification (default is 2 seconds). See
5435 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5436 for the accepted syntax.
5439 Process each channel separately, instead of combined. By default is disabled.
5442 @subsection Examples
5446 Detect 5 seconds of silence with -50dB noise tolerance:
5448 silencedetect=n=-50dB:d=5
5452 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5453 tolerance in @file{silence.mp3}:
5455 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5459 @section silenceremove
5461 Remove silence from the beginning, middle or end of the audio.
5463 The filter accepts the following options:
5467 This value is used to indicate if audio should be trimmed at beginning of
5468 the audio. A value of zero indicates no silence should be trimmed from the
5469 beginning. When specifying a non-zero value, it trims audio up until it
5470 finds non-silence. Normally, when trimming silence from beginning of audio
5471 the @var{start_periods} will be @code{1} but it can be increased to higher
5472 values to trim all audio up to specific count of non-silence periods.
5473 Default value is @code{0}.
5475 @item start_duration
5476 Specify the amount of time that non-silence must be detected before it stops
5477 trimming audio. By increasing the duration, bursts of noises can be treated
5478 as silence and trimmed off. Default value is @code{0}.
5480 @item start_threshold
5481 This indicates what sample value should be treated as silence. For digital
5482 audio, a value of @code{0} may be fine but for audio recorded from analog,
5483 you may wish to increase the value to account for background noise.
5484 Can be specified in dB (in case "dB" is appended to the specified value)
5485 or amplitude ratio. Default value is @code{0}.
5488 Specify max duration of silence at beginning that will be kept after
5489 trimming. Default is 0, which is equal to trimming all samples detected
5493 Specify mode of detection of silence end in start of multi-channel audio.
5494 Can be @var{any} or @var{all}. Default is @var{any}.
5495 With @var{any}, any sample that is detected as non-silence will cause
5496 stopped trimming of silence.
5497 With @var{all}, only if all channels are detected as non-silence will cause
5498 stopped trimming of silence.
5501 Set the count for trimming silence from the end of audio.
5502 To remove silence from the middle of a file, specify a @var{stop_periods}
5503 that is negative. This value is then treated as a positive value and is
5504 used to indicate the effect should restart processing as specified by
5505 @var{start_periods}, making it suitable for removing periods of silence
5506 in the middle of the audio.
5507 Default value is @code{0}.
5510 Specify a duration of silence that must exist before audio is not copied any
5511 more. By specifying a higher duration, silence that is wanted can be left in
5513 Default value is @code{0}.
5515 @item stop_threshold
5516 This is the same as @option{start_threshold} but for trimming silence from
5518 Can be specified in dB (in case "dB" is appended to the specified value)
5519 or amplitude ratio. Default value is @code{0}.
5522 Specify max duration of silence at end that will be kept after
5523 trimming. Default is 0, which is equal to trimming all samples detected
5527 Specify mode of detection of silence start in end of multi-channel audio.
5528 Can be @var{any} or @var{all}. Default is @var{any}.
5529 With @var{any}, any sample that is detected as non-silence will cause
5530 stopped trimming of silence.
5531 With @var{all}, only if all channels are detected as non-silence will cause
5532 stopped trimming of silence.
5535 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5536 and works better with digital silence which is exactly 0.
5537 Default value is @code{rms}.
5540 Set duration in number of seconds used to calculate size of window in number
5541 of samples for detecting silence.
5542 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5545 @subsection Examples
5549 The following example shows how this filter can be used to start a recording
5550 that does not contain the delay at the start which usually occurs between
5551 pressing the record button and the start of the performance:
5553 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5557 Trim all silence encountered from beginning to end where there is more than 1
5558 second of silence in audio:
5560 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5564 Trim all digital silence samples, using peak detection, from beginning to end
5565 where there is more than 0 samples of digital silence in audio and digital
5566 silence is detected in all channels at same positions in stream:
5568 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5574 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5575 loudspeakers around the user for binaural listening via headphones (audio
5576 formats up to 9 channels supported).
5577 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5578 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5579 Austrian Academy of Sciences.
5581 To enable compilation of this filter you need to configure FFmpeg with
5582 @code{--enable-libmysofa}.
5584 The filter accepts the following options:
5588 Set the SOFA file used for rendering.
5591 Set gain applied to audio. Value is in dB. Default is 0.
5594 Set rotation of virtual loudspeakers in deg. Default is 0.
5597 Set elevation of virtual speakers in deg. Default is 0.
5600 Set distance in meters between loudspeakers and the listener with near-field
5601 HRTFs. Default is 1.
5604 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5605 processing audio in time domain which is slow.
5606 @var{freq} is processing audio in frequency domain which is fast.
5607 Default is @var{freq}.
5610 Set custom positions of virtual loudspeakers. Syntax for this option is:
5611 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5612 Each virtual loudspeaker is described with short channel name following with
5613 azimuth and elevation in degrees.
5614 Each virtual loudspeaker description is separated by '|'.
5615 For example to override front left and front right channel positions use:
5616 'speakers=FL 45 15|FR 345 15'.
5617 Descriptions with unrecognised channel names are ignored.
5620 Set custom gain for LFE channels. Value is in dB. Default is 0.
5623 Set custom frame size in number of samples. Default is 1024.
5624 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5625 is set to @var{freq}.
5628 Should all IRs be normalized upon importing SOFA file.
5629 By default is enabled.
5632 Should nearest IRs be interpolated with neighbor IRs if exact position
5633 does not match. By default is disabled.
5636 Minphase all IRs upon loading of SOFA file. By default is disabled.
5639 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5642 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5645 @subsection Examples
5649 Using ClubFritz6 sofa file:
5651 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5655 Using ClubFritz12 sofa file and bigger radius with small rotation:
5657 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5661 Similar as above but with custom speaker positions for front left, front right, back left and back right
5662 and also with custom gain:
5664 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5671 This filter expands or compresses each half-cycle of audio samples
5672 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5673 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5675 The filter accepts the following options:
5679 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5680 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5683 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5684 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5685 would be such that local peak value reaches target peak value but never to surpass it and that
5686 ratio between new and previous peak value does not surpass this option value.
5688 @item compression, c
5689 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5690 This option controls maximum local half-cycle of samples compression. This option is used
5691 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5692 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5693 that peak's half-cycle will be compressed by current compression factor.
5696 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5697 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5698 Any half-cycle samples with their local peak value below or same as this option value will be
5699 compressed by current compression factor, otherwise, if greater than threshold value they will be
5700 expanded with expansion factor so that it could reach peak target value but never surpass it.
5703 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5704 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5705 each new half-cycle until it reaches @option{expansion} value.
5706 Setting this options too high may lead to distortions.
5709 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5710 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5711 each new half-cycle until it reaches @option{compression} value.
5714 Specify which channels to filter, by default all available channels are filtered.
5717 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5718 option. When enabled any half-cycle of samples with their local peak value below or same as
5719 @option{threshold} option will be expanded otherwise it will be compressed.
5722 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5723 When disabled each filtered channel gain calculation is independent, otherwise when this option
5724 is enabled the minimum of all possible gains for each filtered channel is used.
5727 @subsection Commands
5729 This filter supports the all above options as @ref{commands}.
5731 @section stereotools
5733 This filter has some handy utilities to manage stereo signals, for converting
5734 M/S stereo recordings to L/R signal while having control over the parameters
5735 or spreading the stereo image of master track.
5737 The filter accepts the following options:
5741 Set input level before filtering for both channels. Defaults is 1.
5742 Allowed range is from 0.015625 to 64.
5745 Set output level after filtering for both channels. Defaults is 1.
5746 Allowed range is from 0.015625 to 64.
5749 Set input balance between both channels. Default is 0.
5750 Allowed range is from -1 to 1.
5753 Set output balance between both channels. Default is 0.
5754 Allowed range is from -1 to 1.
5757 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5758 clipping. Disabled by default.
5761 Mute the left channel. Disabled by default.
5764 Mute the right channel. Disabled by default.
5767 Change the phase of the left channel. Disabled by default.
5770 Change the phase of the right channel. Disabled by default.
5773 Set stereo mode. Available values are:
5777 Left/Right to Left/Right, this is default.
5780 Left/Right to Mid/Side.
5783 Mid/Side to Left/Right.
5786 Left/Right to Left/Left.
5789 Left/Right to Right/Right.
5792 Left/Right to Left + Right.
5795 Left/Right to Right/Left.
5798 Mid/Side to Left/Left.
5801 Mid/Side to Right/Right.
5804 Mid/Side to Right/Left.
5807 Left/Right to Left - Right.
5811 Set level of side signal. Default is 1.
5812 Allowed range is from 0.015625 to 64.
5815 Set balance of side signal. Default is 0.
5816 Allowed range is from -1 to 1.
5819 Set level of the middle signal. Default is 1.
5820 Allowed range is from 0.015625 to 64.
5823 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5826 Set stereo base between mono and inversed channels. Default is 0.
5827 Allowed range is from -1 to 1.
5830 Set delay in milliseconds how much to delay left from right channel and
5831 vice versa. Default is 0. Allowed range is from -20 to 20.
5834 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5837 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5839 @item bmode_in, bmode_out
5840 Set balance mode for balance_in/balance_out option.
5842 Can be one of the following:
5846 Classic balance mode. Attenuate one channel at time.
5847 Gain is raised up to 1.
5850 Similar as classic mode above but gain is raised up to 2.
5853 Equal power distribution, from -6dB to +6dB range.
5857 @subsection Commands
5859 This filter supports the all above options as @ref{commands}.
5861 @subsection Examples
5865 Apply karaoke like effect:
5867 stereotools=mlev=0.015625
5871 Convert M/S signal to L/R:
5873 "stereotools=mode=ms>lr"
5877 @section stereowiden
5879 This filter enhance the stereo effect by suppressing signal common to both
5880 channels and by delaying the signal of left into right and vice versa,
5881 thereby widening the stereo effect.
5883 The filter accepts the following options:
5887 Time in milliseconds of the delay of left signal into right and vice versa.
5888 Default is 20 milliseconds.
5891 Amount of gain in delayed signal into right and vice versa. Gives a delay
5892 effect of left signal in right output and vice versa which gives widening
5893 effect. Default is 0.3.
5896 Cross feed of left into right with inverted phase. This helps in suppressing
5897 the mono. If the value is 1 it will cancel all the signal common to both
5898 channels. Default is 0.3.
5901 Set level of input signal of original channel. Default is 0.8.
5904 @subsection Commands
5906 This filter supports the all above options except @code{delay} as @ref{commands}.
5908 @section superequalizer
5909 Apply 18 band equalizer.
5911 The filter accepts the following options:
5918 Set 131Hz band gain.
5920 Set 185Hz band gain.
5922 Set 262Hz band gain.
5924 Set 370Hz band gain.
5926 Set 523Hz band gain.
5928 Set 740Hz band gain.
5930 Set 1047Hz band gain.
5932 Set 1480Hz band gain.
5934 Set 2093Hz band gain.
5936 Set 2960Hz band gain.
5938 Set 4186Hz band gain.
5940 Set 5920Hz band gain.
5942 Set 8372Hz band gain.
5944 Set 11840Hz band gain.
5946 Set 16744Hz band gain.
5948 Set 20000Hz band gain.
5952 Apply audio surround upmix filter.
5954 This filter allows to produce multichannel output from audio stream.
5956 The filter accepts the following options:
5960 Set output channel layout. By default, this is @var{5.1}.
5962 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5963 for the required syntax.
5966 Set input channel layout. By default, this is @var{stereo}.
5968 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5969 for the required syntax.
5972 Set input volume level. By default, this is @var{1}.
5975 Set output volume level. By default, this is @var{1}.
5978 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5981 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5984 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5987 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5988 In @var{add} mode, LFE channel is created from input audio and added to output.
5989 In @var{sub} mode, LFE channel is created from input audio and added to output but
5990 also all non-LFE output channels are subtracted with output LFE channel.
5993 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5994 Default is @var{90}.
5997 Set front center input volume. By default, this is @var{1}.
6000 Set front center output volume. By default, this is @var{1}.
6003 Set front left input volume. By default, this is @var{1}.
6006 Set front left output volume. By default, this is @var{1}.
6009 Set front right input volume. By default, this is @var{1}.
6012 Set front right output volume. By default, this is @var{1}.
6015 Set side left input volume. By default, this is @var{1}.
6018 Set side left output volume. By default, this is @var{1}.
6021 Set side right input volume. By default, this is @var{1}.
6024 Set side right output volume. By default, this is @var{1}.
6027 Set back left input volume. By default, this is @var{1}.
6030 Set back left output volume. By default, this is @var{1}.
6033 Set back right input volume. By default, this is @var{1}.
6036 Set back right output volume. By default, this is @var{1}.
6039 Set back center input volume. By default, this is @var{1}.
6042 Set back center output volume. By default, this is @var{1}.
6045 Set LFE input volume. By default, this is @var{1}.
6048 Set LFE output volume. By default, this is @var{1}.
6051 Set spread usage of stereo image across X axis for all channels.
6054 Set spread usage of stereo image across Y axis for all channels.
6056 @item fcx, flx, frx, blx, brx, slx, srx, bcx
6057 Set spread usage of stereo image across X axis for each channel.
6059 @item fcy, fly, fry, bly, bry, sly, sry, bcy
6060 Set spread usage of stereo image across Y axis for each channel.
6063 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
6066 Set window function.
6068 It accepts the following values:
6091 Default is @code{hann}.
6094 Set window overlap. If set to 1, the recommended overlap for selected
6095 window function will be picked. Default is @code{0.5}.
6098 @section treble, highshelf
6100 Boost or cut treble (upper) frequencies of the audio using a two-pole
6101 shelving filter with a response similar to that of a standard
6102 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6104 The filter accepts the following options:
6108 Give the gain at whichever is the lower of ~22 kHz and the
6109 Nyquist frequency. Its useful range is about -20 (for a large cut)
6110 to +20 (for a large boost). Beware of clipping when using a positive gain.
6113 Set the filter's central frequency and so can be used
6114 to extend or reduce the frequency range to be boosted or cut.
6115 The default value is @code{3000} Hz.
6118 Set method to specify band-width of filter.
6133 Determine how steep is the filter's shelf transition.
6136 Set number of poles. Default is 2.
6139 How much to use filtered signal in output. Default is 1.
6140 Range is between 0 and 1.
6143 Specify which channels to filter, by default all available are filtered.
6146 Normalize biquad coefficients, by default is disabled.
6147 Enabling it will normalize magnitude response at DC to 0dB.
6150 Set transform type of IIR filter.
6159 Set precison of filtering.
6162 Pick automatic sample format depending on surround filters.
6164 Always use signed 16-bit.
6166 Always use signed 32-bit.
6168 Always use float 32-bit.
6170 Always use float 64-bit.
6174 @subsection Commands
6176 This filter supports the following commands:
6179 Change treble frequency.
6180 Syntax for the command is : "@var{frequency}"
6183 Change treble width_type.
6184 Syntax for the command is : "@var{width_type}"
6187 Change treble width.
6188 Syntax for the command is : "@var{width}"
6192 Syntax for the command is : "@var{gain}"
6196 Syntax for the command is : "@var{mix}"
6201 Sinusoidal amplitude modulation.
6203 The filter accepts the following options:
6207 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6208 (20 Hz or lower) will result in a tremolo effect.
6209 This filter may also be used as a ring modulator by specifying
6210 a modulation frequency higher than 20 Hz.
6211 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6214 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6215 Default value is 0.5.
6220 Sinusoidal phase modulation.
6222 The filter accepts the following options:
6226 Modulation frequency in Hertz.
6227 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6230 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6231 Default value is 0.5.
6236 Adjust the input audio volume.
6238 It accepts the following parameters:
6242 Set audio volume expression.
6244 Output values are clipped to the maximum value.
6246 The output audio volume is given by the relation:
6248 @var{output_volume} = @var{volume} * @var{input_volume}
6251 The default value for @var{volume} is "1.0".
6254 This parameter represents the mathematical precision.
6256 It determines which input sample formats will be allowed, which affects the
6257 precision of the volume scaling.
6261 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6263 32-bit floating-point; this limits input sample format to FLT. (default)
6265 64-bit floating-point; this limits input sample format to DBL.
6269 Choose the behaviour on encountering ReplayGain side data in input frames.
6273 Remove ReplayGain side data, ignoring its contents (the default).
6276 Ignore ReplayGain side data, but leave it in the frame.
6279 Prefer the track gain, if present.
6282 Prefer the album gain, if present.
6285 @item replaygain_preamp
6286 Pre-amplification gain in dB to apply to the selected replaygain gain.
6288 Default value for @var{replaygain_preamp} is 0.0.
6290 @item replaygain_noclip
6291 Prevent clipping by limiting the gain applied.
6293 Default value for @var{replaygain_noclip} is 1.
6296 Set when the volume expression is evaluated.
6298 It accepts the following values:
6301 only evaluate expression once during the filter initialization, or
6302 when the @samp{volume} command is sent
6305 evaluate expression for each incoming frame
6308 Default value is @samp{once}.
6311 The volume expression can contain the following parameters.
6315 frame number (starting at zero)
6318 @item nb_consumed_samples
6319 number of samples consumed by the filter
6321 number of samples in the current frame
6323 original frame position in the file
6329 PTS at start of stream
6331 time at start of stream
6337 last set volume value
6340 Note that when @option{eval} is set to @samp{once} only the
6341 @var{sample_rate} and @var{tb} variables are available, all other
6342 variables will evaluate to NAN.
6344 @subsection Commands
6346 This filter supports the following commands:
6349 Modify the volume expression.
6350 The command accepts the same syntax of the corresponding option.
6352 If the specified expression is not valid, it is kept at its current
6356 @subsection Examples
6360 Halve the input audio volume:
6364 volume=volume=-6.0206dB
6367 In all the above example the named key for @option{volume} can be
6368 omitted, for example like in:
6374 Increase input audio power by 6 decibels using fixed-point precision:
6376 volume=volume=6dB:precision=fixed
6380 Fade volume after time 10 with an annihilation period of 5 seconds:
6382 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6386 @section volumedetect
6388 Detect the volume of the input video.
6390 The filter has no parameters. The input is not modified. Statistics about
6391 the volume will be printed in the log when the input stream end is reached.
6393 In particular it will show the mean volume (root mean square), maximum
6394 volume (on a per-sample basis), and the beginning of a histogram of the
6395 registered volume values (from the maximum value to a cumulated 1/1000 of
6398 All volumes are in decibels relative to the maximum PCM value.
6400 @subsection Examples
6402 Here is an excerpt of the output:
6404 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6405 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6406 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6407 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6408 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6409 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6410 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6411 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6412 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6418 The mean square energy is approximately -27 dB, or 10^-2.7.
6420 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6422 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6425 In other words, raising the volume by +4 dB does not cause any clipping,
6426 raising it by +5 dB causes clipping for 6 samples, etc.
6428 @c man end AUDIO FILTERS
6430 @chapter Audio Sources
6431 @c man begin AUDIO SOURCES
6433 Below is a description of the currently available audio sources.
6437 Buffer audio frames, and make them available to the filter chain.
6439 This source is mainly intended for a programmatic use, in particular
6440 through the interface defined in @file{libavfilter/buffersrc.h}.
6442 It accepts the following parameters:
6446 The timebase which will be used for timestamps of submitted frames. It must be
6447 either a floating-point number or in @var{numerator}/@var{denominator} form.
6450 The sample rate of the incoming audio buffers.
6453 The sample format of the incoming audio buffers.
6454 Either a sample format name or its corresponding integer representation from
6455 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6457 @item channel_layout
6458 The channel layout of the incoming audio buffers.
6459 Either a channel layout name from channel_layout_map in
6460 @file{libavutil/channel_layout.c} or its corresponding integer representation
6461 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6464 The number of channels of the incoming audio buffers.
6465 If both @var{channels} and @var{channel_layout} are specified, then they
6470 @subsection Examples
6473 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6476 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6477 Since the sample format with name "s16p" corresponds to the number
6478 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6481 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6486 Generate an audio signal specified by an expression.
6488 This source accepts in input one or more expressions (one for each
6489 channel), which are evaluated and used to generate a corresponding
6492 This source accepts the following options:
6496 Set the '|'-separated expressions list for each separate channel. In case the
6497 @option{channel_layout} option is not specified, the selected channel layout
6498 depends on the number of provided expressions. Otherwise the last
6499 specified expression is applied to the remaining output channels.
6501 @item channel_layout, c
6502 Set the channel layout. The number of channels in the specified layout
6503 must be equal to the number of specified expressions.
6506 Set the minimum duration of the sourced audio. See
6507 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6508 for the accepted syntax.
6509 Note that the resulting duration may be greater than the specified
6510 duration, as the generated audio is always cut at the end of a
6513 If not specified, or the expressed duration is negative, the audio is
6514 supposed to be generated forever.
6517 Set the number of samples per channel per each output frame,
6520 @item sample_rate, s
6521 Specify the sample rate, default to 44100.
6524 Each expression in @var{exprs} can contain the following constants:
6528 number of the evaluated sample, starting from 0
6531 time of the evaluated sample expressed in seconds, starting from 0
6538 @subsection Examples
6548 Generate a sin signal with frequency of 440 Hz, set sample rate to
6551 aevalsrc="sin(440*2*PI*t):s=8000"
6555 Generate a two channels signal, specify the channel layout (Front
6556 Center + Back Center) explicitly:
6558 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6562 Generate white noise:
6564 aevalsrc="-2+random(0)"
6568 Generate an amplitude modulated signal:
6570 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6574 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6576 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6583 Generate a FIR coefficients using frequency sampling method.
6585 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6587 The filter accepts the following options:
6591 Set number of filter coefficents in output audio stream.
6592 Default value is 1025.
6595 Set frequency points from where magnitude and phase are set.
6596 This must be in non decreasing order, and first element must be 0, while last element
6597 must be 1. Elements are separated by white spaces.
6600 Set magnitude value for every frequency point set by @option{frequency}.
6601 Number of values must be same as number of frequency points.
6602 Values are separated by white spaces.
6605 Set phase value for every frequency point set by @option{frequency}.
6606 Number of values must be same as number of frequency points.
6607 Values are separated by white spaces.
6609 @item sample_rate, r
6610 Set sample rate, default is 44100.
6613 Set number of samples per each frame. Default is 1024.
6616 Set window function. Default is blackman.
6621 The null audio source, return unprocessed audio frames. It is mainly useful
6622 as a template and to be employed in analysis / debugging tools, or as
6623 the source for filters which ignore the input data (for example the sox
6626 This source accepts the following options:
6630 @item channel_layout, cl
6632 Specifies the channel layout, and can be either an integer or a string
6633 representing a channel layout. The default value of @var{channel_layout}
6636 Check the channel_layout_map definition in
6637 @file{libavutil/channel_layout.c} for the mapping between strings and
6638 channel layout values.
6640 @item sample_rate, r
6641 Specifies the sample rate, and defaults to 44100.
6644 Set the number of samples per requested frames.
6647 Set the duration of the sourced audio. See
6648 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6649 for the accepted syntax.
6651 If not specified, or the expressed duration is negative, the audio is
6652 supposed to be generated forever.
6655 @subsection Examples
6659 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6661 anullsrc=r=48000:cl=4
6665 Do the same operation with a more obvious syntax:
6667 anullsrc=r=48000:cl=mono
6671 All the parameters need to be explicitly defined.
6675 Synthesize a voice utterance using the libflite library.
6677 To enable compilation of this filter you need to configure FFmpeg with
6678 @code{--enable-libflite}.
6680 Note that versions of the flite library prior to 2.0 are not thread-safe.
6682 The filter accepts the following options:
6687 If set to 1, list the names of the available voices and exit
6688 immediately. Default value is 0.
6691 Set the maximum number of samples per frame. Default value is 512.
6694 Set the filename containing the text to speak.
6697 Set the text to speak.
6700 Set the voice to use for the speech synthesis. Default value is
6701 @code{kal}. See also the @var{list_voices} option.
6704 @subsection Examples
6708 Read from file @file{speech.txt}, and synthesize the text using the
6709 standard flite voice:
6711 flite=textfile=speech.txt
6715 Read the specified text selecting the @code{slt} voice:
6717 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6721 Input text to ffmpeg:
6723 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6727 Make @file{ffplay} speak the specified text, using @code{flite} and
6728 the @code{lavfi} device:
6730 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6734 For more information about libflite, check:
6735 @url{http://www.festvox.org/flite/}
6739 Generate a noise audio signal.
6741 The filter accepts the following options:
6744 @item sample_rate, r
6745 Specify the sample rate. Default value is 48000 Hz.
6748 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6752 Specify the duration of the generated audio stream. Not specifying this option
6753 results in noise with an infinite length.
6755 @item color, colour, c
6756 Specify the color of noise. Available noise colors are white, pink, brown,
6757 blue, violet and velvet. Default color is white.
6760 Specify a value used to seed the PRNG.
6763 Set the number of samples per each output frame, default is 1024.
6766 @subsection Examples
6771 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6773 anoisesrc=d=60:c=pink:r=44100:a=0.5
6779 Generate odd-tap Hilbert transform FIR coefficients.
6781 The resulting stream can be used with @ref{afir} filter for phase-shifting
6782 the signal by 90 degrees.
6784 This is used in many matrix coding schemes and for analytic signal generation.
6785 The process is often written as a multiplication by i (or j), the imaginary unit.
6787 The filter accepts the following options:
6791 @item sample_rate, s
6792 Set sample rate, default is 44100.
6795 Set length of FIR filter, default is 22051.
6798 Set number of samples per each frame.
6801 Set window function to be used when generating FIR coefficients.
6806 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6808 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6810 The filter accepts the following options:
6813 @item sample_rate, r
6814 Set sample rate, default is 44100.
6817 Set number of samples per each frame. Default is 1024.
6820 Set high-pass frequency. Default is 0.
6823 Set low-pass frequency. Default is 0.
6824 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6825 is higher than 0 then filter will create band-pass filter coefficients,
6826 otherwise band-reject filter coefficients.
6829 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6832 Set Kaiser window beta.
6835 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6838 Enable rounding, by default is disabled.
6841 Set number of taps for high-pass filter.
6844 Set number of taps for low-pass filter.
6849 Generate an audio signal made of a sine wave with amplitude 1/8.
6851 The audio signal is bit-exact.
6853 The filter accepts the following options:
6858 Set the carrier frequency. Default is 440 Hz.
6860 @item beep_factor, b
6861 Enable a periodic beep every second with frequency @var{beep_factor} times
6862 the carrier frequency. Default is 0, meaning the beep is disabled.
6864 @item sample_rate, r
6865 Specify the sample rate, default is 44100.
6868 Specify the duration of the generated audio stream.
6870 @item samples_per_frame
6871 Set the number of samples per output frame.
6873 The expression can contain the following constants:
6877 The (sequential) number of the output audio frame, starting from 0.
6880 The PTS (Presentation TimeStamp) of the output audio frame,
6881 expressed in @var{TB} units.
6884 The PTS of the output audio frame, expressed in seconds.
6887 The timebase of the output audio frames.
6890 Default is @code{1024}.
6893 @subsection Examples
6898 Generate a simple 440 Hz sine wave:
6904 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6908 sine=frequency=220:beep_factor=4:duration=5
6912 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6915 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6919 @c man end AUDIO SOURCES
6921 @chapter Audio Sinks
6922 @c man begin AUDIO SINKS
6924 Below is a description of the currently available audio sinks.
6926 @section abuffersink
6928 Buffer audio frames, and make them available to the end of filter chain.
6930 This sink is mainly intended for programmatic use, in particular
6931 through the interface defined in @file{libavfilter/buffersink.h}
6932 or the options system.
6934 It accepts a pointer to an AVABufferSinkContext structure, which
6935 defines the incoming buffers' formats, to be passed as the opaque
6936 parameter to @code{avfilter_init_filter} for initialization.
6939 Null audio sink; do absolutely nothing with the input audio. It is
6940 mainly useful as a template and for use in analysis / debugging
6943 @c man end AUDIO SINKS
6945 @chapter Video Filters
6946 @c man begin VIDEO FILTERS
6948 When you configure your FFmpeg build, you can disable any of the
6949 existing filters using @code{--disable-filters}.
6950 The configure output will show the video filters included in your
6953 Below is a description of the currently available video filters.
6957 Mark a region of interest in a video frame.
6959 The frame data is passed through unchanged, but metadata is attached
6960 to the frame indicating regions of interest which can affect the
6961 behaviour of later encoding. Multiple regions can be marked by
6962 applying the filter multiple times.
6966 Region distance in pixels from the left edge of the frame.
6968 Region distance in pixels from the top edge of the frame.
6970 Region width in pixels.
6972 Region height in pixels.
6974 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6975 and may contain the following variables:
6978 Width of the input frame.
6980 Height of the input frame.
6984 Quantisation offset to apply within the region.
6986 This must be a real value in the range -1 to +1. A value of zero
6987 indicates no quality change. A negative value asks for better quality
6988 (less quantisation), while a positive value asks for worse quality
6989 (greater quantisation).
6991 The range is calibrated so that the extreme values indicate the
6992 largest possible offset - if the rest of the frame is encoded with the
6993 worst possible quality, an offset of -1 indicates that this region
6994 should be encoded with the best possible quality anyway. Intermediate
6995 values are then interpolated in some codec-dependent way.
6997 For example, in 10-bit H.264 the quantisation parameter varies between
6998 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6999 this region should be encoded with a QP around one-tenth of the full
7000 range better than the rest of the frame. So, if most of the frame
7001 were to be encoded with a QP of around 30, this region would get a QP
7002 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
7003 An extreme value of -1 would indicate that this region should be
7004 encoded with the best possible quality regardless of the treatment of
7005 the rest of the frame - that is, should be encoded at a QP of -12.
7007 If set to true, remove any existing regions of interest marked on the
7008 frame before adding the new one.
7011 @subsection Examples
7015 Mark the centre quarter of the frame as interesting.
7017 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
7020 Mark the 100-pixel-wide region on the left edge of the frame as very
7021 uninteresting (to be encoded at much lower quality than the rest of
7024 addroi=0:0:100:ih:+1/5
7028 @section alphaextract
7030 Extract the alpha component from the input as a grayscale video. This
7031 is especially useful with the @var{alphamerge} filter.
7035 Add or replace the alpha component of the primary input with the
7036 grayscale value of a second input. This is intended for use with
7037 @var{alphaextract} to allow the transmission or storage of frame
7038 sequences that have alpha in a format that doesn't support an alpha
7041 For example, to reconstruct full frames from a normal YUV-encoded video
7042 and a separate video created with @var{alphaextract}, you might use:
7044 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
7049 Amplify differences between current pixel and pixels of adjacent frames in
7050 same pixel location.
7052 This filter accepts the following options:
7056 Set frame radius. Default is 2. Allowed range is from 1 to 63.
7057 For example radius of 3 will instruct filter to calculate average of 7 frames.
7060 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
7063 Set threshold for difference amplification. Any difference greater or equal to
7064 this value will not alter source pixel. Default is 10.
7065 Allowed range is from 0 to 65535.
7068 Set tolerance for difference amplification. Any difference lower to
7069 this value will not alter source pixel. Default is 0.
7070 Allowed range is from 0 to 65535.
7073 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7074 This option controls maximum possible value that will decrease source pixel value.
7077 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7078 This option controls maximum possible value that will increase source pixel value.
7081 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
7084 @subsection Commands
7086 This filter supports the following @ref{commands} that corresponds to option of same name:
7098 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7099 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7100 Substation Alpha) subtitles files.
7102 This filter accepts the following option in addition to the common options from
7103 the @ref{subtitles} filter:
7107 Set the shaping engine
7109 Available values are:
7112 The default libass shaping engine, which is the best available.
7114 Fast, font-agnostic shaper that can do only substitutions
7116 Slower shaper using OpenType for substitutions and positioning
7119 The default is @code{auto}.
7123 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7125 The filter accepts the following options:
7129 Set threshold A for 1st plane. Default is 0.02.
7130 Valid range is 0 to 0.3.
7133 Set threshold B for 1st plane. Default is 0.04.
7134 Valid range is 0 to 5.
7137 Set threshold A for 2nd plane. Default is 0.02.
7138 Valid range is 0 to 0.3.
7141 Set threshold B for 2nd plane. Default is 0.04.
7142 Valid range is 0 to 5.
7145 Set threshold A for 3rd plane. Default is 0.02.
7146 Valid range is 0 to 0.3.
7149 Set threshold B for 3rd plane. Default is 0.04.
7150 Valid range is 0 to 5.
7152 Threshold A is designed to react on abrupt changes in the input signal and
7153 threshold B is designed to react on continuous changes in the input signal.
7156 Set number of frames filter will use for averaging. Default is 9. Must be odd
7157 number in range [5, 129].
7160 Set what planes of frame filter will use for averaging. Default is all.
7163 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7164 Alternatively can be set to @code{s} serial.
7166 Parallel can be faster then serial, while other way around is never true.
7167 Parallel will abort early on first change being greater then thresholds, while serial
7168 will continue processing other side of frames if they are equal or below thresholds.
7173 Set sigma for 1st plane, 2nd plane or 3rd plane. Default is 32767.
7174 Valid range is from 0 to 32767.
7175 This options controls weight for each pixel in radius defined by size.
7176 Default value means every pixel have same weight.
7177 Setting this option to 0 effectively disables filtering.
7180 @subsection Commands
7181 This filter supports same @ref{commands} as options except option @code{s}.
7182 The command accepts the same syntax of the corresponding option.
7186 Apply average blur filter.
7188 The filter accepts the following options:
7192 Set horizontal radius size.
7195 Set which planes to filter. By default all planes are filtered.
7198 Set vertical radius size, if zero it will be same as @code{sizeX}.
7199 Default is @code{0}.
7202 @subsection Commands
7203 This filter supports same commands as options.
7204 The command accepts the same syntax of the corresponding option.
7206 If the specified expression is not valid, it is kept at its current
7211 Compute the bounding box for the non-black pixels in the input frame
7214 This filter computes the bounding box containing all the pixels with a
7215 luminance value greater than the minimum allowed value.
7216 The parameters describing the bounding box are printed on the filter
7219 The filter accepts the following option:
7223 Set the minimal luminance value. Default is @code{16}.
7226 @subsection Commands
7228 This filter supports the all above options as @ref{commands}.
7231 Apply bilateral filter, spatial smoothing while preserving edges.
7233 The filter accepts the following options:
7236 Set sigma of gaussian function to calculate spatial weight.
7237 Allowed range is 0 to 512. Default is 0.1.
7240 Set sigma of gaussian function to calculate range weight.
7241 Allowed range is 0 to 1. Default is 0.1.
7244 Set planes to filter. Default is first only.
7247 @subsection Commands
7249 This filter supports the all above options as @ref{commands}.
7251 @section bitplanenoise
7253 Show and measure bit plane noise.
7255 The filter accepts the following options:
7259 Set which plane to analyze. Default is @code{1}.
7262 Filter out noisy pixels from @code{bitplane} set above.
7263 Default is disabled.
7266 @section blackdetect
7268 Detect video intervals that are (almost) completely black. Can be
7269 useful to detect chapter transitions, commercials, or invalid
7272 The filter outputs its detection analysis to both the log as well as
7273 frame metadata. If a black segment of at least the specified minimum
7274 duration is found, a line with the start and end timestamps as well
7275 as duration is printed to the log with level @code{info}. In addition,
7276 a log line with level @code{debug} is printed per frame showing the
7277 black amount detected for that frame.
7279 The filter also attaches metadata to the first frame of a black
7280 segment with key @code{lavfi.black_start} and to the first frame
7281 after the black segment ends with key @code{lavfi.black_end}. The
7282 value is the frame's timestamp. This metadata is added regardless
7283 of the minimum duration specified.
7285 The filter accepts the following options:
7288 @item black_min_duration, d
7289 Set the minimum detected black duration expressed in seconds. It must
7290 be a non-negative floating point number.
7292 Default value is 2.0.
7294 @item picture_black_ratio_th, pic_th
7295 Set the threshold for considering a picture "black".
7296 Express the minimum value for the ratio:
7298 @var{nb_black_pixels} / @var{nb_pixels}
7301 for which a picture is considered black.
7302 Default value is 0.98.
7304 @item pixel_black_th, pix_th
7305 Set the threshold for considering a pixel "black".
7307 The threshold expresses the maximum pixel luminance value for which a
7308 pixel is considered "black". The provided value is scaled according to
7309 the following equation:
7311 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7314 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7315 the input video format, the range is [0-255] for YUV full-range
7316 formats and [16-235] for YUV non full-range formats.
7318 Default value is 0.10.
7321 The following example sets the maximum pixel threshold to the minimum
7322 value, and detects only black intervals of 2 or more seconds:
7324 blackdetect=d=2:pix_th=0.00
7329 Detect frames that are (almost) completely black. Can be useful to
7330 detect chapter transitions or commercials. Output lines consist of
7331 the frame number of the detected frame, the percentage of blackness,
7332 the position in the file if known or -1 and the timestamp in seconds.
7334 In order to display the output lines, you need to set the loglevel at
7335 least to the AV_LOG_INFO value.
7337 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7338 The value represents the percentage of pixels in the picture that
7339 are below the threshold value.
7341 It accepts the following parameters:
7346 The percentage of the pixels that have to be below the threshold; it defaults to
7349 @item threshold, thresh
7350 The threshold below which a pixel value is considered black; it defaults to
7358 Blend two video frames into each other.
7360 The @code{blend} filter takes two input streams and outputs one
7361 stream, the first input is the "top" layer and second input is
7362 "bottom" layer. By default, the output terminates when the longest input terminates.
7364 The @code{tblend} (time blend) filter takes two consecutive frames
7365 from one single stream, and outputs the result obtained by blending
7366 the new frame on top of the old frame.
7368 A description of the accepted options follows.
7376 Set blend mode for specific pixel component or all pixel components in case
7377 of @var{all_mode}. Default value is @code{normal}.
7379 Available values for component modes are:
7421 Set blend opacity for specific pixel component or all pixel components in case
7422 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7429 Set blend expression for specific pixel component or all pixel components in case
7430 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7432 The expressions can use the following variables:
7436 The sequential number of the filtered frame, starting from @code{0}.
7440 the coordinates of the current sample
7444 the width and height of currently filtered plane
7448 Width and height scale for the plane being filtered. It is the
7449 ratio between the dimensions of the current plane to the luma plane,
7450 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7451 the luma plane and @code{0.5,0.5} for the chroma planes.
7454 Time of the current frame, expressed in seconds.
7457 Value of pixel component at current location for first video frame (top layer).
7460 Value of pixel component at current location for second video frame (bottom layer).
7464 The @code{blend} filter also supports the @ref{framesync} options.
7466 @subsection Examples
7470 Apply transition from bottom layer to top layer in first 10 seconds:
7472 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7476 Apply linear horizontal transition from top layer to bottom layer:
7478 blend=all_expr='A*(X/W)+B*(1-X/W)'
7482 Apply 1x1 checkerboard effect:
7484 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7488 Apply uncover left effect:
7490 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7494 Apply uncover down effect:
7496 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7500 Apply uncover up-left effect:
7502 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7506 Split diagonally video and shows top and bottom layer on each side:
7508 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7512 Display differences between the current and the previous frame:
7514 tblend=all_mode=grainextract
7518 @subsection Commands
7519 This filter supports same @ref{commands} as options.
7523 Denoise frames using Block-Matching 3D algorithm.
7525 The filter accepts the following options.
7529 Set denoising strength. Default value is 1.
7530 Allowed range is from 0 to 999.9.
7531 The denoising algorithm is very sensitive to sigma, so adjust it
7532 according to the source.
7535 Set local patch size. This sets dimensions in 2D.
7538 Set sliding step for processing blocks. Default value is 4.
7539 Allowed range is from 1 to 64.
7540 Smaller values allows processing more reference blocks and is slower.
7543 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7544 When set to 1, no block matching is done. Larger values allows more blocks
7546 Allowed range is from 1 to 256.
7549 Set radius for search block matching. Default is 9.
7550 Allowed range is from 1 to INT32_MAX.
7553 Set step between two search locations for block matching. Default is 1.
7554 Allowed range is from 1 to 64. Smaller is slower.
7557 Set threshold of mean square error for block matching. Valid range is 0 to
7561 Set thresholding parameter for hard thresholding in 3D transformed domain.
7562 Larger values results in stronger hard-thresholding filtering in frequency
7566 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7567 Default is @code{basic}.
7570 If enabled, filter will use 2nd stream for block matching.
7571 Default is disabled for @code{basic} value of @var{estim} option,
7572 and always enabled if value of @var{estim} is @code{final}.
7575 Set planes to filter. Default is all available except alpha.
7578 @subsection Examples
7582 Basic filtering with bm3d:
7584 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7588 Same as above, but filtering only luma:
7590 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7594 Same as above, but with both estimation modes:
7596 split[a][b],[a]bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic[a],[b][a]bm3d=sigma=3:block=4:bstep=2:group=16:estim=final:ref=1
7600 Same as above, but prefilter with @ref{nlmeans} filter instead:
7602 split[a][b],[a]nlmeans=s=3:r=7:p=3[a],[b][a]bm3d=sigma=3:block=4:bstep=2:group=16:estim=final:ref=1
7608 Apply a boxblur algorithm to the input video.
7610 It accepts the following parameters:
7614 @item luma_radius, lr
7615 @item luma_power, lp
7616 @item chroma_radius, cr
7617 @item chroma_power, cp
7618 @item alpha_radius, ar
7619 @item alpha_power, ap
7623 A description of the accepted options follows.
7626 @item luma_radius, lr
7627 @item chroma_radius, cr
7628 @item alpha_radius, ar
7629 Set an expression for the box radius in pixels used for blurring the
7630 corresponding input plane.
7632 The radius value must be a non-negative number, and must not be
7633 greater than the value of the expression @code{min(w,h)/2} for the
7634 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7637 Default value for @option{luma_radius} is "2". If not specified,
7638 @option{chroma_radius} and @option{alpha_radius} default to the
7639 corresponding value set for @option{luma_radius}.
7641 The expressions can contain the following constants:
7645 The input width and height in pixels.
7649 The input chroma image width and height in pixels.
7653 The horizontal and vertical chroma subsample values. For example, for the
7654 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7657 @item luma_power, lp
7658 @item chroma_power, cp
7659 @item alpha_power, ap
7660 Specify how many times the boxblur filter is applied to the
7661 corresponding plane.
7663 Default value for @option{luma_power} is 2. If not specified,
7664 @option{chroma_power} and @option{alpha_power} default to the
7665 corresponding value set for @option{luma_power}.
7667 A value of 0 will disable the effect.
7670 @subsection Examples
7674 Apply a boxblur filter with the luma, chroma, and alpha radii
7677 boxblur=luma_radius=2:luma_power=1
7682 Set the luma radius to 2, and alpha and chroma radius to 0:
7684 boxblur=2:1:cr=0:ar=0
7688 Set the luma and chroma radii to a fraction of the video dimension:
7690 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7696 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7697 Deinterlacing Filter").
7699 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7700 interpolation algorithms.
7701 It accepts the following parameters:
7705 The interlacing mode to adopt. It accepts one of the following values:
7709 Output one frame for each frame.
7711 Output one frame for each field.
7714 The default value is @code{send_field}.
7717 The picture field parity assumed for the input interlaced video. It accepts one
7718 of the following values:
7722 Assume the top field is first.
7724 Assume the bottom field is first.
7726 Enable automatic detection of field parity.
7729 The default value is @code{auto}.
7730 If the interlacing is unknown or the decoder does not export this information,
7731 top field first will be assumed.
7734 Specify which frames to deinterlace. Accepts one of the following
7739 Deinterlace all frames.
7741 Only deinterlace frames marked as interlaced.
7744 The default value is @code{all}.
7749 Apply Contrast Adaptive Sharpen filter to video stream.
7751 The filter accepts the following options:
7755 Set the sharpening strength. Default value is 0.
7758 Set planes to filter. Default value is to filter all
7759 planes except alpha plane.
7762 @subsection Commands
7763 This filter supports same @ref{commands} as options.
7766 Remove all color information for all colors except for certain one.
7768 The filter accepts the following options:
7772 The color which will not be replaced with neutral chroma.
7775 Similarity percentage with the above color.
7776 0.01 matches only the exact key color, while 1.0 matches everything.
7780 0.0 makes pixels either fully gray, or not gray at all.
7781 Higher values result in more preserved color.
7784 Signals that the color passed is already in YUV instead of RGB.
7786 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7787 This can be used to pass exact YUV values as hexadecimal numbers.
7790 @subsection Commands
7791 This filter supports same @ref{commands} as options.
7792 The command accepts the same syntax of the corresponding option.
7794 If the specified expression is not valid, it is kept at its current
7798 YUV colorspace color/chroma keying.
7800 The filter accepts the following options:
7804 The color which will be replaced with transparency.
7807 Similarity percentage with the key color.
7809 0.01 matches only the exact key color, while 1.0 matches everything.
7814 0.0 makes pixels either fully transparent, or not transparent at all.
7816 Higher values result in semi-transparent pixels, with a higher transparency
7817 the more similar the pixels color is to the key color.
7820 Signals that the color passed is already in YUV instead of RGB.
7822 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7823 This can be used to pass exact YUV values as hexadecimal numbers.
7826 @subsection Commands
7827 This filter supports same @ref{commands} as options.
7828 The command accepts the same syntax of the corresponding option.
7830 If the specified expression is not valid, it is kept at its current
7833 @subsection Examples
7837 Make every green pixel in the input image transparent:
7839 ffmpeg -i input.png -vf chromakey=green out.png
7843 Overlay a greenscreen-video on top of a static black background.
7845 ffmpeg -f lavfi -i color=c=black:s=1280x720 -i video.mp4 -shortest -filter_complex "[1:v]chromakey=0x70de77:0.1:0.2[ckout];[0:v][ckout]overlay[out]" -map "[out]" output.mkv
7850 Reduce chrominance noise.
7852 The filter accepts the following options:
7856 Set threshold for averaging chrominance values.
7857 Sum of absolute difference of Y, U and V pixel components of current
7858 pixel and neighbour pixels lower than this threshold will be used in
7859 averaging. Luma component is left unchanged and is copied to output.
7860 Default value is 30. Allowed range is from 1 to 200.
7863 Set horizontal radius of rectangle used for averaging.
7864 Allowed range is from 1 to 100. Default value is 5.
7867 Set vertical radius of rectangle used for averaging.
7868 Allowed range is from 1 to 100. Default value is 5.
7871 Set horizontal step when averaging. Default value is 1.
7872 Allowed range is from 1 to 50.
7873 Mostly useful to speed-up filtering.
7876 Set vertical step when averaging. Default value is 1.
7877 Allowed range is from 1 to 50.
7878 Mostly useful to speed-up filtering.
7881 Set Y threshold for averaging chrominance values.
7882 Set finer control for max allowed difference between Y components
7883 of current pixel and neigbour pixels.
7884 Default value is 200. Allowed range is from 1 to 200.
7887 Set U threshold for averaging chrominance values.
7888 Set finer control for max allowed difference between U components
7889 of current pixel and neigbour pixels.
7890 Default value is 200. Allowed range is from 1 to 200.
7893 Set V threshold for averaging chrominance values.
7894 Set finer control for max allowed difference between V components
7895 of current pixel and neigbour pixels.
7896 Default value is 200. Allowed range is from 1 to 200.
7899 @subsection Commands
7900 This filter supports same @ref{commands} as options.
7901 The command accepts the same syntax of the corresponding option.
7903 @section chromashift
7904 Shift chroma pixels horizontally and/or vertically.
7906 The filter accepts the following options:
7909 Set amount to shift chroma-blue horizontally.
7911 Set amount to shift chroma-blue vertically.
7913 Set amount to shift chroma-red horizontally.
7915 Set amount to shift chroma-red vertically.
7917 Set edge mode, can be @var{smear}, default, or @var{warp}.
7920 @subsection Commands
7922 This filter supports the all above options as @ref{commands}.
7926 Display CIE color diagram with pixels overlaid onto it.
7928 The filter accepts the following options:
7943 @item uhdtv, rec2020
7957 Set what gamuts to draw.
7959 See @code{system} option for available values.
7962 Set ciescope size, by default set to 512.
7965 Set intensity used to map input pixel values to CIE diagram.
7968 Set contrast used to draw tongue colors that are out of active color system gamut.
7971 Correct gamma displayed on scope, by default enabled.
7974 Show white point on CIE diagram, by default disabled.
7977 Set input gamma. Used only with XYZ input color space.
7982 Visualize information exported by some codecs.
7984 Some codecs can export information through frames using side-data or other
7985 means. For example, some MPEG based codecs export motion vectors through the
7986 @var{export_mvs} flag in the codec @option{flags2} option.
7988 The filter accepts the following option:
7992 Set motion vectors to visualize.
7994 Available flags for @var{mv} are:
7998 forward predicted MVs of P-frames
8000 forward predicted MVs of B-frames
8002 backward predicted MVs of B-frames
8006 Display quantization parameters using the chroma planes.
8009 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
8011 Available flags for @var{mv_type} are:
8015 forward predicted MVs
8017 backward predicted MVs
8020 @item frame_type, ft
8021 Set frame type to visualize motion vectors of.
8023 Available flags for @var{frame_type} are:
8027 intra-coded frames (I-frames)
8029 predicted frames (P-frames)
8031 bi-directionally predicted frames (B-frames)
8035 @subsection Examples
8039 Visualize forward predicted MVs of all frames using @command{ffplay}:
8041 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
8045 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
8047 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
8051 @section colorbalance
8052 Modify intensity of primary colors (red, green and blue) of input frames.
8054 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
8055 regions for the red-cyan, green-magenta or blue-yellow balance.
8057 A positive adjustment value shifts the balance towards the primary color, a negative
8058 value towards the complementary color.
8060 The filter accepts the following options:
8066 Adjust red, green and blue shadows (darkest pixels).
8071 Adjust red, green and blue midtones (medium pixels).
8076 Adjust red, green and blue highlights (brightest pixels).
8078 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8081 Preserve lightness when changing color balance. Default is disabled.
8084 @subsection Examples
8088 Add red color cast to shadows:
8094 @subsection Commands
8096 This filter supports the all above options as @ref{commands}.
8098 @section colorcontrast
8100 Adjust color contrast between RGB components.
8102 The filter accepts the following options:
8106 Set the red-cyan contrast. Defaults is 0.0. Allowed range is from -1.0 to 1.0.
8109 Set the green-magenta contrast. Defaults is 0.0. Allowed range is from -1.0 to 1.0.
8112 Set the blue-yellow contrast. Defaults is 0.0. Allowed range is from -1.0 to 1.0.
8117 Set the weight of each @code{rc}, @code{gm}, @code{by} option value. Default value is 0.0.
8118 Allowed range is from 0.0 to 1.0. If all weights are 0.0 filtering is disabled.
8121 Set the amount of preserving lightness. Default value is 0.0. Allowed range is from 0.0 to 1.0.
8124 @subsection Commands
8126 This filter supports the all above options as @ref{commands}.
8128 @section colorcorrect
8130 Adjust color white balance selectively for blacks and whites.
8131 This filter operates in YUV colorspace.
8133 The filter accepts the following options:
8137 Set the red shadow spot. Allowed range is from -1.0 to 1.0.
8141 Set the blue shadow spot. Allowed range is from -1.0 to 1.0.
8145 Set the red highlight spot. Allowed range is from -1.0 to 1.0.
8149 Set the red highlight spot. Allowed range is from -1.0 to 1.0.
8153 Set the amount of saturation. Allowed range is from -3.0 to 3.0.
8157 @subsection Commands
8159 This filter supports the all above options as @ref{commands}.
8161 @section colorchannelmixer
8163 Adjust video input frames by re-mixing color channels.
8165 This filter modifies a color channel by adding the values associated to
8166 the other channels of the same pixels. For example if the value to
8167 modify is red, the output value will be:
8169 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8172 The filter accepts the following options:
8179 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8180 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8186 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8187 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8193 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8194 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8200 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8201 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8203 Allowed ranges for options are @code{[-2.0, 2.0]}.
8206 Preserve lightness when changing colors. Allowed range is from @code{[0.0, 1.0]}.
8207 Default is @code{0.0}, thus disabled.
8210 @subsection Examples
8214 Convert source to grayscale:
8216 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8219 Simulate sepia tones:
8221 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8225 @subsection Commands
8227 This filter supports the all above options as @ref{commands}.
8230 Overlay a solid color on the video stream.
8232 The filter accepts the following options:
8236 Set the color hue. Allowed range is from 0 to 360.
8240 Set the color saturation. Allowed range is from 0 to 1.
8241 Default value is 0.5.
8244 Set the color lightness. Allowed range is from 0 to 1.
8245 Default value is 0.5.
8248 Set the mix of source lightness. By default is set to 1.0.
8249 Allowed range is from 0.0 to 1.0.
8252 @subsection Commands
8254 This filter supports the all above options as @ref{commands}.
8257 RGB colorspace color keying.
8259 The filter accepts the following options:
8263 The color which will be replaced with transparency.
8266 Similarity percentage with the key color.
8268 0.01 matches only the exact key color, while 1.0 matches everything.
8273 0.0 makes pixels either fully transparent, or not transparent at all.
8275 Higher values result in semi-transparent pixels, with a higher transparency
8276 the more similar the pixels color is to the key color.
8279 @subsection Examples
8283 Make every green pixel in the input image transparent:
8285 ffmpeg -i input.png -vf colorkey=green out.png
8289 Overlay a greenscreen-video on top of a static background image.
8291 ffmpeg -i background.png -i video.mp4 -filter_complex "[1:v]colorkey=0x3BBD1E:0.3:0.2[ckout];[0:v][ckout]overlay[out]" -map "[out]" output.flv
8295 @subsection Commands
8296 This filter supports same @ref{commands} as options.
8297 The command accepts the same syntax of the corresponding option.
8299 If the specified expression is not valid, it is kept at its current
8303 Remove all color information for all RGB colors except for certain one.
8305 The filter accepts the following options:
8309 The color which will not be replaced with neutral gray.
8312 Similarity percentage with the above color.
8313 0.01 matches only the exact key color, while 1.0 matches everything.
8316 Blend percentage. 0.0 makes pixels fully gray.
8317 Higher values result in more preserved color.
8320 @subsection Commands
8321 This filter supports same @ref{commands} as options.
8322 The command accepts the same syntax of the corresponding option.
8324 If the specified expression is not valid, it is kept at its current
8327 @section colorlevels
8329 Adjust video input frames using levels.
8331 The filter accepts the following options:
8338 Adjust red, green, blue and alpha input black point.
8339 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8345 Adjust red, green, blue and alpha input white point.
8346 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8348 Input levels are used to lighten highlights (bright tones), darken shadows
8349 (dark tones), change the balance of bright and dark tones.
8355 Adjust red, green, blue and alpha output black point.
8356 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8362 Adjust red, green, blue and alpha output white point.
8363 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8365 Output levels allows manual selection of a constrained output level range.
8368 @subsection Examples
8372 Make video output darker:
8374 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8380 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8384 Make video output lighter:
8386 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8390 Increase brightness:
8392 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8396 @subsection Commands
8398 This filter supports the all above options as @ref{commands}.
8400 @section colormatrix
8402 Convert color matrix.
8404 The filter accepts the following options:
8409 Specify the source and destination color matrix. Both values must be
8412 The accepted values are:
8440 For example to convert from BT.601 to SMPTE-240M, use the command:
8442 colormatrix=bt601:smpte240m
8447 Convert colorspace, transfer characteristics or color primaries.
8448 Input video needs to have an even size.
8450 The filter accepts the following options:
8455 Specify all color properties at once.
8457 The accepted values are:
8487 Specify output colorspace.
8489 The accepted values are:
8498 BT.470BG or BT.601-6 625
8501 SMPTE-170M or BT.601-6 525
8510 BT.2020 with non-constant luminance
8516 Specify output transfer characteristics.
8518 The accepted values are:
8530 Constant gamma of 2.2
8533 Constant gamma of 2.8
8536 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8554 BT.2020 for 10-bits content
8557 BT.2020 for 12-bits content
8563 Specify output color primaries.
8565 The accepted values are:
8574 BT.470BG or BT.601-6 625
8577 SMPTE-170M or BT.601-6 525
8601 Specify output color range.
8603 The accepted values are:
8606 TV (restricted) range
8609 MPEG (restricted) range
8620 Specify output color format.
8622 The accepted values are:
8625 YUV 4:2:0 planar 8-bits
8628 YUV 4:2:0 planar 10-bits
8631 YUV 4:2:0 planar 12-bits
8634 YUV 4:2:2 planar 8-bits
8637 YUV 4:2:2 planar 10-bits
8640 YUV 4:2:2 planar 12-bits
8643 YUV 4:4:4 planar 8-bits
8646 YUV 4:4:4 planar 10-bits
8649 YUV 4:4:4 planar 12-bits
8654 Do a fast conversion, which skips gamma/primary correction. This will take
8655 significantly less CPU, but will be mathematically incorrect. To get output
8656 compatible with that produced by the colormatrix filter, use fast=1.
8659 Specify dithering mode.
8661 The accepted values are:
8667 Floyd-Steinberg dithering
8671 Whitepoint adaptation mode.
8673 The accepted values are:
8676 Bradford whitepoint adaptation
8679 von Kries whitepoint adaptation
8682 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8686 Override all input properties at once. Same accepted values as @ref{all}.
8689 Override input colorspace. Same accepted values as @ref{space}.
8692 Override input color primaries. Same accepted values as @ref{primaries}.
8695 Override input transfer characteristics. Same accepted values as @ref{trc}.
8698 Override input color range. Same accepted values as @ref{range}.
8702 The filter converts the transfer characteristics, color space and color
8703 primaries to the specified user values. The output value, if not specified,
8704 is set to a default value based on the "all" property. If that property is
8705 also not specified, the filter will log an error. The output color range and
8706 format default to the same value as the input color range and format. The
8707 input transfer characteristics, color space, color primaries and color range
8708 should be set on the input data. If any of these are missing, the filter will
8709 log an error and no conversion will take place.
8711 For example to convert the input to SMPTE-240M, use the command:
8713 colorspace=smpte240m
8716 @section colortemperature
8717 Adjust color temperature in video to simulate variations in ambient color temperature.
8719 The filter accepts the following options:
8723 Set the temperature in Kelvin. Allowed range is from 1000 to 40000.
8724 Default value is 6500 K.
8727 Set mixing with filtered output. Allowed range is from 0 to 1.
8731 Set the amount of preserving lightness. Allowed range is from 0 to 1.
8735 @subsection Commands
8736 This filter supports same @ref{commands} as options.
8738 @section convolution
8740 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8742 The filter accepts the following options:
8749 Set matrix for each plane.
8750 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8751 and from 1 to 49 odd number of signed integers in @var{row} mode.
8757 Set multiplier for calculated value for each plane.
8758 If unset or 0, it will be sum of all matrix elements.
8764 Set bias for each plane. This value is added to the result of the multiplication.
8765 Useful for making the overall image brighter or darker. Default is 0.0.
8771 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8772 Default is @var{square}.
8775 @subsection Commands
8777 This filter supports the all above options as @ref{commands}.
8779 @subsection Examples
8785 convolution="0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0"
8791 convolution="1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1/9:1/9:1/9:1/9"
8797 convolution="0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:5:1:1:1:0:128:128:128"
8803 convolution="0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:5:5:5:1:0:128:128:128"
8807 Apply laplacian edge detector which includes diagonals:
8809 convolution="1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:5:5:5:1:0:128:128:0"
8815 convolution="-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2"
8821 Apply 2D convolution of video stream in frequency domain using second stream
8824 The filter accepts the following options:
8828 Set which planes to process.
8831 Set which impulse video frames will be processed, can be @var{first}
8832 or @var{all}. Default is @var{all}.
8835 The @code{convolve} filter also supports the @ref{framesync} options.
8839 Copy the input video source unchanged to the output. This is mainly useful for
8844 Video filtering on GPU using Apple's CoreImage API on OSX.
8846 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8847 processed by video hardware. However, software-based OpenGL implementations
8848 exist which means there is no guarantee for hardware processing. It depends on
8851 There are many filters and image generators provided by Apple that come with a
8852 large variety of options. The filter has to be referenced by its name along
8855 The coreimage filter accepts the following options:
8858 List all available filters and generators along with all their respective
8859 options as well as possible minimum and maximum values along with the default
8866 Specify all filters by their respective name and options.
8867 Use @var{list_filters} to determine all valid filter names and options.
8868 Numerical options are specified by a float value and are automatically clamped
8869 to their respective value range. Vector and color options have to be specified
8870 by a list of space separated float values. Character escaping has to be done.
8871 A special option name @code{default} is available to use default options for a
8874 It is required to specify either @code{default} or at least one of the filter options.
8875 All omitted options are used with their default values.
8876 The syntax of the filter string is as follows:
8878 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8882 Specify a rectangle where the output of the filter chain is copied into the
8883 input image. It is given by a list of space separated float values:
8885 output_rect=x\ y\ width\ height
8887 If not given, the output rectangle equals the dimensions of the input image.
8888 The output rectangle is automatically cropped at the borders of the input
8889 image. Negative values are valid for each component.
8891 output_rect=25\ 25\ 100\ 100
8895 Several filters can be chained for successive processing without GPU-HOST
8896 transfers allowing for fast processing of complex filter chains.
8897 Currently, only filters with zero (generators) or exactly one (filters) input
8898 image and one output image are supported. Also, transition filters are not yet
8901 Some filters generate output images with additional padding depending on the
8902 respective filter kernel. The padding is automatically removed to ensure the
8903 filter output has the same size as the input image.
8905 For image generators, the size of the output image is determined by the
8906 previous output image of the filter chain or the input image of the whole
8907 filterchain, respectively. The generators do not use the pixel information of
8908 this image to generate their output. However, the generated output is
8909 blended onto this image, resulting in partial or complete coverage of the
8912 The @ref{coreimagesrc} video source can be used for generating input images
8913 which are directly fed into the filter chain. By using it, providing input
8914 images by another video source or an input video is not required.
8916 @subsection Examples
8921 List all filters available:
8923 coreimage=list_filters=true
8927 Use the CIBoxBlur filter with default options to blur an image:
8929 coreimage=filter=CIBoxBlur@@default
8933 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8934 its center at 100x100 and a radius of 50 pixels:
8936 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8940 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8941 given as complete and escaped command-line for Apple's standard bash shell:
8943 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8949 Cover a rectangular object
8951 It accepts the following options:
8955 Filepath of the optional cover image, needs to be in yuv420.
8960 It accepts the following values:
8963 cover it by the supplied image
8965 cover it by interpolating the surrounding pixels
8968 Default value is @var{blur}.
8971 @subsection Examples
8975 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8977 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8983 Crop the input video to given dimensions.
8985 It accepts the following parameters:
8989 The width of the output video. It defaults to @code{iw}.
8990 This expression is evaluated only once during the filter
8991 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8994 The height of the output video. It defaults to @code{ih}.
8995 This expression is evaluated only once during the filter
8996 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8999 The horizontal position, in the input video, of the left edge of the output
9000 video. It defaults to @code{(in_w-out_w)/2}.
9001 This expression is evaluated per-frame.
9004 The vertical position, in the input video, of the top edge of the output video.
9005 It defaults to @code{(in_h-out_h)/2}.
9006 This expression is evaluated per-frame.
9009 If set to 1 will force the output display aspect ratio
9010 to be the same of the input, by changing the output sample aspect
9011 ratio. It defaults to 0.
9014 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
9015 width/height/x/y as specified and will not be rounded to nearest smaller value.
9019 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
9020 expressions containing the following constants:
9025 The computed values for @var{x} and @var{y}. They are evaluated for
9030 The input width and height.
9034 These are the same as @var{in_w} and @var{in_h}.
9038 The output (cropped) width and height.
9042 These are the same as @var{out_w} and @var{out_h}.
9045 same as @var{iw} / @var{ih}
9048 input sample aspect ratio
9051 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
9055 horizontal and vertical chroma subsample values. For example for the
9056 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9059 The number of the input frame, starting from 0.
9062 the position in the file of the input frame, NAN if unknown
9065 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
9069 The expression for @var{out_w} may depend on the value of @var{out_h},
9070 and the expression for @var{out_h} may depend on @var{out_w}, but they
9071 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
9072 evaluated after @var{out_w} and @var{out_h}.
9074 The @var{x} and @var{y} parameters specify the expressions for the
9075 position of the top-left corner of the output (non-cropped) area. They
9076 are evaluated for each frame. If the evaluated value is not valid, it
9077 is approximated to the nearest valid value.
9079 The expression for @var{x} may depend on @var{y}, and the expression
9080 for @var{y} may depend on @var{x}.
9082 @subsection Examples
9086 Crop area with size 100x100 at position (12,34).
9091 Using named options, the example above becomes:
9093 crop=w=100:h=100:x=12:y=34
9097 Crop the central input area with size 100x100:
9103 Crop the central input area with size 2/3 of the input video:
9105 crop=2/3*in_w:2/3*in_h
9109 Crop the input video central square:
9116 Delimit the rectangle with the top-left corner placed at position
9117 100:100 and the right-bottom corner corresponding to the right-bottom
9118 corner of the input image.
9120 crop=in_w-100:in_h-100:100:100
9124 Crop 10 pixels from the left and right borders, and 20 pixels from
9125 the top and bottom borders
9127 crop=in_w-2*10:in_h-2*20
9131 Keep only the bottom right quarter of the input image:
9133 crop=in_w/2:in_h/2:in_w/2:in_h/2
9137 Crop height for getting Greek harmony:
9139 crop=in_w:1/PHI*in_w
9143 Apply trembling effect:
9145 crop=in_w/2:in_h/2:(in_w-out_w)/2+((in_w-out_w)/2)*sin(n/10):(in_h-out_h)/2 +((in_h-out_h)/2)*sin(n/7)
9149 Apply erratic camera effect depending on timestamp:
9151 crop=in_w/2:in_h/2:(in_w-out_w)/2+((in_w-out_w)/2)*sin(t*10):(in_h-out_h)/2 +((in_h-out_h)/2)*sin(t*13)"
9155 Set x depending on the value of y:
9157 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
9161 @subsection Commands
9163 This filter supports the following commands:
9169 Set width/height of the output video and the horizontal/vertical position
9171 The command accepts the same syntax of the corresponding option.
9173 If the specified expression is not valid, it is kept at its current
9179 Auto-detect the crop size.
9181 It calculates the necessary cropping parameters and prints the
9182 recommended parameters via the logging system. The detected dimensions
9183 correspond to the non-black area of the input video.
9185 It accepts the following parameters:
9190 Set higher black value threshold, which can be optionally specified
9191 from nothing (0) to everything (255 for 8-bit based formats). An intensity
9192 value greater to the set value is considered non-black. It defaults to 24.
9193 You can also specify a value between 0.0 and 1.0 which will be scaled depending
9194 on the bitdepth of the pixel format.
9197 The value which the width/height should be divisible by. It defaults to
9198 16. The offset is automatically adjusted to center the video. Use 2 to
9199 get only even dimensions (needed for 4:2:2 video). 16 is best when
9200 encoding to most video codecs.
9203 Set the number of initial frames for which evaluation is skipped.
9204 Default is 2. Range is 0 to INT_MAX.
9206 @item reset_count, reset
9207 Set the counter that determines after how many frames cropdetect will
9208 reset the previously detected largest video area and start over to
9209 detect the current optimal crop area. Default value is 0.
9211 This can be useful when channel logos distort the video area. 0
9212 indicates 'never reset', and returns the largest area encountered during
9219 Delay video filtering until a given wallclock timestamp. The filter first
9220 passes on @option{preroll} amount of frames, then it buffers at most
9221 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9222 it forwards the buffered frames and also any subsequent frames coming in its
9225 The filter can be used synchronize the output of multiple ffmpeg processes for
9226 realtime output devices like decklink. By putting the delay in the filtering
9227 chain and pre-buffering frames the process can pass on data to output almost
9228 immediately after the target wallclock timestamp is reached.
9230 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9236 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9239 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9242 The maximum duration of content to buffer before waiting for the cue expressed
9243 in seconds. Default is 0.
9250 Apply color adjustments using curves.
9252 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9253 component (red, green and blue) has its values defined by @var{N} key points
9254 tied from each other using a smooth curve. The x-axis represents the pixel
9255 values from the input frame, and the y-axis the new pixel values to be set for
9258 By default, a component curve is defined by the two points @var{(0;0)} and
9259 @var{(1;1)}. This creates a straight line where each original pixel value is
9260 "adjusted" to its own value, which means no change to the image.
9262 The filter allows you to redefine these two points and add some more. A new
9263 curve (using a natural cubic spline interpolation) will be define to pass
9264 smoothly through all these new coordinates. The new defined points needs to be
9265 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9266 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9267 the vector spaces, the values will be clipped accordingly.
9269 The filter accepts the following options:
9273 Select one of the available color presets. This option can be used in addition
9274 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9275 options takes priority on the preset values.
9276 Available presets are:
9279 @item color_negative
9282 @item increase_contrast
9284 @item linear_contrast
9285 @item medium_contrast
9287 @item strong_contrast
9290 Default is @code{none}.
9292 Set the master key points. These points will define a second pass mapping. It
9293 is sometimes called a "luminance" or "value" mapping. It can be used with
9294 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9295 post-processing LUT.
9297 Set the key points for the red component.
9299 Set the key points for the green component.
9301 Set the key points for the blue component.
9303 Set the key points for all components (not including master).
9304 Can be used in addition to the other key points component
9305 options. In this case, the unset component(s) will fallback on this
9306 @option{all} setting.
9308 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9310 Save Gnuplot script of the curves in specified file.
9313 To avoid some filtergraph syntax conflicts, each key points list need to be
9314 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9316 @subsection Commands
9318 This filter supports same @ref{commands} as options.
9320 @subsection Examples
9324 Increase slightly the middle level of blue:
9326 curves=blue='0/0 0.5/0.58 1/1'
9332 curves=r='0/0.11 .42/.51 1/0.95':g='0/0 0.50/0.48 1/1':b='0/0.22 .49/.44 1/0.8'
9334 Here we obtain the following coordinates for each components:
9337 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9339 @code{(0;0) (0.50;0.48) (1;1)}
9341 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9345 The previous example can also be achieved with the associated built-in preset:
9347 curves=preset=vintage
9357 Use a Photoshop preset and redefine the points of the green component:
9359 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9363 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9364 and @command{gnuplot}:
9366 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9367 gnuplot -p /tmp/curves.plt
9373 Video data analysis filter.
9375 This filter shows hexadecimal pixel values of part of video.
9377 The filter accepts the following options:
9381 Set output video size.
9384 Set x offset from where to pick pixels.
9387 Set y offset from where to pick pixels.
9390 Set scope mode, can be one of the following:
9393 Draw hexadecimal pixel values with white color on black background.
9396 Draw hexadecimal pixel values with input video pixel color on black
9400 Draw hexadecimal pixel values on color background picked from input video,
9401 the text color is picked in such way so its always visible.
9405 Draw rows and columns numbers on left and top of video.
9408 Set background opacity.
9411 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9414 Set pixel components to display. By default all pixel components are displayed.
9417 @subsection Commands
9419 This filter supports same @ref{commands} as options excluding @code{size} option.
9422 Apply Directional blur filter.
9424 The filter accepts the following options:
9428 Set angle of directional blur. Default is @code{45}.
9431 Set radius of directional blur. Default is @code{5}.
9434 Set which planes to filter. By default all planes are filtered.
9437 @subsection Commands
9438 This filter supports same @ref{commands} as options.
9439 The command accepts the same syntax of the corresponding option.
9441 If the specified expression is not valid, it is kept at its current
9446 Denoise frames using 2D DCT (frequency domain filtering).
9448 This filter is not designed for real time.
9450 The filter accepts the following options:
9454 Set the noise sigma constant.
9456 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9457 coefficient (absolute value) below this threshold with be dropped.
9459 If you need a more advanced filtering, see @option{expr}.
9461 Default is @code{0}.
9464 Set number overlapping pixels for each block. Since the filter can be slow, you
9465 may want to reduce this value, at the cost of a less effective filter and the
9466 risk of various artefacts.
9468 If the overlapping value doesn't permit processing the whole input width or
9469 height, a warning will be displayed and according borders won't be denoised.
9471 Default value is @var{blocksize}-1, which is the best possible setting.
9474 Set the coefficient factor expression.
9476 For each coefficient of a DCT block, this expression will be evaluated as a
9477 multiplier value for the coefficient.
9479 If this is option is set, the @option{sigma} option will be ignored.
9481 The absolute value of the coefficient can be accessed through the @var{c}
9485 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9486 @var{blocksize}, which is the width and height of the processed blocks.
9488 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9489 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9490 on the speed processing. Also, a larger block size does not necessarily means a
9494 @subsection Examples
9496 Apply a denoise with a @option{sigma} of @code{4.5}:
9501 The same operation can be achieved using the expression system:
9503 dctdnoiz=e='gte(c, 4.5*3)'
9506 Violent denoise using a block size of @code{16x16}:
9513 Remove banding artifacts from input video.
9514 It works by replacing banded pixels with average value of referenced pixels.
9516 The filter accepts the following options:
9523 Set banding detection threshold for each plane. Default is 0.02.
9524 Valid range is 0.00003 to 0.5.
9525 If difference between current pixel and reference pixel is less than threshold,
9526 it will be considered as banded.
9529 Banding detection range in pixels. Default is 16. If positive, random number
9530 in range 0 to set value will be used. If negative, exact absolute value
9532 The range defines square of four pixels around current pixel.
9535 Set direction in radians from which four pixel will be compared. If positive,
9536 random direction from 0 to set direction will be picked. If negative, exact of
9537 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9538 will pick only pixels on same row and -PI/2 will pick only pixels on same
9542 If enabled, current pixel is compared with average value of all four
9543 surrounding pixels. The default is enabled. If disabled current pixel is
9544 compared with all four surrounding pixels. The pixel is considered banded
9545 if only all four differences with surrounding pixels are less than threshold.
9548 If enabled, current pixel is changed if and only if all pixel components are banded,
9549 e.g. banding detection threshold is triggered for all color components.
9550 The default is disabled.
9553 @subsection Commands
9555 This filter supports the all above options as @ref{commands}.
9559 Remove blocking artifacts from input video.
9561 The filter accepts the following options:
9565 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9566 This controls what kind of deblocking is applied.
9569 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9575 Set blocking detection thresholds. Allowed range is 0 to 1.
9576 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9577 Using higher threshold gives more deblocking strength.
9578 Setting @var{alpha} controls threshold detection at exact edge of block.
9579 Remaining options controls threshold detection near the edge. Each one for
9580 below/above or left/right. Setting any of those to @var{0} disables
9584 Set planes to filter. Default is to filter all available planes.
9587 @subsection Examples
9591 Deblock using weak filter and block size of 4 pixels.
9593 deblock=filter=weak:block=4
9597 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9598 deblocking more edges.
9600 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9604 Similar as above, but filter only first plane.
9606 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9610 Similar as above, but filter only second and third plane.
9612 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9616 @subsection Commands
9618 This filter supports the all above options as @ref{commands}.
9623 Drop duplicated frames at regular intervals.
9625 The filter accepts the following options:
9629 Set the number of frames from which one will be dropped. Setting this to
9630 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9631 Default is @code{5}.
9634 Set the threshold for duplicate detection. If the difference metric for a frame
9635 is less than or equal to this value, then it is declared as duplicate. Default
9639 Set scene change threshold. Default is @code{15}.
9643 Set the size of the x and y-axis blocks used during metric calculations.
9644 Larger blocks give better noise suppression, but also give worse detection of
9645 small movements. Must be a power of two. Default is @code{32}.
9648 Mark main input as a pre-processed input and activate clean source input
9649 stream. This allows the input to be pre-processed with various filters to help
9650 the metrics calculation while keeping the frame selection lossless. When set to
9651 @code{1}, the first stream is for the pre-processed input, and the second
9652 stream is the clean source from where the kept frames are chosen. Default is
9656 Set whether or not chroma is considered in the metric calculations. Default is
9662 Apply 2D deconvolution of video stream in frequency domain using second stream
9665 The filter accepts the following options:
9669 Set which planes to process.
9672 Set which impulse video frames will be processed, can be @var{first}
9673 or @var{all}. Default is @var{all}.
9676 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9677 and height are not same and not power of 2 or if stream prior to convolving
9681 The @code{deconvolve} filter also supports the @ref{framesync} options.
9685 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9687 It accepts the following options:
9691 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9692 @var{rainbows} for cross-color reduction.
9695 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9698 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9701 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9704 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9709 Apply deflate effect to the video.
9711 This filter replaces the pixel by the local(3x3) average by taking into account
9712 only values lower than the pixel.
9714 It accepts the following options:
9721 Limit the maximum change for each plane, default is 65535.
9722 If 0, plane will remain unchanged.
9725 @subsection Commands
9727 This filter supports the all above options as @ref{commands}.
9731 Remove temporal frame luminance variations.
9733 It accepts the following options:
9737 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9740 Set averaging mode to smooth temporal luminance variations.
9742 Available values are:
9767 Do not actually modify frame. Useful when one only wants metadata.
9772 Remove judder produced by partially interlaced telecined content.
9774 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9775 source was partially telecined content then the output of @code{pullup,dejudder}
9776 will have a variable frame rate. May change the recorded frame rate of the
9777 container. Aside from that change, this filter will not affect constant frame
9780 The option available in this filter is:
9784 Specify the length of the window over which the judder repeats.
9786 Accepts any integer greater than 1. Useful values are:
9790 If the original was telecined from 24 to 30 fps (Film to NTSC).
9793 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9796 If a mixture of the two.
9799 The default is @samp{4}.
9804 Suppress a TV station logo by a simple interpolation of the surrounding
9805 pixels. Just set a rectangle covering the logo and watch it disappear
9806 (and sometimes something even uglier appear - your mileage may vary).
9808 It accepts the following parameters:
9813 Specify the top left corner coordinates of the logo. They must be
9818 Specify the width and height of the logo to clear. They must be
9822 When set to 1, a green rectangle is drawn on the screen to simplify
9823 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9824 The default value is 0.
9826 The rectangle is drawn on the outermost pixels which will be (partly)
9827 replaced with interpolated values. The values of the next pixels
9828 immediately outside this rectangle in each direction will be used to
9829 compute the interpolated pixel values inside the rectangle.
9833 @subsection Examples
9837 Set a rectangle covering the area with top left corner coordinates 0,0
9840 delogo=x=0:y=0:w=100:h=77
9848 Remove the rain in the input image/video by applying the derain methods based on
9849 convolutional neural networks. Supported models:
9853 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9854 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9857 Training as well as model generation scripts are provided in
9858 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9860 Native model files (.model) can be generated from TensorFlow model
9861 files (.pb) by using tools/python/convert.py
9863 The filter accepts the following options:
9867 Specify which filter to use. This option accepts the following values:
9871 Derain filter. To conduct derain filter, you need to use a derain model.
9874 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9876 Default value is @samp{derain}.
9879 Specify which DNN backend to use for model loading and execution. This option accepts
9880 the following values:
9884 Native implementation of DNN loading and execution.
9887 TensorFlow backend. To enable this backend you
9888 need to install the TensorFlow for C library (see
9889 @url{https://www.tensorflow.org/install/lang_c}) and configure FFmpeg with
9890 @code{--enable-libtensorflow}
9892 Default value is @samp{native}.
9895 Set path to model file specifying network architecture and its parameters.
9896 Note that different backends use different file formats. TensorFlow and native
9897 backend can load files for only its format.
9900 It can also be finished with @ref{dnn_processing} filter.
9904 Attempt to fix small changes in horizontal and/or vertical shift. This
9905 filter helps remove camera shake from hand-holding a camera, bumping a
9906 tripod, moving on a vehicle, etc.
9908 The filter accepts the following options:
9916 Specify a rectangular area where to limit the search for motion
9918 If desired the search for motion vectors can be limited to a
9919 rectangular area of the frame defined by its top left corner, width
9920 and height. These parameters have the same meaning as the drawbox
9921 filter which can be used to visualise the position of the bounding
9924 This is useful when simultaneous movement of subjects within the frame
9925 might be confused for camera motion by the motion vector search.
9927 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9928 then the full frame is used. This allows later options to be set
9929 without specifying the bounding box for the motion vector search.
9931 Default - search the whole frame.
9935 Specify the maximum extent of movement in x and y directions in the
9936 range 0-64 pixels. Default 16.
9939 Specify how to generate pixels to fill blanks at the edge of the
9940 frame. Available values are:
9943 Fill zeroes at blank locations
9945 Original image at blank locations
9947 Extruded edge value at blank locations
9949 Mirrored edge at blank locations
9951 Default value is @samp{mirror}.
9954 Specify the blocksize to use for motion search. Range 4-128 pixels,
9958 Specify the contrast threshold for blocks. Only blocks with more than
9959 the specified contrast (difference between darkest and lightest
9960 pixels) will be considered. Range 1-255, default 125.
9963 Specify the search strategy. Available values are:
9966 Set exhaustive search
9968 Set less exhaustive search.
9970 Default value is @samp{exhaustive}.
9973 If set then a detailed log of the motion search is written to the
9980 Remove unwanted contamination of foreground colors, caused by reflected color of
9981 greenscreen or bluescreen.
9983 This filter accepts the following options:
9987 Set what type of despill to use.
9990 Set how spillmap will be generated.
9993 Set how much to get rid of still remaining spill.
9996 Controls amount of red in spill area.
9999 Controls amount of green in spill area.
10000 Should be -1 for greenscreen.
10003 Controls amount of blue in spill area.
10004 Should be -1 for bluescreen.
10007 Controls brightness of spill area, preserving colors.
10010 Modify alpha from generated spillmap.
10013 @subsection Commands
10015 This filter supports the all above options as @ref{commands}.
10017 @section detelecine
10019 Apply an exact inverse of the telecine operation. It requires a predefined
10020 pattern specified using the pattern option which must be the same as that passed
10021 to the telecine filter.
10023 This filter accepts the following options:
10032 The default value is @code{top}.
10036 A string of numbers representing the pulldown pattern you wish to apply.
10037 The default value is @code{23}.
10040 A number representing position of the first frame with respect to the telecine
10041 pattern. This is to be used if the stream is cut. The default value is @code{0}.
10046 Apply dilation effect to the video.
10048 This filter replaces the pixel by the local(3x3) maximum.
10050 It accepts the following options:
10057 Limit the maximum change for each plane, default is 65535.
10058 If 0, plane will remain unchanged.
10061 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
10064 Flags to local 3x3 coordinates maps like this:
10071 @subsection Commands
10073 This filter supports the all above options as @ref{commands}.
10077 Displace pixels as indicated by second and third input stream.
10079 It takes three input streams and outputs one stream, the first input is the
10080 source, and second and third input are displacement maps.
10082 The second input specifies how much to displace pixels along the
10083 x-axis, while the third input specifies how much to displace pixels
10085 If one of displacement map streams terminates, last frame from that
10086 displacement map will be used.
10088 Note that once generated, displacements maps can be reused over and over again.
10090 A description of the accepted options follows.
10094 Set displace behavior for pixels that are out of range.
10096 Available values are:
10099 Missing pixels are replaced by black pixels.
10102 Adjacent pixels will spread out to replace missing pixels.
10105 Out of range pixels are wrapped so they point to pixels of other side.
10108 Out of range pixels will be replaced with mirrored pixels.
10110 Default is @samp{smear}.
10114 @subsection Examples
10118 Add ripple effect to rgb input of video size hd720:
10120 ffmpeg -i INPUT -f lavfi -i nullsrc=s=hd720,lutrgb=128:128:128 -f lavfi -i nullsrc=s=hd720,geq='r=128+30*sin(2*PI*X/400+T):g=128+30*sin(2*PI*X/400+T):b=128+30*sin(2*PI*X/400+T)' -lavfi '[0][1][2]displace' OUTPUT
10124 Add wave effect to rgb input of video size hd720:
10126 ffmpeg -i INPUT -f lavfi -i nullsrc=hd720,geq='r=128+80*(sin(sqrt((X-W/2)*(X-W/2)+(Y-H/2)*(Y-H/2))/220*2*PI+T)):g=128+80*(sin(sqrt((X-W/2)*(X-W/2)+(Y-H/2)*(Y-H/2))/220*2*PI+T)):b=128+80*(sin(sqrt((X-W/2)*(X-W/2)+(Y-H/2)*(Y-H/2))/220*2*PI+T))' -lavfi '[1]split[x][y],[0][x][y]displace' OUTPUT
10130 @section dnn_detect
10132 Do object detection with deep neural networks.
10134 The filter accepts the following options:
10138 Specify which DNN backend to use for model loading and execution. This option accepts
10139 only openvino now, tensorflow backends will be added.
10142 Set path to model file specifying network architecture and its parameters.
10143 Note that different backends use different file formats.
10146 Set the input name of the dnn network.
10149 Set the output name of the dnn network.
10152 Set the confidence threshold (default: 0.5).
10155 Set path to label file specifying the mapping between label id and name.
10156 Each label name is written in one line, tailing spaces and empty lines are skipped.
10157 The first line is the name of label id 0 (usually it is 'background'),
10158 and the second line is the name of label id 1, etc.
10159 The label id is considered as name if the label file is not provided.
10161 @item backend_configs
10162 Set the configs to be passed into backend
10165 use DNN async execution if set (default: set),
10166 roll back to sync execution if the backend does not support async.
10170 @anchor{dnn_processing}
10171 @section dnn_processing
10173 Do image processing with deep neural networks. It works together with another filter
10174 which converts the pixel format of the Frame to what the dnn network requires.
10176 The filter accepts the following options:
10180 Specify which DNN backend to use for model loading and execution. This option accepts
10181 the following values:
10185 Native implementation of DNN loading and execution.
10188 TensorFlow backend. To enable this backend you
10189 need to install the TensorFlow for C library (see
10190 @url{https://www.tensorflow.org/install/lang_c}) and configure FFmpeg with
10191 @code{--enable-libtensorflow}
10194 OpenVINO backend. To enable this backend you
10195 need to build and install the OpenVINO for C library (see
10196 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
10197 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
10198 be needed if the header files and libraries are not installed into system path)
10202 Default value is @samp{native}.
10205 Set path to model file specifying network architecture and its parameters.
10206 Note that different backends use different file formats. TensorFlow, OpenVINO and native
10207 backend can load files for only its format.
10209 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
10212 Set the input name of the dnn network.
10215 Set the output name of the dnn network.
10218 use DNN async execution if set (default: set),
10219 roll back to sync execution if the backend does not support async.
10223 @subsection Examples
10227 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
10229 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
10233 Halve the pixel value of the frame with format gray32f:
10235 ffmpeg -i input.jpg -vf format=grayf32,dnn_processing=model=halve_gray_float.model:input=dnn_in:output=dnn_out:dnn_backend=native -y out.native.png
10239 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
10241 ./ffmpeg -i 480p.jpg -vf format=yuv420p,scale=w=iw*2:h=ih*2,dnn_processing=dnn_backend=tensorflow:model=srcnn.pb:input=x:output=y -y srcnn.jpg
10245 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
10247 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
10254 Draw a colored box on the input image.
10256 It accepts the following parameters:
10261 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10265 The expressions which specify the width and height of the box; if 0 they are interpreted as
10266 the input width and height. It defaults to 0.
10269 Specify the color of the box to write. For the general syntax of this option,
10270 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10271 value @code{invert} is used, the box edge color is the same as the
10272 video with inverted luma.
10275 The expression which sets the thickness of the box edge.
10276 A value of @code{fill} will create a filled box. Default value is @code{3}.
10278 See below for the list of accepted constants.
10281 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10282 will overwrite the video's color and alpha pixels.
10283 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10286 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10287 following constants:
10291 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10295 horizontal and vertical chroma subsample values. For example for the
10296 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10300 The input width and height.
10303 The input sample aspect ratio.
10307 The x and y offset coordinates where the box is drawn.
10311 The width and height of the drawn box.
10314 The thickness of the drawn box.
10316 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10317 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10321 @subsection Examples
10325 Draw a black box around the edge of the input image:
10331 Draw a box with color red and an opacity of 50%:
10333 drawbox=10:20:200:60:red@@0.5
10336 The previous example can be specified as:
10338 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10342 Fill the box with pink color:
10344 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10348 Draw a 2-pixel red 2.40:1 mask:
10350 drawbox=x=-t:y=0.5*(ih-iw/2.4)-t:w=iw+t*2:h=iw/2.4+t*2:t=2:c=red
10354 @subsection Commands
10355 This filter supports same commands as options.
10356 The command accepts the same syntax of the corresponding option.
10358 If the specified expression is not valid, it is kept at its current
10363 Draw a graph using input video metadata.
10365 It accepts the following parameters:
10369 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10372 Set 1st foreground color expression.
10375 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10378 Set 2nd foreground color expression.
10381 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10384 Set 3rd foreground color expression.
10387 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10390 Set 4th foreground color expression.
10393 Set minimal value of metadata value.
10396 Set maximal value of metadata value.
10399 Set graph background color. Default is white.
10404 Available values for mode is:
10411 Default is @code{line}.
10416 Available values for slide is:
10419 Draw new frame when right border is reached.
10422 Replace old columns with new ones.
10425 Scroll from right to left.
10428 Scroll from left to right.
10431 Draw single picture.
10434 Default is @code{frame}.
10437 Set size of graph video. For the syntax of this option, check the
10438 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10439 The default value is @code{900x256}.
10442 Set the output frame rate. Default value is @code{25}.
10444 The foreground color expressions can use the following variables:
10447 Minimal value of metadata value.
10450 Maximal value of metadata value.
10453 Current metadata key value.
10456 The color is defined as 0xAABBGGRR.
10459 Example using metadata from @ref{signalstats} filter:
10461 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10464 Example using metadata from @ref{ebur128} filter:
10466 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10471 Draw a grid on the input image.
10473 It accepts the following parameters:
10478 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10482 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10483 input width and height, respectively, minus @code{thickness}, so image gets
10484 framed. Default to 0.
10487 Specify the color of the grid. For the general syntax of this option,
10488 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10489 value @code{invert} is used, the grid color is the same as the
10490 video with inverted luma.
10493 The expression which sets the thickness of the grid line. Default value is @code{1}.
10495 See below for the list of accepted constants.
10498 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10499 will overwrite the video's color and alpha pixels.
10500 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10503 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10504 following constants:
10508 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10512 horizontal and vertical chroma subsample values. For example for the
10513 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10517 The input grid cell width and height.
10520 The input sample aspect ratio.
10524 The x and y coordinates of some point of grid intersection (meant to configure offset).
10528 The width and height of the drawn cell.
10531 The thickness of the drawn cell.
10533 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10534 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10538 @subsection Examples
10542 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10544 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10548 Draw a white 3x3 grid with an opacity of 50%:
10550 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10554 @subsection Commands
10555 This filter supports same commands as options.
10556 The command accepts the same syntax of the corresponding option.
10558 If the specified expression is not valid, it is kept at its current
10564 Draw a text string or text from a specified file on top of a video, using the
10565 libfreetype library.
10567 To enable compilation of this filter, you need to configure FFmpeg with
10568 @code{--enable-libfreetype}.
10569 To enable default font fallback and the @var{font} option you need to
10570 configure FFmpeg with @code{--enable-libfontconfig}.
10571 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10572 @code{--enable-libfribidi}.
10576 It accepts the following parameters:
10581 Used to draw a box around text using the background color.
10582 The value must be either 1 (enable) or 0 (disable).
10583 The default value of @var{box} is 0.
10586 Set the width of the border to be drawn around the box using @var{boxcolor}.
10587 The default value of @var{boxborderw} is 0.
10590 The color to be used for drawing box around text. For the syntax of this
10591 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10593 The default value of @var{boxcolor} is "white".
10596 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10597 The default value of @var{line_spacing} is 0.
10600 Set the width of the border to be drawn around the text using @var{bordercolor}.
10601 The default value of @var{borderw} is 0.
10604 Set the color to be used for drawing border around text. For the syntax of this
10605 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10607 The default value of @var{bordercolor} is "black".
10610 Select how the @var{text} is expanded. Can be either @code{none},
10611 @code{strftime} (deprecated) or
10612 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10616 Set a start time for the count. Value is in microseconds. Only applied
10617 in the deprecated strftime expansion mode. To emulate in normal expansion
10618 mode use the @code{pts} function, supplying the start time (in seconds)
10619 as the second argument.
10622 If true, check and fix text coords to avoid clipping.
10625 The color to be used for drawing fonts. For the syntax of this option, check
10626 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10628 The default value of @var{fontcolor} is "black".
10630 @item fontcolor_expr
10631 String which is expanded the same way as @var{text} to obtain dynamic
10632 @var{fontcolor} value. By default this option has empty value and is not
10633 processed. When this option is set, it overrides @var{fontcolor} option.
10636 The font family to be used for drawing text. By default Sans.
10639 The font file to be used for drawing text. The path must be included.
10640 This parameter is mandatory if the fontconfig support is disabled.
10643 Draw the text applying alpha blending. The value can
10644 be a number between 0.0 and 1.0.
10645 The expression accepts the same variables @var{x, y} as well.
10646 The default value is 1.
10647 Please see @var{fontcolor_expr}.
10650 The font size to be used for drawing text.
10651 The default value of @var{fontsize} is 16.
10654 If set to 1, attempt to shape the text (for example, reverse the order of
10655 right-to-left text and join Arabic characters) before drawing it.
10656 Otherwise, just draw the text exactly as given.
10657 By default 1 (if supported).
10659 @item ft_load_flags
10660 The flags to be used for loading the fonts.
10662 The flags map the corresponding flags supported by libfreetype, and are
10663 a combination of the following values:
10670 @item vertical_layout
10671 @item force_autohint
10674 @item ignore_global_advance_width
10676 @item ignore_transform
10678 @item linear_design
10682 Default value is "default".
10684 For more information consult the documentation for the FT_LOAD_*
10688 The color to be used for drawing a shadow behind the drawn text. For the
10689 syntax of this option, check the @ref{color syntax,,"Color" section in the
10690 ffmpeg-utils manual,ffmpeg-utils}.
10692 The default value of @var{shadowcolor} is "black".
10696 The x and y offsets for the text shadow position with respect to the
10697 position of the text. They can be either positive or negative
10698 values. The default value for both is "0".
10701 The starting frame number for the n/frame_num variable. The default value
10705 The size in number of spaces to use for rendering the tab.
10706 Default value is 4.
10709 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10710 format. It can be used with or without text parameter. @var{timecode_rate}
10711 option must be specified.
10713 @item timecode_rate, rate, r
10714 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10715 integer. Minimum value is "1".
10716 Drop-frame timecode is supported for frame rates 30 & 60.
10719 If set to 1, the output of the timecode option will wrap around at 24 hours.
10720 Default is 0 (disabled).
10723 The text string to be drawn. The text must be a sequence of UTF-8
10724 encoded characters.
10725 This parameter is mandatory if no file is specified with the parameter
10729 A text file containing text to be drawn. The text must be a sequence
10730 of UTF-8 encoded characters.
10732 This parameter is mandatory if no text string is specified with the
10733 parameter @var{text}.
10735 If both @var{text} and @var{textfile} are specified, an error is thrown.
10738 If set to 1, the @var{textfile} will be reloaded before each frame.
10739 Be sure to update it atomically, or it may be read partially, or even fail.
10743 The expressions which specify the offsets where text will be drawn
10744 within the video frame. They are relative to the top/left border of the
10747 The default value of @var{x} and @var{y} is "0".
10749 See below for the list of accepted constants and functions.
10752 The parameters for @var{x} and @var{y} are expressions containing the
10753 following constants and functions:
10757 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10761 horizontal and vertical chroma subsample values. For example for the
10762 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10765 the height of each text line
10773 @item max_glyph_a, ascent
10774 the maximum distance from the baseline to the highest/upper grid
10775 coordinate used to place a glyph outline point, for all the rendered
10777 It is a positive value, due to the grid's orientation with the Y axis
10780 @item max_glyph_d, descent
10781 the maximum distance from the baseline to the lowest grid coordinate
10782 used to place a glyph outline point, for all the rendered glyphs.
10783 This is a negative value, due to the grid's orientation, with the Y axis
10787 maximum glyph height, that is the maximum height for all the glyphs
10788 contained in the rendered text, it is equivalent to @var{ascent} -
10792 maximum glyph width, that is the maximum width for all the glyphs
10793 contained in the rendered text
10796 the number of input frame, starting from 0
10798 @item rand(min, max)
10799 return a random number included between @var{min} and @var{max}
10802 The input sample aspect ratio.
10805 timestamp expressed in seconds, NAN if the input timestamp is unknown
10808 the height of the rendered text
10811 the width of the rendered text
10815 the x and y offset coordinates where the text is drawn.
10817 These parameters allow the @var{x} and @var{y} expressions to refer
10818 to each other, so you can for example specify @code{y=x/dar}.
10821 A one character description of the current frame's picture type.
10824 The current packet's position in the input file or stream
10825 (in bytes, from the start of the input). A value of -1 indicates
10826 this info is not available.
10829 The current packet's duration, in seconds.
10832 The current packet's size (in bytes).
10835 @anchor{drawtext_expansion}
10836 @subsection Text expansion
10838 If @option{expansion} is set to @code{strftime},
10839 the filter recognizes strftime() sequences in the provided text and
10840 expands them accordingly. Check the documentation of strftime(). This
10841 feature is deprecated.
10843 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10845 If @option{expansion} is set to @code{normal} (which is the default),
10846 the following expansion mechanism is used.
10848 The backslash character @samp{\}, followed by any character, always expands to
10849 the second character.
10851 Sequences of the form @code{%@{...@}} are expanded. The text between the
10852 braces is a function name, possibly followed by arguments separated by ':'.
10853 If the arguments contain special characters or delimiters (':' or '@}'),
10854 they should be escaped.
10856 Note that they probably must also be escaped as the value for the
10857 @option{text} option in the filter argument string and as the filter
10858 argument in the filtergraph description, and possibly also for the shell,
10859 that makes up to four levels of escaping; using a text file avoids these
10862 The following functions are available:
10867 The expression evaluation result.
10869 It must take one argument specifying the expression to be evaluated,
10870 which accepts the same constants and functions as the @var{x} and
10871 @var{y} values. Note that not all constants should be used, for
10872 example the text size is not known when evaluating the expression, so
10873 the constants @var{text_w} and @var{text_h} will have an undefined
10876 @item expr_int_format, eif
10877 Evaluate the expression's value and output as formatted integer.
10879 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10880 The second argument specifies the output format. Allowed values are @samp{x},
10881 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10882 @code{printf} function.
10883 The third parameter is optional and sets the number of positions taken by the output.
10884 It can be used to add padding with zeros from the left.
10887 The time at which the filter is running, expressed in UTC.
10888 It can accept an argument: a strftime() format string.
10891 The time at which the filter is running, expressed in the local time zone.
10892 It can accept an argument: a strftime() format string.
10895 Frame metadata. Takes one or two arguments.
10897 The first argument is mandatory and specifies the metadata key.
10899 The second argument is optional and specifies a default value, used when the
10900 metadata key is not found or empty.
10902 Available metadata can be identified by inspecting entries
10903 starting with TAG included within each frame section
10904 printed by running @code{ffprobe -show_frames}.
10906 String metadata generated in filters leading to
10907 the drawtext filter are also available.
10910 The frame number, starting from 0.
10913 A one character description of the current picture type.
10916 The timestamp of the current frame.
10917 It can take up to three arguments.
10919 The first argument is the format of the timestamp; it defaults to @code{flt}
10920 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10921 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10922 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10923 @code{localtime} stands for the timestamp of the frame formatted as
10924 local time zone time.
10926 The second argument is an offset added to the timestamp.
10928 If the format is set to @code{hms}, a third argument @code{24HH} may be
10929 supplied to present the hour part of the formatted timestamp in 24h format
10932 If the format is set to @code{localtime} or @code{gmtime},
10933 a third argument may be supplied: a strftime() format string.
10934 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10937 @subsection Commands
10939 This filter supports altering parameters via commands:
10942 Alter existing filter parameters.
10944 Syntax for the argument is the same as for filter invocation, e.g.
10947 fontsize=56:fontcolor=green:text='Hello World'
10950 Full filter invocation with sendcmd would look like this:
10953 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10957 If the entire argument can't be parsed or applied as valid values then the filter will
10958 continue with its existing parameters.
10960 @subsection Examples
10964 Draw "Test Text" with font FreeSerif, using the default values for the
10965 optional parameters.
10968 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10972 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10973 and y=50 (counting from the top-left corner of the screen), text is
10974 yellow with a red box around it. Both the text and the box have an
10978 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10979 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10982 Note that the double quotes are not necessary if spaces are not used
10983 within the parameter list.
10986 Show the text at the center of the video frame:
10988 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10992 Show the text at a random position, switching to a new position every 30 seconds:
10994 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=if(eq(mod(t\,30)\,0)\,rand(0\,(w-text_w))\,x):y=if(eq(mod(t\,30)\,0)\,rand(0\,(h-text_h))\,y)"
10998 Show a text line sliding from right to left in the last row of the video
10999 frame. The file @file{LONG_LINE} is assumed to contain a single line
11002 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
11006 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
11008 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
11012 Draw a single green letter "g", at the center of the input video.
11013 The glyph baseline is placed at half screen height.
11015 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
11019 Show text for 1 second every 3 seconds:
11021 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
11025 Use fontconfig to set the font. Note that the colons need to be escaped.
11027 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
11031 Draw "Test Text" with font size dependent on height of the video.
11033 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
11037 Print the date of a real-time encoding (see strftime(3)):
11039 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
11043 Show text fading in and out (appearing/disappearing):
11046 DS=1.0 # display start
11047 DE=10.0 # display end
11048 FID=1.5 # fade in duration
11049 FOD=5 # fade out duration
11050 ffplay -f lavfi "color,drawtext=text=TEST:fontsize=50:fontfile=FreeSerif.ttf:fontcolor_expr=ff0000%@{eif\\\\: clip(255*(1*between(t\\, $DS + $FID\\, $DE - $FOD) + ((t - $DS)/$FID)*between(t\\, $DS\\, $DS + $FID) + (-(t - $DE)/$FOD)*between(t\\, $DE - $FOD\\, $DE) )\\, 0\\, 255) \\\\: x\\\\: 2 @}"
11054 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
11055 and the @option{fontsize} value are included in the @option{y} offset.
11057 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
11058 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
11062 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
11063 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
11064 must have option @option{-export_path_metadata 1} for the special metadata fields
11065 to be available for filters.
11067 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
11072 For more information about libfreetype, check:
11073 @url{http://www.freetype.org/}.
11075 For more information about fontconfig, check:
11076 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
11078 For more information about libfribidi, check:
11079 @url{http://fribidi.org/}.
11081 @section edgedetect
11083 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
11085 The filter accepts the following options:
11090 Set low and high threshold values used by the Canny thresholding
11093 The high threshold selects the "strong" edge pixels, which are then
11094 connected through 8-connectivity with the "weak" edge pixels selected
11095 by the low threshold.
11097 @var{low} and @var{high} threshold values must be chosen in the range
11098 [0,1], and @var{low} should be lesser or equal to @var{high}.
11100 Default value for @var{low} is @code{20/255}, and default value for @var{high}
11104 Define the drawing mode.
11108 Draw white/gray wires on black background.
11111 Mix the colors to create a paint/cartoon effect.
11114 Apply Canny edge detector on all selected planes.
11116 Default value is @var{wires}.
11119 Select planes for filtering. By default all available planes are filtered.
11122 @subsection Examples
11126 Standard edge detection with custom values for the hysteresis thresholding:
11128 edgedetect=low=0.1:high=0.4
11132 Painting effect without thresholding:
11134 edgedetect=mode=colormix:high=0
11140 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
11142 For each input image, the filter will compute the optimal mapping from
11143 the input to the output given the codebook length, that is the number
11144 of distinct output colors.
11146 This filter accepts the following options.
11149 @item codebook_length, l
11150 Set codebook length. The value must be a positive integer, and
11151 represents the number of distinct output colors. Default value is 256.
11154 Set the maximum number of iterations to apply for computing the optimal
11155 mapping. The higher the value the better the result and the higher the
11156 computation time. Default value is 1.
11159 Set a random seed, must be an integer included between 0 and
11160 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
11161 will try to use a good random seed on a best effort basis.
11164 Set pal8 output pixel format. This option does not work with codebook
11165 length greater than 256. Default is disabled.
11170 Measure graylevel entropy in histogram of color channels of video frames.
11172 It accepts the following parameters:
11176 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
11178 @var{diff} mode measures entropy of histogram delta values, absolute differences
11179 between neighbour histogram values.
11183 Apply the EPX magnification filter which is designed for pixel art.
11185 It accepts the following option:
11189 Set the scaling dimension: @code{2} for @code{2xEPX}, @code{3} for
11191 Default is @code{3}.
11195 Set brightness, contrast, saturation and approximate gamma adjustment.
11197 The filter accepts the following options:
11201 Set the contrast expression. The value must be a float value in range
11202 @code{-1000.0} to @code{1000.0}. The default value is "1".
11205 Set the brightness expression. The value must be a float value in
11206 range @code{-1.0} to @code{1.0}. The default value is "0".
11209 Set the saturation expression. The value must be a float in
11210 range @code{0.0} to @code{3.0}. The default value is "1".
11213 Set the gamma expression. The value must be a float in range
11214 @code{0.1} to @code{10.0}. The default value is "1".
11217 Set the gamma expression for red. The value must be a float in
11218 range @code{0.1} to @code{10.0}. The default value is "1".
11221 Set the gamma expression for green. The value must be a float in range
11222 @code{0.1} to @code{10.0}. The default value is "1".
11225 Set the gamma expression for blue. The value must be a float in range
11226 @code{0.1} to @code{10.0}. The default value is "1".
11229 Set the gamma weight expression. It can be used to reduce the effect
11230 of a high gamma value on bright image areas, e.g. keep them from
11231 getting overamplified and just plain white. The value must be a float
11232 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
11233 gamma correction all the way down while @code{1.0} leaves it at its
11234 full strength. Default is "1".
11237 Set when the expressions for brightness, contrast, saturation and
11238 gamma expressions are evaluated.
11240 It accepts the following values:
11243 only evaluate expressions once during the filter initialization or
11244 when a command is processed
11247 evaluate expressions for each incoming frame
11250 Default value is @samp{init}.
11253 The expressions accept the following parameters:
11256 frame count of the input frame starting from 0
11259 byte position of the corresponding packet in the input file, NAN if
11263 frame rate of the input video, NAN if the input frame rate is unknown
11266 timestamp expressed in seconds, NAN if the input timestamp is unknown
11269 @subsection Commands
11270 The filter supports the following commands:
11274 Set the contrast expression.
11277 Set the brightness expression.
11280 Set the saturation expression.
11283 Set the gamma expression.
11286 Set the gamma_r expression.
11289 Set gamma_g expression.
11292 Set gamma_b expression.
11295 Set gamma_weight expression.
11297 The command accepts the same syntax of the corresponding option.
11299 If the specified expression is not valid, it is kept at its current
11306 Apply erosion effect to the video.
11308 This filter replaces the pixel by the local(3x3) minimum.
11310 It accepts the following options:
11317 Limit the maximum change for each plane, default is 65535.
11318 If 0, plane will remain unchanged.
11321 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11324 Flags to local 3x3 coordinates maps like this:
11331 @subsection Commands
11333 This filter supports the all above options as @ref{commands}.
11337 Deinterlace the input video ("estdif" stands for "Edge Slope
11338 Tracing Deinterlacing Filter").
11340 Spatial only filter that uses edge slope tracing algorithm
11341 to interpolate missing lines.
11342 It accepts the following parameters:
11346 The interlacing mode to adopt. It accepts one of the following values:
11350 Output one frame for each frame.
11352 Output one frame for each field.
11355 The default value is @code{field}.
11358 The picture field parity assumed for the input interlaced video. It accepts one
11359 of the following values:
11363 Assume the top field is first.
11365 Assume the bottom field is first.
11367 Enable automatic detection of field parity.
11370 The default value is @code{auto}.
11371 If the interlacing is unknown or the decoder does not export this information,
11372 top field first will be assumed.
11375 Specify which frames to deinterlace. Accepts one of the following
11380 Deinterlace all frames.
11382 Only deinterlace frames marked as interlaced.
11385 The default value is @code{all}.
11388 Specify the search radius for edge slope tracing. Default value is 1.
11389 Allowed range is from 1 to 15.
11392 Specify the search radius for best edge matching. Default value is 2.
11393 Allowed range is from 0 to 15.
11396 Specify the interpolation used. Default is 4-point interpolation. It accepts one
11397 of the following values:
11401 Two-point interpolation.
11403 Four-point interpolation.
11405 Six-point interpolation.
11409 @subsection Commands
11410 This filter supports same @ref{commands} as options.
11413 Adjust exposure of the video stream.
11415 The filter accepts the following options:
11419 Set the exposure correction in EV. Allowed range is from -3.0 to 3.0 EV
11420 Default value is 0 EV.
11423 Set the black level correction. Allowed range is from -1.0 to 1.0.
11424 Default value is 0.
11427 @subsection Commands
11429 This filter supports same @ref{commands} as options.
11431 @section extractplanes
11433 Extract color channel components from input video stream into
11434 separate grayscale video streams.
11436 The filter accepts the following option:
11440 Set plane(s) to extract.
11442 Available values for planes are:
11453 Choosing planes not available in the input will result in an error.
11454 That means you cannot select @code{r}, @code{g}, @code{b} planes
11455 with @code{y}, @code{u}, @code{v} planes at same time.
11458 @subsection Examples
11462 Extract luma, u and v color channel component from input video frame
11463 into 3 grayscale outputs:
11465 ffmpeg -i video.avi -filter_complex 'extractplanes=y+u+v[y][u][v]' -map '[y]' y.avi -map '[u]' u.avi -map '[v]' v.avi
11471 Apply a fade-in/out effect to the input video.
11473 It accepts the following parameters:
11477 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11479 Default is @code{in}.
11481 @item start_frame, s
11482 Specify the number of the frame to start applying the fade
11483 effect at. Default is 0.
11486 The number of frames that the fade effect lasts. At the end of the
11487 fade-in effect, the output video will have the same intensity as the input video.
11488 At the end of the fade-out transition, the output video will be filled with the
11489 selected @option{color}.
11493 If set to 1, fade only alpha channel, if one exists on the input.
11494 Default value is 0.
11496 @item start_time, st
11497 Specify the timestamp (in seconds) of the frame to start to apply the fade
11498 effect. If both start_frame and start_time are specified, the fade will start at
11499 whichever comes last. Default is 0.
11502 The number of seconds for which the fade effect has to last. At the end of the
11503 fade-in effect the output video will have the same intensity as the input video,
11504 at the end of the fade-out transition the output video will be filled with the
11505 selected @option{color}.
11506 If both duration and nb_frames are specified, duration is used. Default is 0
11507 (nb_frames is used by default).
11510 Specify the color of the fade. Default is "black".
11513 @subsection Examples
11517 Fade in the first 30 frames of video:
11522 The command above is equivalent to:
11528 Fade out the last 45 frames of a 200-frame video:
11531 fade=type=out:start_frame=155:nb_frames=45
11535 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11537 fade=in:0:25, fade=out:975:25
11541 Make the first 5 frames yellow, then fade in from frame 5-24:
11543 fade=in:5:20:color=yellow
11547 Fade in alpha over first 25 frames of video:
11549 fade=in:0:25:alpha=1
11553 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11555 fade=t=in:st=5.5:d=0.5
11561 Denoise frames using 3D FFT (frequency domain filtering).
11563 The filter accepts the following options:
11567 Set the noise sigma constant. This sets denoising strength.
11568 Default value is 1. Allowed range is from 0 to 30.
11569 Using very high sigma with low overlap may give blocking artifacts.
11572 Set amount of denoising. By default all detected noise is reduced.
11573 Default value is 1. Allowed range is from 0 to 1.
11576 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11577 Actual size of block in pixels is 2 to power of @var{block}, so by default
11578 block size in pixels is 2^4 which is 16.
11581 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11584 Set number of previous frames to use for denoising. By default is set to 0.
11587 Set number of next frames to to use for denoising. By default is set to 0.
11590 Set planes which will be filtered, by default are all available filtered
11595 Apply arbitrary expressions to samples in frequency domain
11599 Adjust the dc value (gain) of the luma plane of the image. The filter
11600 accepts an integer value in range @code{0} to @code{1000}. The default
11601 value is set to @code{0}.
11604 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11605 filter accepts an integer value in range @code{0} to @code{1000}. The
11606 default value is set to @code{0}.
11609 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11610 filter accepts an integer value in range @code{0} to @code{1000}. The
11611 default value is set to @code{0}.
11614 Set the frequency domain weight expression for the luma plane.
11617 Set the frequency domain weight expression for the 1st chroma plane.
11620 Set the frequency domain weight expression for the 2nd chroma plane.
11623 Set when the expressions are evaluated.
11625 It accepts the following values:
11628 Only evaluate expressions once during the filter initialization.
11631 Evaluate expressions for each incoming frame.
11634 Default value is @samp{init}.
11636 The filter accepts the following variables:
11639 The coordinates of the current sample.
11643 The width and height of the image.
11646 The number of input frame, starting from 0.
11649 @subsection Examples
11655 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11661 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11667 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11673 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11680 Extract a single field from an interlaced image using stride
11681 arithmetic to avoid wasting CPU time. The output frames are marked as
11684 The filter accepts the following options:
11688 Specify whether to extract the top (if the value is @code{0} or
11689 @code{top}) or the bottom field (if the value is @code{1} or
11695 Create new frames by copying the top and bottom fields from surrounding frames
11696 supplied as numbers by the hint file.
11700 Set file containing hints: absolute/relative frame numbers.
11702 There must be one line for each frame in a clip. Each line must contain two
11703 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11704 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11705 is current frame number for @code{absolute} mode or out of [-1, 1] range
11706 for @code{relative} mode. First number tells from which frame to pick up top
11707 field and second number tells from which frame to pick up bottom field.
11709 If optionally followed by @code{+} output frame will be marked as interlaced,
11710 else if followed by @code{-} output frame will be marked as progressive, else
11711 it will be marked same as input frame.
11712 If optionally followed by @code{t} output frame will use only top field, or in
11713 case of @code{b} it will use only bottom field.
11714 If line starts with @code{#} or @code{;} that line is skipped.
11717 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11720 Example of first several lines of @code{hint} file for @code{relative} mode:
11722 0,0 - # first frame
11723 1,0 - # second frame, use third's frame top field and second's frame bottom field
11724 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11739 @section fieldmatch
11741 Field matching filter for inverse telecine. It is meant to reconstruct the
11742 progressive frames from a telecined stream. The filter does not drop duplicated
11743 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11744 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11746 The separation of the field matching and the decimation is notably motivated by
11747 the possibility of inserting a de-interlacing filter fallback between the two.
11748 If the source has mixed telecined and real interlaced content,
11749 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11750 But these remaining combed frames will be marked as interlaced, and thus can be
11751 de-interlaced by a later filter such as @ref{yadif} before decimation.
11753 In addition to the various configuration options, @code{fieldmatch} can take an
11754 optional second stream, activated through the @option{ppsrc} option. If
11755 enabled, the frames reconstruction will be based on the fields and frames from
11756 this second stream. This allows the first input to be pre-processed in order to
11757 help the various algorithms of the filter, while keeping the output lossless
11758 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11759 or brightness/contrast adjustments can help.
11761 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11762 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11763 which @code{fieldmatch} is based on. While the semantic and usage are very
11764 close, some behaviour and options names can differ.
11766 The @ref{decimate} filter currently only works for constant frame rate input.
11767 If your input has mixed telecined (30fps) and progressive content with a lower
11768 framerate like 24fps use the following filterchain to produce the necessary cfr
11769 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11771 The filter accepts the following options:
11775 Specify the assumed field order of the input stream. Available values are:
11779 Auto detect parity (use FFmpeg's internal parity value).
11781 Assume bottom field first.
11783 Assume top field first.
11786 Note that it is sometimes recommended not to trust the parity announced by the
11789 Default value is @var{auto}.
11792 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11793 sense that it won't risk creating jerkiness due to duplicate frames when
11794 possible, but if there are bad edits or blended fields it will end up
11795 outputting combed frames when a good match might actually exist. On the other
11796 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11797 but will almost always find a good frame if there is one. The other values are
11798 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11799 jerkiness and creating duplicate frames versus finding good matches in sections
11800 with bad edits, orphaned fields, blended fields, etc.
11802 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11804 Available values are:
11808 2-way matching (p/c)
11810 2-way matching, and trying 3rd match if still combed (p/c + n)
11812 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11814 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11815 still combed (p/c + n + u/b)
11817 3-way matching (p/c/n)
11819 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11820 detected as combed (p/c/n + u/b)
11823 The parenthesis at the end indicate the matches that would be used for that
11824 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11827 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11830 Default value is @var{pc_n}.
11833 Mark the main input stream as a pre-processed input, and enable the secondary
11834 input stream as the clean source to pick the fields from. See the filter
11835 introduction for more details. It is similar to the @option{clip2} feature from
11838 Default value is @code{0} (disabled).
11841 Set the field to match from. It is recommended to set this to the same value as
11842 @option{order} unless you experience matching failures with that setting. In
11843 certain circumstances changing the field that is used to match from can have a
11844 large impact on matching performance. Available values are:
11848 Automatic (same value as @option{order}).
11850 Match from the bottom field.
11852 Match from the top field.
11855 Default value is @var{auto}.
11858 Set whether or not chroma is included during the match comparisons. In most
11859 cases it is recommended to leave this enabled. You should set this to @code{0}
11860 only if your clip has bad chroma problems such as heavy rainbowing or other
11861 artifacts. Setting this to @code{0} could also be used to speed things up at
11862 the cost of some accuracy.
11864 Default value is @code{1}.
11868 These define an exclusion band which excludes the lines between @option{y0} and
11869 @option{y1} from being included in the field matching decision. An exclusion
11870 band can be used to ignore subtitles, a logo, or other things that may
11871 interfere with the matching. @option{y0} sets the starting scan line and
11872 @option{y1} sets the ending line; all lines in between @option{y0} and
11873 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11874 @option{y0} and @option{y1} to the same value will disable the feature.
11875 @option{y0} and @option{y1} defaults to @code{0}.
11878 Set the scene change detection threshold as a percentage of maximum change on
11879 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11880 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11881 @option{scthresh} is @code{[0.0, 100.0]}.
11883 Default value is @code{12.0}.
11886 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11887 account the combed scores of matches when deciding what match to use as the
11888 final match. Available values are:
11892 No final matching based on combed scores.
11894 Combed scores are only used when a scene change is detected.
11896 Use combed scores all the time.
11899 Default is @var{sc}.
11902 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11903 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11904 Available values are:
11908 No forced calculation.
11910 Force p/c/n calculations.
11912 Force p/c/n/u/b calculations.
11915 Default value is @var{none}.
11918 This is the area combing threshold used for combed frame detection. This
11919 essentially controls how "strong" or "visible" combing must be to be detected.
11920 Larger values mean combing must be more visible and smaller values mean combing
11921 can be less visible or strong and still be detected. Valid settings are from
11922 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11923 be detected as combed). This is basically a pixel difference value. A good
11924 range is @code{[8, 12]}.
11926 Default value is @code{9}.
11929 Sets whether or not chroma is considered in the combed frame decision. Only
11930 disable this if your source has chroma problems (rainbowing, etc.) that are
11931 causing problems for the combed frame detection with chroma enabled. Actually,
11932 using @option{chroma}=@var{0} is usually more reliable, except for the case
11933 where there is chroma only combing in the source.
11935 Default value is @code{0}.
11939 Respectively set the x-axis and y-axis size of the window used during combed
11940 frame detection. This has to do with the size of the area in which
11941 @option{combpel} pixels are required to be detected as combed for a frame to be
11942 declared combed. See the @option{combpel} parameter description for more info.
11943 Possible values are any number that is a power of 2 starting at 4 and going up
11946 Default value is @code{16}.
11949 The number of combed pixels inside any of the @option{blocky} by
11950 @option{blockx} size blocks on the frame for the frame to be detected as
11951 combed. While @option{cthresh} controls how "visible" the combing must be, this
11952 setting controls "how much" combing there must be in any localized area (a
11953 window defined by the @option{blockx} and @option{blocky} settings) on the
11954 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11955 which point no frames will ever be detected as combed). This setting is known
11956 as @option{MI} in TFM/VFM vocabulary.
11958 Default value is @code{80}.
11961 @anchor{p/c/n/u/b meaning}
11962 @subsection p/c/n/u/b meaning
11964 @subsubsection p/c/n
11966 We assume the following telecined stream:
11969 Top fields: 1 2 2 3 4
11970 Bottom fields: 1 2 3 4 4
11973 The numbers correspond to the progressive frame the fields relate to. Here, the
11974 first two frames are progressive, the 3rd and 4th are combed, and so on.
11976 When @code{fieldmatch} is configured to run a matching from bottom
11977 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11982 B 1 2 3 4 4 <-- matching reference
11991 As a result of the field matching, we can see that some frames get duplicated.
11992 To perform a complete inverse telecine, you need to rely on a decimation filter
11993 after this operation. See for instance the @ref{decimate} filter.
11995 The same operation now matching from top fields (@option{field}=@var{top})
12000 T 1 2 2 3 4 <-- matching reference
12010 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
12011 basically, they refer to the frame and field of the opposite parity:
12014 @item @var{p} matches the field of the opposite parity in the previous frame
12015 @item @var{c} matches the field of the opposite parity in the current frame
12016 @item @var{n} matches the field of the opposite parity in the next frame
12021 The @var{u} and @var{b} matching are a bit special in the sense that they match
12022 from the opposite parity flag. In the following examples, we assume that we are
12023 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
12024 'x' is placed above and below each matched fields.
12026 With bottom matching (@option{field}=@var{bottom}):
12031 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
12032 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
12040 With top matching (@option{field}=@var{top}):
12045 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
12046 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
12054 @subsection Examples
12056 Simple IVTC of a top field first telecined stream:
12058 fieldmatch=order=tff:combmatch=none, decimate
12061 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
12063 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
12066 @section fieldorder
12068 Transform the field order of the input video.
12070 It accepts the following parameters:
12075 The output field order. Valid values are @var{tff} for top field first or @var{bff}
12076 for bottom field first.
12079 The default value is @samp{tff}.
12081 The transformation is done by shifting the picture content up or down
12082 by one line, and filling the remaining line with appropriate picture content.
12083 This method is consistent with most broadcast field order converters.
12085 If the input video is not flagged as being interlaced, or it is already
12086 flagged as being of the required output field order, then this filter does
12087 not alter the incoming video.
12089 It is very useful when converting to or from PAL DV material,
12090 which is bottom field first.
12094 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
12097 @section fifo, afifo
12099 Buffer input images and send them when they are requested.
12101 It is mainly useful when auto-inserted by the libavfilter
12104 It does not take parameters.
12106 @section fillborders
12108 Fill borders of the input video, without changing video stream dimensions.
12109 Sometimes video can have garbage at the four edges and you may not want to
12110 crop video input to keep size multiple of some number.
12112 This filter accepts the following options:
12116 Number of pixels to fill from left border.
12119 Number of pixels to fill from right border.
12122 Number of pixels to fill from top border.
12125 Number of pixels to fill from bottom border.
12130 It accepts the following values:
12133 fill pixels using outermost pixels
12136 fill pixels using mirroring (half sample symmetric)
12139 fill pixels with constant value
12142 fill pixels using reflecting (whole sample symmetric)
12145 fill pixels using wrapping
12148 fade pixels to constant value
12151 Default is @var{smear}.
12154 Set color for pixels in fixed or fade mode. Default is @var{black}.
12157 @subsection Commands
12158 This filter supports same @ref{commands} as options.
12159 The command accepts the same syntax of the corresponding option.
12161 If the specified expression is not valid, it is kept at its current
12166 Find a rectangular object
12168 It accepts the following options:
12172 Filepath of the object image, needs to be in gray8.
12175 Detection threshold, default is 0.5.
12178 Number of mipmaps, default is 3.
12180 @item xmin, ymin, xmax, ymax
12181 Specifies the rectangle in which to search.
12184 Discard frames where object is not detected. Default is disabled.
12187 @subsection Examples
12191 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
12193 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
12199 Flood area with values of same pixel components with another values.
12201 It accepts the following options:
12204 Set pixel x coordinate.
12207 Set pixel y coordinate.
12210 Set source #0 component value.
12213 Set source #1 component value.
12216 Set source #2 component value.
12219 Set source #3 component value.
12222 Set destination #0 component value.
12225 Set destination #1 component value.
12228 Set destination #2 component value.
12231 Set destination #3 component value.
12237 Convert the input video to one of the specified pixel formats.
12238 Libavfilter will try to pick one that is suitable as input to
12241 It accepts the following parameters:
12245 A '|'-separated list of pixel format names, such as
12246 "pix_fmts=yuv420p|monow|rgb24".
12250 @subsection Examples
12254 Convert the input video to the @var{yuv420p} format
12256 format=pix_fmts=yuv420p
12259 Convert the input video to any of the formats in the list
12261 format=pix_fmts=yuv420p|yuv444p|yuv410p
12268 Convert the video to specified constant frame rate by duplicating or dropping
12269 frames as necessary.
12271 It accepts the following parameters:
12275 The desired output frame rate. The default is @code{25}.
12278 Assume the first PTS should be the given value, in seconds. This allows for
12279 padding/trimming at the start of stream. By default, no assumption is made
12280 about the first frame's expected PTS, so no padding or trimming is done.
12281 For example, this could be set to 0 to pad the beginning with duplicates of
12282 the first frame if a video stream starts after the audio stream or to trim any
12283 frames with a negative PTS.
12286 Timestamp (PTS) rounding method.
12288 Possible values are:
12295 round towards -infinity
12297 round towards +infinity
12301 The default is @code{near}.
12304 Action performed when reading the last frame.
12306 Possible values are:
12309 Use same timestamp rounding method as used for other frames.
12311 Pass through last frame if input duration has not been reached yet.
12313 The default is @code{round}.
12317 Alternatively, the options can be specified as a flat string:
12318 @var{fps}[:@var{start_time}[:@var{round}]].
12320 See also the @ref{setpts} filter.
12322 @subsection Examples
12326 A typical usage in order to set the fps to 25:
12332 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
12334 fps=fps=film:round=near
12340 Pack two different video streams into a stereoscopic video, setting proper
12341 metadata on supported codecs. The two views should have the same size and
12342 framerate and processing will stop when the shorter video ends. Please note
12343 that you may conveniently adjust view properties with the @ref{scale} and
12346 It accepts the following parameters:
12350 The desired packing format. Supported values are:
12355 The views are next to each other (default).
12358 The views are on top of each other.
12361 The views are packed by line.
12364 The views are packed by column.
12367 The views are temporally interleaved.
12376 # Convert left and right views into a frame-sequential video
12377 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12379 # Convert views into a side-by-side video with the same output resolution as the input
12380 ffmpeg -i LEFT -i RIGHT -filter_complex [0:v]scale=w=iw/2[left],[1:v]scale=w=iw/2[right],[left][right]framepack=sbs OUTPUT
12385 Change the frame rate by interpolating new video output frames from the source
12388 This filter is not designed to function correctly with interlaced media. If
12389 you wish to change the frame rate of interlaced media then you are required
12390 to deinterlace before this filter and re-interlace after this filter.
12392 A description of the accepted options follows.
12396 Specify the output frames per second. This option can also be specified
12397 as a value alone. The default is @code{50}.
12400 Specify the start of a range where the output frame will be created as a
12401 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12402 the default is @code{15}.
12405 Specify the end of a range where the output frame will be created as a
12406 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12407 the default is @code{240}.
12410 Specify the level at which a scene change is detected as a value between
12411 0 and 100 to indicate a new scene; a low value reflects a low
12412 probability for the current frame to introduce a new scene, while a higher
12413 value means the current frame is more likely to be one.
12414 The default is @code{8.2}.
12417 Specify flags influencing the filter process.
12419 Available value for @var{flags} is:
12422 @item scene_change_detect, scd
12423 Enable scene change detection using the value of the option @var{scene}.
12424 This flag is enabled by default.
12430 Select one frame every N-th frame.
12432 This filter accepts the following option:
12435 Select frame after every @code{step} frames.
12436 Allowed values are positive integers higher than 0. Default value is @code{1}.
12439 @section freezedetect
12441 Detect frozen video.
12443 This filter logs a message and sets frame metadata when it detects that the
12444 input video has no significant change in content during a specified duration.
12445 Video freeze detection calculates the mean average absolute difference of all
12446 the components of video frames and compares it to a noise floor.
12448 The printed times and duration are expressed in seconds. The
12449 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12450 whose timestamp equals or exceeds the detection duration and it contains the
12451 timestamp of the first frame of the freeze. The
12452 @code{lavfi.freezedetect.freeze_duration} and
12453 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12456 The filter accepts the following options:
12460 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12461 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12465 Set freeze duration until notification (default is 2 seconds).
12468 @section freezeframes
12470 Freeze video frames.
12472 This filter freezes video frames using frame from 2nd input.
12474 The filter accepts the following options:
12478 Set number of first frame from which to start freeze.
12481 Set number of last frame from which to end freeze.
12484 Set number of frame from 2nd input which will be used instead of replaced frames.
12490 Apply a frei0r effect to the input video.
12492 To enable the compilation of this filter, you need to install the frei0r
12493 header and configure FFmpeg with @code{--enable-frei0r}.
12495 It accepts the following parameters:
12500 The name of the frei0r effect to load. If the environment variable
12501 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12502 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12503 Otherwise, the standard frei0r paths are searched, in this order:
12504 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12505 @file{/usr/lib/frei0r-1/}.
12507 @item filter_params
12508 A '|'-separated list of parameters to pass to the frei0r effect.
12512 A frei0r effect parameter can be a boolean (its value is either
12513 "y" or "n"), a double, a color (specified as
12514 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12515 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12516 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12517 a position (specified as @var{X}/@var{Y}, where
12518 @var{X} and @var{Y} are floating point numbers) and/or a string.
12520 The number and types of parameters depend on the loaded effect. If an
12521 effect parameter is not specified, the default value is set.
12523 @subsection Examples
12527 Apply the distort0r effect, setting the first two double parameters:
12529 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12533 Apply the colordistance effect, taking a color as the first parameter:
12535 frei0r=colordistance:0.2/0.3/0.4
12536 frei0r=colordistance:violet
12537 frei0r=colordistance:0x112233
12541 Apply the perspective effect, specifying the top left and top right image
12544 frei0r=perspective:0.2/0.2|0.8/0.2
12548 For more information, see
12549 @url{http://frei0r.dyne.org}
12551 @subsection Commands
12553 This filter supports the @option{filter_params} option as @ref{commands}.
12557 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12559 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12560 processing filter, one of them is performed once per block, not per pixel.
12561 This allows for much higher speed.
12563 The filter accepts the following options:
12567 Set quality. This option defines the number of levels for averaging. It accepts
12568 an integer in the range 4-5. Default value is @code{4}.
12571 Force a constant quantization parameter. It accepts an integer in range 0-63.
12572 If not set, the filter will use the QP from the video stream (if available).
12575 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12576 more details but also more artifacts, while higher values make the image smoother
12577 but also blurrier. Default value is @code{0} − PSNR optimal.
12579 @item use_bframe_qp
12580 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12581 option may cause flicker since the B-Frames have often larger QP. Default is
12582 @code{0} (not enabled).
12588 Apply Gaussian blur filter.
12590 The filter accepts the following options:
12594 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12597 Set number of steps for Gaussian approximation. Default is @code{1}.
12600 Set which planes to filter. By default all planes are filtered.
12603 Set vertical sigma, if negative it will be same as @code{sigma}.
12604 Default is @code{-1}.
12607 @subsection Commands
12608 This filter supports same commands as options.
12609 The command accepts the same syntax of the corresponding option.
12611 If the specified expression is not valid, it is kept at its current
12616 Apply generic equation to each pixel.
12618 The filter accepts the following options:
12621 @item lum_expr, lum
12622 Set the luminance expression.
12624 Set the chrominance blue expression.
12626 Set the chrominance red expression.
12627 @item alpha_expr, a
12628 Set the alpha expression.
12630 Set the red expression.
12631 @item green_expr, g
12632 Set the green expression.
12634 Set the blue expression.
12637 The colorspace is selected according to the specified options. If one
12638 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12639 options is specified, the filter will automatically select a YCbCr
12640 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12641 @option{blue_expr} options is specified, it will select an RGB
12644 If one of the chrominance expression is not defined, it falls back on the other
12645 one. If no alpha expression is specified it will evaluate to opaque value.
12646 If none of chrominance expressions are specified, they will evaluate
12647 to the luminance expression.
12649 The expressions can use the following variables and functions:
12653 The sequential number of the filtered frame, starting from @code{0}.
12657 The coordinates of the current sample.
12661 The width and height of the image.
12665 Width and height scale depending on the currently filtered plane. It is the
12666 ratio between the corresponding luma plane number of pixels and the current
12667 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12668 @code{0.5,0.5} for chroma planes.
12671 Time of the current frame, expressed in seconds.
12674 Return the value of the pixel at location (@var{x},@var{y}) of the current
12678 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12682 Return the value of the pixel at location (@var{x},@var{y}) of the
12683 blue-difference chroma plane. Return 0 if there is no such plane.
12686 Return the value of the pixel at location (@var{x},@var{y}) of the
12687 red-difference chroma plane. Return 0 if there is no such plane.
12692 Return the value of the pixel at location (@var{x},@var{y}) of the
12693 red/green/blue component. Return 0 if there is no such component.
12696 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12697 plane. Return 0 if there is no such plane.
12699 @item psum(x,y), lumsum(x, y), cbsum(x,y), crsum(x,y), rsum(x,y), gsum(x,y), bsum(x,y), alphasum(x,y)
12700 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12701 sums of samples within a rectangle. See the functions without the sum postfix.
12703 @item interpolation
12704 Set one of interpolation methods:
12709 Default is bilinear.
12712 For functions, if @var{x} and @var{y} are outside the area, the value will be
12713 automatically clipped to the closer edge.
12715 Please note that this filter can use multiple threads in which case each slice
12716 will have its own expression state. If you want to use only a single expression
12717 state because your expressions depend on previous state then you should limit
12718 the number of filter threads to 1.
12720 @subsection Examples
12724 Flip the image horizontally:
12730 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12731 wavelength of 100 pixels:
12733 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12737 Generate a fancy enigmatic moving light:
12739 nullsrc=s=256x256,geq=random(1)/hypot(X-cos(N*0.07)*W/2-W/2\,Y-sin(N*0.09)*H/2-H/2)^2*1000000*sin(N*0.02):128:128
12743 Generate a quick emboss effect:
12745 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12749 Modify RGB components depending on pixel position:
12751 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12755 Create a radial gradient that is the same size as the input (also see
12756 the @ref{vignette} filter):
12758 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12764 Fix the banding artifacts that are sometimes introduced into nearly flat
12765 regions by truncation to 8-bit color depth.
12766 Interpolate the gradients that should go where the bands are, and
12769 It is designed for playback only. Do not use it prior to
12770 lossy compression, because compression tends to lose the dither and
12771 bring back the bands.
12773 It accepts the following parameters:
12778 The maximum amount by which the filter will change any one pixel. This is also
12779 the threshold for detecting nearly flat regions. Acceptable values range from
12780 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12784 The neighborhood to fit the gradient to. A larger radius makes for smoother
12785 gradients, but also prevents the filter from modifying the pixels near detailed
12786 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12787 values will be clipped to the valid range.
12791 Alternatively, the options can be specified as a flat string:
12792 @var{strength}[:@var{radius}]
12794 @subsection Examples
12798 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12804 Specify radius, omitting the strength (which will fall-back to the default
12812 @anchor{graphmonitor}
12813 @section graphmonitor
12814 Show various filtergraph stats.
12816 With this filter one can debug complete filtergraph.
12817 Especially issues with links filling with queued frames.
12819 The filter accepts the following options:
12823 Set video output size. Default is @var{hd720}.
12826 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12829 Set output mode, can be @var{fulll} or @var{compact}.
12830 In @var{compact} mode only filters with some queued frames have displayed stats.
12833 Set flags which enable which stats are shown in video.
12835 Available values for flags are:
12838 Display number of queued frames in each link.
12840 @item frame_count_in
12841 Display number of frames taken from filter.
12843 @item frame_count_out
12844 Display number of frames given out from filter.
12847 Display current filtered frame pts.
12850 Display current filtered frame time.
12853 Display time base for filter link.
12856 Display used format for filter link.
12859 Display video size or number of audio channels in case of audio used by filter link.
12862 Display video frame rate or sample rate in case of audio used by filter link.
12865 Display link output status.
12869 Set upper limit for video rate of output stream, Default value is @var{25}.
12870 This guarantee that output video frame rate will not be higher than this value.
12874 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12875 and corrects the scene colors accordingly.
12877 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12879 The filter accepts the following options:
12883 The order of differentiation to be applied on the scene. Must be chosen in the range
12884 [0,2] and default value is 1.
12887 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12888 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12889 max value instead of calculating Minkowski distance.
12892 The standard deviation of Gaussian blur to be applied on the scene. Must be
12893 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12894 can't be equal to 0 if @var{difford} is greater than 0.
12897 @subsection Examples
12903 greyedge=difford=1:minknorm=5:sigma=2
12909 greyedge=difford=1:minknorm=0:sigma=2
12917 Apply a Hald CLUT to a video stream.
12919 First input is the video stream to process, and second one is the Hald CLUT.
12920 The Hald CLUT input can be a simple picture or a complete video stream.
12922 The filter accepts the following options:
12926 Force termination when the shortest input terminates. Default is @code{0}.
12928 Continue applying the last CLUT after the end of the stream. A value of
12929 @code{0} disable the filter after the last frame of the CLUT is reached.
12930 Default is @code{1}.
12933 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12934 filters share the same internals).
12936 This filter also supports the @ref{framesync} options.
12938 More information about the Hald CLUT can be found on Eskil Steenberg's website
12939 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12941 @subsection Commands
12943 This filter supports the @code{interp} option as @ref{commands}.
12945 @subsection Workflow examples
12947 @subsubsection Hald CLUT video stream
12949 Generate an identity Hald CLUT stream altered with various effects:
12951 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "hue=H=2*PI*t:s=sin(2*PI*t)+1, curves=cross_process" -t 10 -c:v ffv1 clut.nut
12954 Note: make sure you use a lossless codec.
12956 Then use it with @code{haldclut} to apply it on some random stream:
12958 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12961 The Hald CLUT will be applied to the 10 first seconds (duration of
12962 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12963 to the remaining frames of the @code{mandelbrot} stream.
12965 @subsubsection Hald CLUT with preview
12967 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12968 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12969 biggest possible square starting at the top left of the picture. The remaining
12970 padding pixels (bottom or right) will be ignored. This area can be used to add
12971 a preview of the Hald CLUT.
12973 Typically, the following generated Hald CLUT will be supported by the
12974 @code{haldclut} filter:
12977 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12978 pad=iw+320 [padded_clut];
12979 smptebars=s=320x256, split [a][b];
12980 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12981 [main][b] overlay=W-320" -frames:v 1 clut.png
12984 It contains the original and a preview of the effect of the CLUT: SMPTE color
12985 bars are displayed on the right-top, and below the same color bars processed by
12988 Then, the effect of this Hald CLUT can be visualized with:
12990 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12995 Flip the input video horizontally.
12997 For example, to horizontally flip the input video with @command{ffmpeg}:
12999 ffmpeg -i in.avi -vf "hflip" out.avi
13003 This filter applies a global color histogram equalization on a
13006 It can be used to correct video that has a compressed range of pixel
13007 intensities. The filter redistributes the pixel intensities to
13008 equalize their distribution across the intensity range. It may be
13009 viewed as an "automatically adjusting contrast filter". This filter is
13010 useful only for correcting degraded or poorly captured source
13013 The filter accepts the following options:
13017 Determine the amount of equalization to be applied. As the strength
13018 is reduced, the distribution of pixel intensities more-and-more
13019 approaches that of the input frame. The value must be a float number
13020 in the range [0,1] and defaults to 0.200.
13023 Set the maximum intensity that can generated and scale the output
13024 values appropriately. The strength should be set as desired and then
13025 the intensity can be limited if needed to avoid washing-out. The value
13026 must be a float number in the range [0,1] and defaults to 0.210.
13029 Set the antibanding level. If enabled the filter will randomly vary
13030 the luminance of output pixels by a small amount to avoid banding of
13031 the histogram. Possible values are @code{none}, @code{weak} or
13032 @code{strong}. It defaults to @code{none}.
13038 Compute and draw a color distribution histogram for the input video.
13040 The computed histogram is a representation of the color component
13041 distribution in an image.
13043 Standard histogram displays the color components distribution in an image.
13044 Displays color graph for each color component. Shows distribution of
13045 the Y, U, V, A or R, G, B components, depending on input format, in the
13046 current frame. Below each graph a color component scale meter is shown.
13048 The filter accepts the following options:
13052 Set height of level. Default value is @code{200}.
13053 Allowed range is [50, 2048].
13056 Set height of color scale. Default value is @code{12}.
13057 Allowed range is [0, 40].
13061 It accepts the following values:
13064 Per color component graphs are placed below each other.
13067 Per color component graphs are placed side by side.
13070 Presents information identical to that in the @code{parade}, except
13071 that the graphs representing color components are superimposed directly
13074 Default is @code{stack}.
13077 Set mode. Can be either @code{linear}, or @code{logarithmic}.
13078 Default is @code{linear}.
13081 Set what color components to display.
13082 Default is @code{7}.
13085 Set foreground opacity. Default is @code{0.7}.
13088 Set background opacity. Default is @code{0.5}.
13091 @subsection Examples
13096 Calculate and draw histogram:
13098 ffplay -i input -vf histogram
13106 This is a high precision/quality 3d denoise filter. It aims to reduce
13107 image noise, producing smooth images and making still images really
13108 still. It should enhance compressibility.
13110 It accepts the following optional parameters:
13114 A non-negative floating point number which specifies spatial luma strength.
13115 It defaults to 4.0.
13117 @item chroma_spatial
13118 A non-negative floating point number which specifies spatial chroma strength.
13119 It defaults to 3.0*@var{luma_spatial}/4.0.
13122 A floating point number which specifies luma temporal strength. It defaults to
13123 6.0*@var{luma_spatial}/4.0.
13126 A floating point number which specifies chroma temporal strength. It defaults to
13127 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
13130 @subsection Commands
13131 This filter supports same @ref{commands} as options.
13132 The command accepts the same syntax of the corresponding option.
13134 If the specified expression is not valid, it is kept at its current
13137 @anchor{hwdownload}
13138 @section hwdownload
13140 Download hardware frames to system memory.
13142 The input must be in hardware frames, and the output a non-hardware format.
13143 Not all formats will be supported on the output - it may be necessary to insert
13144 an additional @option{format} filter immediately following in the graph to get
13145 the output in a supported format.
13149 Map hardware frames to system memory or to another device.
13151 This filter has several different modes of operation; which one is used depends
13152 on the input and output formats:
13155 Hardware frame input, normal frame output
13157 Map the input frames to system memory and pass them to the output. If the
13158 original hardware frame is later required (for example, after overlaying
13159 something else on part of it), the @option{hwmap} filter can be used again
13160 in the next mode to retrieve it.
13162 Normal frame input, hardware frame output
13164 If the input is actually a software-mapped hardware frame, then unmap it -
13165 that is, return the original hardware frame.
13167 Otherwise, a device must be provided. Create new hardware surfaces on that
13168 device for the output, then map them back to the software format at the input
13169 and give those frames to the preceding filter. This will then act like the
13170 @option{hwupload} filter, but may be able to avoid an additional copy when
13171 the input is already in a compatible format.
13173 Hardware frame input and output
13175 A device must be supplied for the output, either directly or with the
13176 @option{derive_device} option. The input and output devices must be of
13177 different types and compatible - the exact meaning of this is
13178 system-dependent, but typically it means that they must refer to the same
13179 underlying hardware context (for example, refer to the same graphics card).
13181 If the input frames were originally created on the output device, then unmap
13182 to retrieve the original frames.
13184 Otherwise, map the frames to the output device - create new hardware frames
13185 on the output corresponding to the frames on the input.
13188 The following additional parameters are accepted:
13192 Set the frame mapping mode. Some combination of:
13195 The mapped frame should be readable.
13197 The mapped frame should be writeable.
13199 The mapping will always overwrite the entire frame.
13201 This may improve performance in some cases, as the original contents of the
13202 frame need not be loaded.
13204 The mapping must not involve any copying.
13206 Indirect mappings to copies of frames are created in some cases where either
13207 direct mapping is not possible or it would have unexpected properties.
13208 Setting this flag ensures that the mapping is direct and will fail if that is
13211 Defaults to @var{read+write} if not specified.
13213 @item derive_device @var{type}
13214 Rather than using the device supplied at initialisation, instead derive a new
13215 device of type @var{type} from the device the input frames exist on.
13218 In a hardware to hardware mapping, map in reverse - create frames in the sink
13219 and map them back to the source. This may be necessary in some cases where
13220 a mapping in one direction is required but only the opposite direction is
13221 supported by the devices being used.
13223 This option is dangerous - it may break the preceding filter in undefined
13224 ways if there are any additional constraints on that filter's output.
13225 Do not use it without fully understanding the implications of its use.
13231 Upload system memory frames to hardware surfaces.
13233 The device to upload to must be supplied when the filter is initialised. If
13234 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
13235 option or with the @option{derive_device} option. The input and output devices
13236 must be of different types and compatible - the exact meaning of this is
13237 system-dependent, but typically it means that they must refer to the same
13238 underlying hardware context (for example, refer to the same graphics card).
13240 The following additional parameters are accepted:
13243 @item derive_device @var{type}
13244 Rather than using the device supplied at initialisation, instead derive a new
13245 device of type @var{type} from the device the input frames exist on.
13248 @anchor{hwupload_cuda}
13249 @section hwupload_cuda
13251 Upload system memory frames to a CUDA device.
13253 It accepts the following optional parameters:
13257 The number of the CUDA device to use
13262 Apply a high-quality magnification filter designed for pixel art. This filter
13263 was originally created by Maxim Stepin.
13265 It accepts the following option:
13269 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
13270 @code{hq3x} and @code{4} for @code{hq4x}.
13271 Default is @code{3}.
13275 Stack input videos horizontally.
13277 All streams must be of same pixel format and of same height.
13279 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
13280 to create same output.
13282 The filter accepts the following option:
13286 Set number of input streams. Default is 2.
13289 If set to 1, force the output to terminate when the shortest input
13290 terminates. Default value is 0.
13295 Modify the hue and/or the saturation of the input.
13297 It accepts the following parameters:
13301 Specify the hue angle as a number of degrees. It accepts an expression,
13302 and defaults to "0".
13305 Specify the saturation in the [-10,10] range. It accepts an expression and
13309 Specify the hue angle as a number of radians. It accepts an
13310 expression, and defaults to "0".
13313 Specify the brightness in the [-10,10] range. It accepts an expression and
13317 @option{h} and @option{H} are mutually exclusive, and can't be
13318 specified at the same time.
13320 The @option{b}, @option{h}, @option{H} and @option{s} option values are
13321 expressions containing the following constants:
13325 frame count of the input frame starting from 0
13328 presentation timestamp of the input frame expressed in time base units
13331 frame rate of the input video, NAN if the input frame rate is unknown
13334 timestamp expressed in seconds, NAN if the input timestamp is unknown
13337 time base of the input video
13340 @subsection Examples
13344 Set the hue to 90 degrees and the saturation to 1.0:
13350 Same command but expressing the hue in radians:
13356 Rotate hue and make the saturation swing between 0
13357 and 2 over a period of 1 second:
13359 hue="H=2*PI*t: s=sin(2*PI*t)+1"
13363 Apply a 3 seconds saturation fade-in effect starting at 0:
13365 hue="s=min(t/3\,1)"
13368 The general fade-in expression can be written as:
13370 hue="s=min(0\, max((t-START)/DURATION\, 1))"
13374 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13376 hue="s=max(0\, min(1\, (8-t)/3))"
13379 The general fade-out expression can be written as:
13381 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13386 @subsection Commands
13388 This filter supports the following commands:
13394 Modify the hue and/or the saturation and/or brightness of the input video.
13395 The command accepts the same syntax of the corresponding option.
13397 If the specified expression is not valid, it is kept at its current
13401 @section hysteresis
13403 Grow first stream into second stream by connecting components.
13404 This makes it possible to build more robust edge masks.
13406 This filter accepts the following options:
13410 Set which planes will be processed as bitmap, unprocessed planes will be
13411 copied from first stream.
13412 By default value 0xf, all planes will be processed.
13415 Set threshold which is used in filtering. If pixel component value is higher than
13416 this value filter algorithm for connecting components is activated.
13417 By default value is 0.
13420 The @code{hysteresis} filter also supports the @ref{framesync} options.
13424 Obtain the identity score between two input videos.
13426 This filter takes two input videos.
13428 Both input videos must have the same resolution and pixel format for
13429 this filter to work correctly. Also it assumes that both inputs
13430 have the same number of frames, which are compared one by one.
13432 The obtained per component, average, min and max identity score is printed through
13433 the logging system.
13435 The filter stores the calculated identity scores of each frame in frame metadata.
13437 In the below example the input file @file{main.mpg} being processed is compared
13438 with the reference file @file{ref.mpg}.
13441 ffmpeg -i main.mpg -i ref.mpg -lavfi identity -f null -
13446 Detect video interlacing type.
13448 This filter tries to detect if the input frames are interlaced, progressive,
13449 top or bottom field first. It will also try to detect fields that are
13450 repeated between adjacent frames (a sign of telecine).
13452 Single frame detection considers only immediately adjacent frames when classifying each frame.
13453 Multiple frame detection incorporates the classification history of previous frames.
13455 The filter will log these metadata values:
13458 @item single.current_frame
13459 Detected type of current frame using single-frame detection. One of:
13460 ``tff'' (top field first), ``bff'' (bottom field first),
13461 ``progressive'', or ``undetermined''
13464 Cumulative number of frames detected as top field first using single-frame detection.
13467 Cumulative number of frames detected as top field first using multiple-frame detection.
13470 Cumulative number of frames detected as bottom field first using single-frame detection.
13472 @item multiple.current_frame
13473 Detected type of current frame using multiple-frame detection. One of:
13474 ``tff'' (top field first), ``bff'' (bottom field first),
13475 ``progressive'', or ``undetermined''
13478 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13480 @item single.progressive
13481 Cumulative number of frames detected as progressive using single-frame detection.
13483 @item multiple.progressive
13484 Cumulative number of frames detected as progressive using multiple-frame detection.
13486 @item single.undetermined
13487 Cumulative number of frames that could not be classified using single-frame detection.
13489 @item multiple.undetermined
13490 Cumulative number of frames that could not be classified using multiple-frame detection.
13492 @item repeated.current_frame
13493 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13495 @item repeated.neither
13496 Cumulative number of frames with no repeated field.
13499 Cumulative number of frames with the top field repeated from the previous frame's top field.
13501 @item repeated.bottom
13502 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13505 The filter accepts the following options:
13509 Set interlacing threshold.
13511 Set progressive threshold.
13513 Threshold for repeated field detection.
13515 Number of frames after which a given frame's contribution to the
13516 statistics is halved (i.e., it contributes only 0.5 to its
13517 classification). The default of 0 means that all frames seen are given
13518 full weight of 1.0 forever.
13519 @item analyze_interlaced_flag
13520 When this is not 0 then idet will use the specified number of frames to determine
13521 if the interlaced flag is accurate, it will not count undetermined frames.
13522 If the flag is found to be accurate it will be used without any further
13523 computations, if it is found to be inaccurate it will be cleared without any
13524 further computations. This allows inserting the idet filter as a low computational
13525 method to clean up the interlaced flag
13530 Deinterleave or interleave fields.
13532 This filter allows one to process interlaced images fields without
13533 deinterlacing them. Deinterleaving splits the input frame into 2
13534 fields (so called half pictures). Odd lines are moved to the top
13535 half of the output image, even lines to the bottom half.
13536 You can process (filter) them independently and then re-interleave them.
13538 The filter accepts the following options:
13542 @item chroma_mode, c
13543 @item alpha_mode, a
13544 Available values for @var{luma_mode}, @var{chroma_mode} and
13545 @var{alpha_mode} are:
13551 @item deinterleave, d
13552 Deinterleave fields, placing one above the other.
13554 @item interleave, i
13555 Interleave fields. Reverse the effect of deinterleaving.
13557 Default value is @code{none}.
13559 @item luma_swap, ls
13560 @item chroma_swap, cs
13561 @item alpha_swap, as
13562 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13565 @subsection Commands
13567 This filter supports the all above options as @ref{commands}.
13571 Apply inflate effect to the video.
13573 This filter replaces the pixel by the local(3x3) average by taking into account
13574 only values higher than the pixel.
13576 It accepts the following options:
13583 Limit the maximum change for each plane, default is 65535.
13584 If 0, plane will remain unchanged.
13587 @subsection Commands
13589 This filter supports the all above options as @ref{commands}.
13593 Simple interlacing filter from progressive contents. This interleaves upper (or
13594 lower) lines from odd frames with lower (or upper) lines from even frames,
13595 halving the frame rate and preserving image height.
13598 Original Original New Frame
13599 Frame 'j' Frame 'j+1' (tff)
13600 ========== =========== ==================
13601 Line 0 --------------------> Frame 'j' Line 0
13602 Line 1 Line 1 ----> Frame 'j+1' Line 1
13603 Line 2 ---------------------> Frame 'j' Line 2
13604 Line 3 Line 3 ----> Frame 'j+1' Line 3
13606 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13609 It accepts the following optional parameters:
13613 This determines whether the interlaced frame is taken from the even
13614 (tff - default) or odd (bff) lines of the progressive frame.
13617 Vertical lowpass filter to avoid twitter interlacing and
13618 reduce moire patterns.
13622 Disable vertical lowpass filter
13625 Enable linear filter (default)
13628 Enable complex filter. This will slightly less reduce twitter and moire
13629 but better retain detail and subjective sharpness impression.
13636 Deinterlace input video by applying Donald Graft's adaptive kernel
13637 deinterling. Work on interlaced parts of a video to produce
13638 progressive frames.
13640 The description of the accepted parameters follows.
13644 Set the threshold which affects the filter's tolerance when
13645 determining if a pixel line must be processed. It must be an integer
13646 in the range [0,255] and defaults to 10. A value of 0 will result in
13647 applying the process on every pixels.
13650 Paint pixels exceeding the threshold value to white if set to 1.
13654 Set the fields order. Swap fields if set to 1, leave fields alone if
13658 Enable additional sharpening if set to 1. Default is 0.
13661 Enable twoway sharpening if set to 1. Default is 0.
13664 @subsection Examples
13668 Apply default values:
13670 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13674 Enable additional sharpening:
13680 Paint processed pixels in white:
13687 Apply kirsch operator to input video stream.
13689 The filter accepts the following option:
13693 Set which planes will be processed, unprocessed planes will be copied.
13694 By default value 0xf, all planes will be processed.
13697 Set value which will be multiplied with filtered result.
13700 Set value which will be added to filtered result.
13703 @subsection Commands
13705 This filter supports the all above options as @ref{commands}.
13709 Slowly update darker pixels.
13711 This filter makes short flashes of light appear longer.
13712 This filter accepts the following options:
13716 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13719 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13722 @subsection Commands
13724 This filter supports the all above options as @ref{commands}.
13726 @section lenscorrection
13728 Correct radial lens distortion
13730 This filter can be used to correct for radial distortion as can result from the use
13731 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13732 one can use tools available for example as part of opencv or simply trial-and-error.
13733 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13734 and extract the k1 and k2 coefficients from the resulting matrix.
13736 Note that effectively the same filter is available in the open-source tools Krita and
13737 Digikam from the KDE project.
13739 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13740 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13741 brightness distribution, so you may want to use both filters together in certain
13742 cases, though you will have to take care of ordering, i.e. whether vignetting should
13743 be applied before or after lens correction.
13745 @subsection Options
13747 The filter accepts the following options:
13751 Relative x-coordinate of the focal point of the image, and thereby the center of the
13752 distortion. This value has a range [0,1] and is expressed as fractions of the image
13753 width. Default is 0.5.
13755 Relative y-coordinate of the focal point of the image, and thereby the center of the
13756 distortion. This value has a range [0,1] and is expressed as fractions of the image
13757 height. Default is 0.5.
13759 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13760 no correction. Default is 0.
13762 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13763 0 means no correction. Default is 0.
13765 Set interpolation type. Can be @code{nearest} or @code{bilinear}.
13766 Default is @code{nearest}.
13768 Specify the color of the unmapped pixels. For the syntax of this option,
13769 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
13770 manual,ffmpeg-utils}. Default color is @code{black@@0}.
13773 The formula that generates the correction is:
13775 @var{r_src} = @var{r_tgt} * (1 + @var{k1} * (@var{r_tgt} / @var{r_0})^2 + @var{k2} * (@var{r_tgt} / @var{r_0})^4)
13777 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13778 distances from the focal point in the source and target images, respectively.
13780 @subsection Commands
13782 This filter supports the all above options as @ref{commands}.
13786 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13788 The @code{lensfun} filter requires the camera make, camera model, and lens model
13789 to apply the lens correction. The filter will load the lensfun database and
13790 query it to find the corresponding camera and lens entries in the database. As
13791 long as these entries can be found with the given options, the filter can
13792 perform corrections on frames. Note that incomplete strings will result in the
13793 filter choosing the best match with the given options, and the filter will
13794 output the chosen camera and lens models (logged with level "info"). You must
13795 provide the make, camera model, and lens model as they are required.
13797 The filter accepts the following options:
13801 The make of the camera (for example, "Canon"). This option is required.
13804 The model of the camera (for example, "Canon EOS 100D"). This option is
13808 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13809 option is required.
13812 The type of correction to apply. The following values are valid options:
13816 Enables fixing lens vignetting.
13819 Enables fixing lens geometry. This is the default.
13822 Enables fixing chromatic aberrations.
13825 Enables fixing lens vignetting and lens geometry.
13828 Enables fixing lens vignetting and chromatic aberrations.
13831 Enables fixing both lens geometry and chromatic aberrations.
13834 Enables all possible corrections.
13838 The focal length of the image/video (zoom; expected constant for video). For
13839 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13840 range should be chosen when using that lens. Default 18.
13843 The aperture of the image/video (expected constant for video). Note that
13844 aperture is only used for vignetting correction. Default 3.5.
13846 @item focus_distance
13847 The focus distance of the image/video (expected constant for video). Note that
13848 focus distance is only used for vignetting and only slightly affects the
13849 vignetting correction process. If unknown, leave it at the default value (which
13853 The scale factor which is applied after transformation. After correction the
13854 video is no longer necessarily rectangular. This parameter controls how much of
13855 the resulting image is visible. The value 0 means that a value will be chosen
13856 automatically such that there is little or no unmapped area in the output
13857 image. 1.0 means that no additional scaling is done. Lower values may result
13858 in more of the corrected image being visible, while higher values may avoid
13859 unmapped areas in the output.
13861 @item target_geometry
13862 The target geometry of the output image/video. The following values are valid
13866 @item rectilinear (default)
13869 @item equirectangular
13870 @item fisheye_orthographic
13871 @item fisheye_stereographic
13872 @item fisheye_equisolid
13873 @item fisheye_thoby
13876 Apply the reverse of image correction (instead of correcting distortion, apply
13879 @item interpolation
13880 The type of interpolation used when correcting distortion. The following values
13885 @item linear (default)
13890 @subsection Examples
13894 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13895 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13899 ffmpeg -i input.mov -vf lensfun=make=Canon:model="Canon EOS 100D":lens_model="Canon EF-S 18-55mm f/3.5-5.6 IS STM":focal_length=18:aperture=8 -c:v h264 -b:v 8000k output.mov
13903 Apply the same as before, but only for the first 5 seconds of video.
13906 ffmpeg -i input.mov -vf lensfun=make=Canon:model="Canon EOS 100D":lens_model="Canon EF-S 18-55mm f/3.5-5.6 IS STM":focal_length=18:aperture=8:enable='lte(t\,5)' -c:v h264 -b:v 8000k output.mov
13913 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13914 score between two input videos.
13916 The obtained VMAF score is printed through the logging system.
13918 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13919 After installing the library it can be enabled using:
13920 @code{./configure --enable-libvmaf}.
13921 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13923 The filter has following options:
13927 Set the model path which is to be used for SVM.
13928 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13931 Set the file path to be used to store logs.
13934 Set the format of the log file (csv, json or xml).
13936 @item enable_transform
13937 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13938 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13939 Default value: @code{false}
13942 Invokes the phone model which will generate VMAF scores higher than in the
13943 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13944 Default value: @code{false}
13947 Enables computing psnr along with vmaf.
13948 Default value: @code{false}
13951 Enables computing ssim along with vmaf.
13952 Default value: @code{false}
13955 Enables computing ms_ssim along with vmaf.
13956 Default value: @code{false}
13959 Set the pool method to be used for computing vmaf.
13960 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13963 Set number of threads to be used when computing vmaf.
13964 Default value: @code{0}, which makes use of all available logical processors.
13967 Set interval for frame subsampling used when computing vmaf.
13968 Default value: @code{1}
13970 @item enable_conf_interval
13971 Enables confidence interval.
13972 Default value: @code{false}
13975 This filter also supports the @ref{framesync} options.
13977 @subsection Examples
13980 On the below examples the input file @file{main.mpg} being processed is
13981 compared with the reference file @file{ref.mpg}.
13984 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13988 Example with options:
13990 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13994 Example with options and different containers:
13996 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]libvmaf=psnr=1:log_fmt=json" -f null -
14002 Limits the pixel components values to the specified range [min, max].
14004 The filter accepts the following options:
14008 Lower bound. Defaults to the lowest allowed value for the input.
14011 Upper bound. Defaults to the highest allowed value for the input.
14014 Specify which planes will be processed. Defaults to all available.
14017 @subsection Commands
14019 This filter supports the all above options as @ref{commands}.
14025 The filter accepts the following options:
14029 Set the number of loops. Setting this value to -1 will result in infinite loops.
14033 Set maximal size in number of frames. Default is 0.
14036 Set first frame of loop. Default is 0.
14039 @subsection Examples
14043 Loop single first frame infinitely:
14045 loop=loop=-1:size=1:start=0
14049 Loop single first frame 10 times:
14051 loop=loop=10:size=1:start=0
14055 Loop 10 first frames 5 times:
14057 loop=loop=5:size=10:start=0
14063 Apply a 1D LUT to an input video.
14065 The filter accepts the following options:
14069 Set the 1D LUT file name.
14071 Currently supported formats:
14080 Select interpolation mode.
14082 Available values are:
14086 Use values from the nearest defined point.
14088 Interpolate values using the linear interpolation.
14090 Interpolate values using the cosine interpolation.
14092 Interpolate values using the cubic interpolation.
14094 Interpolate values using the spline interpolation.
14098 @subsection Commands
14100 This filter supports the all above options as @ref{commands}.
14105 Apply a 3D LUT to an input video.
14107 The filter accepts the following options:
14111 Set the 3D LUT file name.
14113 Currently supported formats:
14127 Select interpolation mode.
14129 Available values are:
14133 Use values from the nearest defined point.
14135 Interpolate values using the 8 points defining a cube.
14137 Interpolate values using a tetrahedron.
14139 Interpolate values using a pyramid.
14141 Interpolate values using a prism.
14145 @subsection Commands
14147 This filter supports the @code{interp} option as @ref{commands}.
14151 Turn certain luma values into transparency.
14153 The filter accepts the following options:
14157 Set the luma which will be used as base for transparency.
14158 Default value is @code{0}.
14161 Set the range of luma values to be keyed out.
14162 Default value is @code{0.01}.
14165 Set the range of softness. Default value is @code{0}.
14166 Use this to control gradual transition from zero to full transparency.
14169 @subsection Commands
14170 This filter supports same @ref{commands} as options.
14171 The command accepts the same syntax of the corresponding option.
14173 If the specified expression is not valid, it is kept at its current
14176 @section lut, lutrgb, lutyuv
14178 Compute a look-up table for binding each pixel component input value
14179 to an output value, and apply it to the input video.
14181 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
14182 to an RGB input video.
14184 These filters accept the following parameters:
14187 set first pixel component expression
14189 set second pixel component expression
14191 set third pixel component expression
14193 set fourth pixel component expression, corresponds to the alpha component
14196 set red component expression
14198 set green component expression
14200 set blue component expression
14202 alpha component expression
14205 set Y/luminance component expression
14207 set U/Cb component expression
14209 set V/Cr component expression
14212 Each of them specifies the expression to use for computing the lookup table for
14213 the corresponding pixel component values.
14215 The exact component associated to each of the @var{c*} options depends on the
14218 The @var{lut} filter requires either YUV or RGB pixel formats in input,
14219 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
14221 The expressions can contain the following constants and functions:
14226 The input width and height.
14229 The input value for the pixel component.
14232 The input value, clipped to the @var{minval}-@var{maxval} range.
14235 The maximum value for the pixel component.
14238 The minimum value for the pixel component.
14241 The negated value for the pixel component value, clipped to the
14242 @var{minval}-@var{maxval} range; it corresponds to the expression
14243 "maxval-clipval+minval".
14246 The computed value in @var{val}, clipped to the
14247 @var{minval}-@var{maxval} range.
14249 @item gammaval(gamma)
14250 The computed gamma correction value of the pixel component value,
14251 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
14253 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
14257 All expressions default to "val".
14259 @subsection Commands
14261 This filter supports same @ref{commands} as options.
14263 @subsection Examples
14267 Negate input video:
14269 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
14270 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
14273 The above is the same as:
14275 lutrgb="r=negval:g=negval:b=negval"
14276 lutyuv="y=negval:u=negval:v=negval"
14286 Remove chroma components, turning the video into a graytone image:
14288 lutyuv="u=128:v=128"
14292 Apply a luma burning effect:
14298 Remove green and blue components:
14304 Set a constant alpha channel value on input:
14306 format=rgba,lutrgb=a="maxval-minval/2"
14310 Correct luminance gamma by a factor of 0.5:
14312 lutyuv=y=gammaval(0.5)
14316 Discard least significant bits of luma:
14318 lutyuv=y='bitand(val, 128+64+32)'
14322 Technicolor like effect:
14324 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
14328 @section lut2, tlut2
14330 The @code{lut2} filter takes two input streams and outputs one
14333 The @code{tlut2} (time lut2) filter takes two consecutive frames
14334 from one single stream.
14336 This filter accepts the following parameters:
14339 set first pixel component expression
14341 set second pixel component expression
14343 set third pixel component expression
14345 set fourth pixel component expression, corresponds to the alpha component
14348 set output bit depth, only available for @code{lut2} filter. By default is 0,
14349 which means bit depth is automatically picked from first input format.
14352 The @code{lut2} filter also supports the @ref{framesync} options.
14354 Each of them specifies the expression to use for computing the lookup table for
14355 the corresponding pixel component values.
14357 The exact component associated to each of the @var{c*} options depends on the
14360 The expressions can contain the following constants:
14365 The input width and height.
14368 The first input value for the pixel component.
14371 The second input value for the pixel component.
14374 The first input video bit depth.
14377 The second input video bit depth.
14380 All expressions default to "x".
14382 @subsection Commands
14384 This filter supports the all above options as @ref{commands} except option @code{d}.
14386 @subsection Examples
14390 Highlight differences between two RGB video streams:
14392 lut2='ifnot(x-y,0,pow(2,bdx)-1):ifnot(x-y,0,pow(2,bdx)-1):ifnot(x-y,0,pow(2,bdx)-1)'
14396 Highlight differences between two YUV video streams:
14398 lut2='ifnot(x-y,0,pow(2,bdx)-1):ifnot(x-y,pow(2,bdx-1),pow(2,bdx)-1):ifnot(x-y,pow(2,bdx-1),pow(2,bdx)-1)'
14402 Show max difference between two video streams:
14404 lut2='if(lt(x,y),0,if(gt(x,y),pow(2,bdx)-1,pow(2,bdx-1))):if(lt(x,y),0,if(gt(x,y),pow(2,bdx)-1,pow(2,bdx-1))):if(lt(x,y),0,if(gt(x,y),pow(2,bdx)-1,pow(2,bdx-1)))'
14408 @section maskedclamp
14410 Clamp the first input stream with the second input and third input stream.
14412 Returns the value of first stream to be between second input
14413 stream - @code{undershoot} and third input stream + @code{overshoot}.
14415 This filter accepts the following options:
14418 Default value is @code{0}.
14421 Default value is @code{0}.
14424 Set which planes will be processed as bitmap, unprocessed planes will be
14425 copied from first stream.
14426 By default value 0xf, all planes will be processed.
14429 @subsection Commands
14431 This filter supports the all above options as @ref{commands}.
14435 Merge the second and third input stream into output stream using absolute differences
14436 between second input stream and first input stream and absolute difference between
14437 third input stream and first input stream. The picked value will be from second input
14438 stream if second absolute difference is greater than first one or from third input stream
14441 This filter accepts the following options:
14444 Set which planes will be processed as bitmap, unprocessed planes will be
14445 copied from first stream.
14446 By default value 0xf, all planes will be processed.
14449 @subsection Commands
14451 This filter supports the all above options as @ref{commands}.
14453 @section maskedmerge
14455 Merge the first input stream with the second input stream using per pixel
14456 weights in the third input stream.
14458 A value of 0 in the third stream pixel component means that pixel component
14459 from first stream is returned unchanged, while maximum value (eg. 255 for
14460 8-bit videos) means that pixel component from second stream is returned
14461 unchanged. Intermediate values define the amount of merging between both
14462 input stream's pixel components.
14464 This filter accepts the following options:
14467 Set which planes will be processed as bitmap, unprocessed planes will be
14468 copied from first stream.
14469 By default value 0xf, all planes will be processed.
14472 @subsection Commands
14474 This filter supports the all above options as @ref{commands}.
14478 Merge the second and third input stream into output stream using absolute differences
14479 between second input stream and first input stream and absolute difference between
14480 third input stream and first input stream. The picked value will be from second input
14481 stream if second absolute difference is less than first one or from third input stream
14484 This filter accepts the following options:
14487 Set which planes will be processed as bitmap, unprocessed planes will be
14488 copied from first stream.
14489 By default value 0xf, all planes will be processed.
14492 @subsection Commands
14494 This filter supports the all above options as @ref{commands}.
14496 @section maskedthreshold
14497 Pick pixels comparing absolute difference of two video streams with fixed
14500 If absolute difference between pixel component of first and second video
14501 stream is equal or lower than user supplied threshold than pixel component
14502 from first video stream is picked, otherwise pixel component from second
14503 video stream is picked.
14505 This filter accepts the following options:
14508 Set threshold used when picking pixels from absolute difference from two input
14512 Set which planes will be processed as bitmap, unprocessed planes will be
14513 copied from second stream.
14514 By default value 0xf, all planes will be processed.
14517 @subsection Commands
14519 This filter supports the all above options as @ref{commands}.
14522 Create mask from input video.
14524 For example it is useful to create motion masks after @code{tblend} filter.
14526 This filter accepts the following options:
14530 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14533 Set high threshold. Any pixel component higher than this value will be set to max value
14534 allowed for current pixel format.
14537 Set planes to filter, by default all available planes are filtered.
14540 Fill all frame pixels with this value.
14543 Set max average pixel value for frame. If sum of all pixel components is higher that this
14544 average, output frame will be completely filled with value set by @var{fill} option.
14545 Typically useful for scene changes when used in combination with @code{tblend} filter.
14548 @subsection Commands
14550 This filter supports the all above options as @ref{commands}.
14554 Apply motion-compensation deinterlacing.
14556 It needs one field per frame as input and must thus be used together
14557 with yadif=1/3 or equivalent.
14559 This filter accepts the following options:
14562 Set the deinterlacing mode.
14564 It accepts one of the following values:
14569 use iterative motion estimation
14571 like @samp{slow}, but use multiple reference frames.
14573 Default value is @samp{fast}.
14576 Set the picture field parity assumed for the input video. It must be
14577 one of the following values:
14581 assume top field first
14583 assume bottom field first
14586 Default value is @samp{bff}.
14589 Set per-block quantization parameter (QP) used by the internal
14592 Higher values should result in a smoother motion vector field but less
14593 optimal individual vectors. Default value is 1.
14598 Pick median pixel from certain rectangle defined by radius.
14600 This filter accepts the following options:
14604 Set horizontal radius size. Default value is @code{1}.
14605 Allowed range is integer from 1 to 127.
14608 Set which planes to process. Default is @code{15}, which is all available planes.
14611 Set vertical radius size. Default value is @code{0}.
14612 Allowed range is integer from 0 to 127.
14613 If it is 0, value will be picked from horizontal @code{radius} option.
14616 Set median percentile. Default value is @code{0.5}.
14617 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14618 minimum values, and @code{1} maximum values.
14621 @subsection Commands
14622 This filter supports same @ref{commands} as options.
14623 The command accepts the same syntax of the corresponding option.
14625 If the specified expression is not valid, it is kept at its current
14628 @section mergeplanes
14630 Merge color channel components from several video streams.
14632 The filter accepts up to 4 input streams, and merge selected input
14633 planes to the output video.
14635 This filter accepts the following options:
14638 Set input to output plane mapping. Default is @code{0}.
14640 The mappings is specified as a bitmap. It should be specified as a
14641 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14642 mapping for the first plane of the output stream. 'A' sets the number of
14643 the input stream to use (from 0 to 3), and 'a' the plane number of the
14644 corresponding input to use (from 0 to 3). The rest of the mappings is
14645 similar, 'Bb' describes the mapping for the output stream second
14646 plane, 'Cc' describes the mapping for the output stream third plane and
14647 'Dd' describes the mapping for the output stream fourth plane.
14650 Set output pixel format. Default is @code{yuva444p}.
14653 @subsection Examples
14657 Merge three gray video streams of same width and height into single video stream:
14659 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14663 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14665 [a0][a1]mergeplanes=0x00010210:yuva444p
14669 Swap Y and A plane in yuva444p stream:
14671 format=yuva444p,mergeplanes=0x03010200:yuva444p
14675 Swap U and V plane in yuv420p stream:
14677 format=yuv420p,mergeplanes=0x000201:yuv420p
14681 Cast a rgb24 clip to yuv444p:
14683 format=rgb24,mergeplanes=0x000102:yuv444p
14689 Estimate and export motion vectors using block matching algorithms.
14690 Motion vectors are stored in frame side data to be used by other filters.
14692 This filter accepts the following options:
14695 Specify the motion estimation method. Accepts one of the following values:
14699 Exhaustive search algorithm.
14701 Three step search algorithm.
14703 Two dimensional logarithmic search algorithm.
14705 New three step search algorithm.
14707 Four step search algorithm.
14709 Diamond search algorithm.
14711 Hexagon-based search algorithm.
14713 Enhanced predictive zonal search algorithm.
14715 Uneven multi-hexagon search algorithm.
14717 Default value is @samp{esa}.
14720 Macroblock size. Default @code{16}.
14723 Search parameter. Default @code{7}.
14726 @section midequalizer
14728 Apply Midway Image Equalization effect using two video streams.
14730 Midway Image Equalization adjusts a pair of images to have the same
14731 histogram, while maintaining their dynamics as much as possible. It's
14732 useful for e.g. matching exposures from a pair of stereo cameras.
14734 This filter has two inputs and one output, which must be of same pixel format, but
14735 may be of different sizes. The output of filter is first input adjusted with
14736 midway histogram of both inputs.
14738 This filter accepts the following option:
14742 Set which planes to process. Default is @code{15}, which is all available planes.
14745 @section minterpolate
14747 Convert the video to specified frame rate using motion interpolation.
14749 This filter accepts the following options:
14752 Specify the output frame rate. This can be rational e.g. @code{60000/1001}. Frames are dropped if @var{fps} is lower than source fps. Default @code{60}.
14755 Motion interpolation mode. Following values are accepted:
14758 Duplicate previous or next frame for interpolating new ones.
14760 Blend source frames. Interpolated frame is mean of previous and next frames.
14762 Motion compensated interpolation. Following options are effective when this mode is selected:
14766 Motion compensation mode. Following values are accepted:
14769 Overlapped block motion compensation.
14771 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14773 Default mode is @samp{obmc}.
14776 Motion estimation mode. Following values are accepted:
14779 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14781 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14783 Default mode is @samp{bilat}.
14786 The algorithm to be used for motion estimation. Following values are accepted:
14789 Exhaustive search algorithm.
14791 Three step search algorithm.
14793 Two dimensional logarithmic search algorithm.
14795 New three step search algorithm.
14797 Four step search algorithm.
14799 Diamond search algorithm.
14801 Hexagon-based search algorithm.
14803 Enhanced predictive zonal search algorithm.
14805 Uneven multi-hexagon search algorithm.
14807 Default algorithm is @samp{epzs}.
14810 Macroblock size. Default @code{16}.
14813 Motion estimation search parameter. Default @code{32}.
14816 Enable variable-size block motion compensation. Motion estimation is applied with smaller block sizes at object boundaries in order to make the them less blur. Default is @code{0} (disabled).
14821 Scene change detection method. Scene change leads motion vectors to be in random direction. Scene change detection replace interpolated frames by duplicate ones. May not be needed for other modes. Following values are accepted:
14824 Disable scene change detection.
14826 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14828 Default method is @samp{fdiff}.
14830 @item scd_threshold
14831 Scene change detection threshold. Default is @code{10.}.
14836 Mix several video input streams into one video stream.
14838 A description of the accepted options follows.
14842 The number of inputs. If unspecified, it defaults to 2.
14845 Specify weight of each input video stream as sequence.
14846 Each weight is separated by space. If number of weights
14847 is smaller than number of @var{frames} last specified
14848 weight will be used for all remaining unset weights.
14851 Specify scale, if it is set it will be multiplied with sum
14852 of each weight multiplied with pixel values to give final destination
14853 pixel value. By default @var{scale} is auto scaled to sum of weights.
14856 Specify how end of stream is determined.
14859 The duration of the longest input. (default)
14862 The duration of the shortest input.
14865 The duration of the first input.
14869 @subsection Commands
14871 This filter supports the following commands:
14875 Syntax is same as option with same name.
14878 @section monochrome
14879 Convert video to gray using custom color filter.
14881 A description of the accepted options follows.
14885 Set the chroma blue spot. Allowed range is from -1 to 1.
14886 Default value is 0.
14889 Set the chroma red spot. Allowed range is from -1 to 1.
14890 Default value is 0.
14893 Set the color filter size. Allowed range is from .1 to 10.
14894 Default value is 1.
14897 Set the highlights strength. Allowed range is from 0 to 1.
14898 Default value is 0.
14901 @subsection Commands
14903 This filter supports the all above options as @ref{commands}.
14905 @section mpdecimate
14907 Drop frames that do not differ greatly from the previous frame in
14908 order to reduce frame rate.
14910 The main use of this filter is for very-low-bitrate encoding
14911 (e.g. streaming over dialup modem), but it could in theory be used for
14912 fixing movies that were inverse-telecined incorrectly.
14914 A description of the accepted options follows.
14918 Set the maximum number of consecutive frames which can be dropped (if
14919 positive), or the minimum interval between dropped frames (if
14920 negative). If the value is 0, the frame is dropped disregarding the
14921 number of previous sequentially dropped frames.
14923 Default value is 0.
14928 Set the dropping threshold values.
14930 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14931 represent actual pixel value differences, so a threshold of 64
14932 corresponds to 1 unit of difference for each pixel, or the same spread
14933 out differently over the block.
14935 A frame is a candidate for dropping if no 8x8 blocks differ by more
14936 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14937 meaning the whole image) differ by more than a threshold of @option{lo}.
14939 Default value for @option{hi} is 64*12, default value for @option{lo} is
14940 64*5, and default value for @option{frac} is 0.33.
14945 Obtain the MSAD (Mean Sum of Absolute Differences) between two input videos.
14947 This filter takes two input videos.
14949 Both input videos must have the same resolution and pixel format for
14950 this filter to work correctly. Also it assumes that both inputs
14951 have the same number of frames, which are compared one by one.
14953 The obtained per component, average, min and max MSAD is printed through
14954 the logging system.
14956 The filter stores the calculated MSAD of each frame in frame metadata.
14958 In the below example the input file @file{main.mpg} being processed is compared
14959 with the reference file @file{ref.mpg}.
14962 ffmpeg -i main.mpg -i ref.mpg -lavfi msad -f null -
14967 Negate (invert) the input video.
14969 It accepts the following option:
14974 With value 1, it negates the alpha component, if present. Default value is 0.
14977 @subsection Commands
14979 This filter supports same @ref{commands} as options.
14984 Denoise frames using Non-Local Means algorithm.
14986 Each pixel is adjusted by looking for other pixels with similar contexts. This
14987 context similarity is defined by comparing their surrounding patches of size
14988 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14991 Note that the research area defines centers for patches, which means some
14992 patches will be made of pixels outside that research area.
14994 The filter accepts the following options.
14998 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
15001 Set patch size. Default is 7. Must be odd number in range [0, 99].
15004 Same as @option{p} but for chroma planes.
15006 The default value is @var{0} and means automatic.
15009 Set research size. Default is 15. Must be odd number in range [0, 99].
15012 Same as @option{r} but for chroma planes.
15014 The default value is @var{0} and means automatic.
15019 Deinterlace video using neural network edge directed interpolation.
15021 This filter accepts the following options:
15025 Mandatory option, without binary file filter can not work.
15026 Currently file can be found here:
15027 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
15030 Set which frames to deinterlace, by default it is @code{all}.
15031 Can be @code{all} or @code{interlaced}.
15034 Set mode of operation.
15036 Can be one of the following:
15040 Use frame flags, both fields.
15042 Use frame flags, single field.
15044 Use top field only.
15046 Use bottom field only.
15048 Use both fields, top first.
15050 Use both fields, bottom first.
15054 Set which planes to process, by default filter process all frames.
15057 Set size of local neighborhood around each pixel, used by the predictor neural
15060 Can be one of the following:
15073 Set the number of neurons in predictor neural network.
15074 Can be one of the following:
15085 Controls the number of different neural network predictions that are blended
15086 together to compute the final output value. Can be @code{fast}, default or
15090 Set which set of weights to use in the predictor.
15091 Can be one of the following:
15095 weights trained to minimize absolute error
15097 weights trained to minimize squared error
15101 Controls whether or not the prescreener neural network is used to decide
15102 which pixels should be processed by the predictor neural network and which
15103 can be handled by simple cubic interpolation.
15104 The prescreener is trained to know whether cubic interpolation will be
15105 sufficient for a pixel or whether it should be predicted by the predictor nn.
15106 The computational complexity of the prescreener nn is much less than that of
15107 the predictor nn. Since most pixels can be handled by cubic interpolation,
15108 using the prescreener generally results in much faster processing.
15109 The prescreener is pretty accurate, so the difference between using it and not
15110 using it is almost always unnoticeable.
15112 Can be one of the following:
15122 Default is @code{new}.
15125 @subsection Commands
15126 This filter supports same @ref{commands} as options, excluding @var{weights} option.
15130 Force libavfilter not to use any of the specified pixel formats for the
15131 input to the next filter.
15133 It accepts the following parameters:
15137 A '|'-separated list of pixel format names, such as
15138 pix_fmts=yuv420p|monow|rgb24".
15142 @subsection Examples
15146 Force libavfilter to use a format different from @var{yuv420p} for the
15147 input to the vflip filter:
15149 noformat=pix_fmts=yuv420p,vflip
15153 Convert the input video to any of the formats not contained in the list:
15155 noformat=yuv420p|yuv444p|yuv410p
15161 Add noise on video input frame.
15163 The filter accepts the following options:
15171 Set noise seed for specific pixel component or all pixel components in case
15172 of @var{all_seed}. Default value is @code{123457}.
15174 @item all_strength, alls
15175 @item c0_strength, c0s
15176 @item c1_strength, c1s
15177 @item c2_strength, c2s
15178 @item c3_strength, c3s
15179 Set noise strength for specific pixel component or all pixel components in case
15180 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
15182 @item all_flags, allf
15183 @item c0_flags, c0f
15184 @item c1_flags, c1f
15185 @item c2_flags, c2f
15186 @item c3_flags, c3f
15187 Set pixel component flags or set flags for all components if @var{all_flags}.
15188 Available values for component flags are:
15191 averaged temporal noise (smoother)
15193 mix random noise with a (semi)regular pattern
15195 temporal noise (noise pattern changes between frames)
15197 uniform noise (gaussian otherwise)
15201 @subsection Examples
15203 Add temporal and uniform noise to input video:
15205 noise=alls=20:allf=t+u
15210 Normalize RGB video (aka histogram stretching, contrast stretching).
15211 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
15213 For each channel of each frame, the filter computes the input range and maps
15214 it linearly to the user-specified output range. The output range defaults
15215 to the full dynamic range from pure black to pure white.
15217 Temporal smoothing can be used on the input range to reduce flickering (rapid
15218 changes in brightness) caused when small dark or bright objects enter or leave
15219 the scene. This is similar to the auto-exposure (automatic gain control) on a
15220 video camera, and, like a video camera, it may cause a period of over- or
15221 under-exposure of the video.
15223 The R,G,B channels can be normalized independently, which may cause some
15224 color shifting, or linked together as a single channel, which prevents
15225 color shifting. Linked normalization preserves hue. Independent normalization
15226 does not, so it can be used to remove some color casts. Independent and linked
15227 normalization can be combined in any ratio.
15229 The normalize filter accepts the following options:
15234 Colors which define the output range. The minimum input value is mapped to
15235 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
15236 The defaults are black and white respectively. Specifying white for
15237 @var{blackpt} and black for @var{whitept} will give color-inverted,
15238 normalized video. Shades of grey can be used to reduce the dynamic range
15239 (contrast). Specifying saturated colors here can create some interesting
15243 The number of previous frames to use for temporal smoothing. The input range
15244 of each channel is smoothed using a rolling average over the current frame
15245 and the @var{smoothing} previous frames. The default is 0 (no temporal
15249 Controls the ratio of independent (color shifting) channel normalization to
15250 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
15251 independent. Defaults to 1.0 (fully independent).
15254 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
15255 expensive no-op. Defaults to 1.0 (full strength).
15259 @subsection Commands
15260 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
15261 The command accepts the same syntax of the corresponding option.
15263 If the specified expression is not valid, it is kept at its current
15266 @subsection Examples
15268 Stretch video contrast to use the full dynamic range, with no temporal
15269 smoothing; may flicker depending on the source content:
15271 normalize=blackpt=black:whitept=white:smoothing=0
15274 As above, but with 50 frames of temporal smoothing; flicker should be
15275 reduced, depending on the source content:
15277 normalize=blackpt=black:whitept=white:smoothing=50
15280 As above, but with hue-preserving linked channel normalization:
15282 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
15285 As above, but with half strength:
15287 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
15290 Map the darkest input color to red, the brightest input color to cyan:
15292 normalize=blackpt=red:whitept=cyan
15297 Pass the video source unchanged to the output.
15300 Optical Character Recognition
15302 This filter uses Tesseract for optical character recognition. To enable
15303 compilation of this filter, you need to configure FFmpeg with
15304 @code{--enable-libtesseract}.
15306 It accepts the following options:
15310 Set datapath to tesseract data. Default is to use whatever was
15311 set at installation.
15314 Set language, default is "eng".
15317 Set character whitelist.
15320 Set character blacklist.
15323 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
15324 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
15328 Apply a video transform using libopencv.
15330 To enable this filter, install the libopencv library and headers and
15331 configure FFmpeg with @code{--enable-libopencv}.
15333 It accepts the following parameters:
15338 The name of the libopencv filter to apply.
15340 @item filter_params
15341 The parameters to pass to the libopencv filter. If not specified, the default
15342 values are assumed.
15346 Refer to the official libopencv documentation for more precise
15348 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
15350 Several libopencv filters are supported; see the following subsections.
15355 Dilate an image by using a specific structuring element.
15356 It corresponds to the libopencv function @code{cvDilate}.
15358 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
15360 @var{struct_el} represents a structuring element, and has the syntax:
15361 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
15363 @var{cols} and @var{rows} represent the number of columns and rows of
15364 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
15365 point, and @var{shape} the shape for the structuring element. @var{shape}
15366 must be "rect", "cross", "ellipse", or "custom".
15368 If the value for @var{shape} is "custom", it must be followed by a
15369 string of the form "=@var{filename}". The file with name
15370 @var{filename} is assumed to represent a binary image, with each
15371 printable character corresponding to a bright pixel. When a custom
15372 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
15373 or columns and rows of the read file are assumed instead.
15375 The default value for @var{struct_el} is "3x3+0x0/rect".
15377 @var{nb_iterations} specifies the number of times the transform is
15378 applied to the image, and defaults to 1.
15382 # Use the default values
15385 # Dilate using a structuring element with a 5x5 cross, iterating two times
15386 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
15388 # Read the shape from the file diamond.shape, iterating two times.
15389 # The file diamond.shape may contain a pattern of characters like this
15395 # The specified columns and rows are ignored
15396 # but the anchor point coordinates are not
15397 ocv=dilate:0x0+2x2/custom=diamond.shape|2
15402 Erode an image by using a specific structuring element.
15403 It corresponds to the libopencv function @code{cvErode}.
15405 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
15406 with the same syntax and semantics as the @ref{dilate} filter.
15410 Smooth the input video.
15412 The filter takes the following parameters:
15413 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
15415 @var{type} is the type of smooth filter to apply, and must be one of
15416 the following values: "blur", "blur_no_scale", "median", "gaussian",
15417 or "bilateral". The default value is "gaussian".
15419 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
15420 depends on the smooth type. @var{param1} and
15421 @var{param2} accept integer positive values or 0. @var{param3} and
15422 @var{param4} accept floating point values.
15424 The default value for @var{param1} is 3. The default value for the
15425 other parameters is 0.
15427 These parameters correspond to the parameters assigned to the
15428 libopencv function @code{cvSmooth}.
15430 @section oscilloscope
15432 2D Video Oscilloscope.
15434 Useful to measure spatial impulse, step responses, chroma delays, etc.
15436 It accepts the following parameters:
15440 Set scope center x position.
15443 Set scope center y position.
15446 Set scope size, relative to frame diagonal.
15449 Set scope tilt/rotation.
15455 Set trace center x position.
15458 Set trace center y position.
15461 Set trace width, relative to width of frame.
15464 Set trace height, relative to height of frame.
15467 Set which components to trace. By default it traces first three components.
15470 Draw trace grid. By default is enabled.
15473 Draw some statistics. By default is enabled.
15476 Draw scope. By default is enabled.
15479 @subsection Commands
15480 This filter supports same @ref{commands} as options.
15481 The command accepts the same syntax of the corresponding option.
15483 If the specified expression is not valid, it is kept at its current
15486 @subsection Examples
15490 Inspect full first row of video frame.
15492 oscilloscope=x=0.5:y=0:s=1
15496 Inspect full last row of video frame.
15498 oscilloscope=x=0.5:y=1:s=1
15502 Inspect full 5th line of video frame of height 1080.
15504 oscilloscope=x=0.5:y=5/1080:s=1
15508 Inspect full last column of video frame.
15510 oscilloscope=x=1:y=0.5:s=1:t=1
15518 Overlay one video on top of another.
15520 It takes two inputs and has one output. The first input is the "main"
15521 video on which the second input is overlaid.
15523 It accepts the following parameters:
15525 A description of the accepted options follows.
15530 Set the expression for the x and y coordinates of the overlaid video
15531 on the main video. Default value is "0" for both expressions. In case
15532 the expression is invalid, it is set to a huge value (meaning that the
15533 overlay will not be displayed within the output visible area).
15536 See @ref{framesync}.
15539 Set when the expressions for @option{x}, and @option{y} are evaluated.
15541 It accepts the following values:
15544 only evaluate expressions once during the filter initialization or
15545 when a command is processed
15548 evaluate expressions for each incoming frame
15551 Default value is @samp{frame}.
15554 See @ref{framesync}.
15557 Set the format for the output video.
15559 It accepts the following values:
15562 force YUV420 output
15565 force YUV420p10 output
15568 force YUV422 output
15571 force YUV422p10 output
15574 force YUV444 output
15577 force packed RGB output
15580 force planar RGB output
15583 automatically pick format
15586 Default value is @samp{yuv420}.
15589 See @ref{framesync}.
15592 Set format of alpha of the overlaid video, it can be @var{straight} or
15593 @var{premultiplied}. Default is @var{straight}.
15596 The @option{x}, and @option{y} expressions can contain the following
15602 The main input width and height.
15606 The overlay input width and height.
15610 The computed values for @var{x} and @var{y}. They are evaluated for
15615 horizontal and vertical chroma subsample values of the output
15616 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15620 the number of input frame, starting from 0
15623 the position in the file of the input frame, NAN if unknown
15626 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15630 This filter also supports the @ref{framesync} options.
15632 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15633 when evaluation is done @emph{per frame}, and will evaluate to NAN
15634 when @option{eval} is set to @samp{init}.
15636 Be aware that frames are taken from each input video in timestamp
15637 order, hence, if their initial timestamps differ, it is a good idea
15638 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15639 have them begin in the same zero timestamp, as the example for
15640 the @var{movie} filter does.
15642 You can chain together more overlays but you should test the
15643 efficiency of such approach.
15645 @subsection Commands
15647 This filter supports the following commands:
15651 Modify the x and y of the overlay input.
15652 The command accepts the same syntax of the corresponding option.
15654 If the specified expression is not valid, it is kept at its current
15658 @subsection Examples
15662 Draw the overlay at 10 pixels from the bottom right corner of the main
15665 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15668 Using named options the example above becomes:
15670 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15674 Insert a transparent PNG logo in the bottom left corner of the input,
15675 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15677 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15681 Insert 2 different transparent PNG logos (second logo on bottom
15682 right corner) using the @command{ffmpeg} tool:
15684 ffmpeg -i input -i logo1 -i logo2 -filter_complex 'overlay=x=10:y=H-h-10,overlay=x=W-w-10:y=H-h-10' output
15688 Add a transparent color layer on top of the main video; @code{WxH}
15689 must specify the size of the main input to the overlay filter:
15691 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15695 Play an original video and a filtered version (here with the deshake
15696 filter) side by side using the @command{ffplay} tool:
15698 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15701 The above command is the same as:
15703 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15707 Make a sliding overlay appearing from the left to the right top part of the
15708 screen starting since time 2:
15710 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15714 Compose output by putting two input videos side to side:
15716 ffmpeg -i left.avi -i right.avi -filter_complex "
15717 nullsrc=size=200x100 [background];
15718 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15719 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15720 [background][left] overlay=shortest=1 [background+left];
15721 [background+left][right] overlay=shortest=1:x=100 [left+right]
15726 Mask 10-20 seconds of a video by applying the delogo filter to a section
15728 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15729 -vf '[in]split[split_main][split_delogo];[split_delogo]trim=start=360:end=371,delogo=0:0:640:480[delogoed];[split_main][delogoed]overlay=eof_action=pass[out]'
15734 Chain several overlays in cascade:
15736 nullsrc=s=200x200 [bg];
15737 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15738 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15739 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15740 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15741 [in3] null, [mid2] overlay=100:100 [out0]
15746 @anchor{overlay_cuda}
15747 @section overlay_cuda
15749 Overlay one video on top of another.
15751 This is the CUDA variant of the @ref{overlay} filter.
15752 It only accepts CUDA frames. The underlying input pixel formats have to match.
15754 It takes two inputs and has one output. The first input is the "main"
15755 video on which the second input is overlaid.
15757 It accepts the following parameters:
15762 Set the x and y coordinates of the overlaid video on the main video.
15763 Default value is "0" for both expressions.
15766 See @ref{framesync}.
15769 See @ref{framesync}.
15772 See @ref{framesync}.
15776 This filter also supports the @ref{framesync} options.
15780 Apply Overcomplete Wavelet denoiser.
15782 The filter accepts the following options:
15788 Larger depth values will denoise lower frequency components more, but
15789 slow down filtering.
15791 Must be an int in the range 8-16, default is @code{8}.
15793 @item luma_strength, ls
15796 Must be a double value in the range 0-1000, default is @code{1.0}.
15798 @item chroma_strength, cs
15799 Set chroma strength.
15801 Must be a double value in the range 0-1000, default is @code{1.0}.
15807 Add paddings to the input image, and place the original input at the
15808 provided @var{x}, @var{y} coordinates.
15810 It accepts the following parameters:
15815 Specify an expression for the size of the output image with the
15816 paddings added. If the value for @var{width} or @var{height} is 0, the
15817 corresponding input size is used for the output.
15819 The @var{width} expression can reference the value set by the
15820 @var{height} expression, and vice versa.
15822 The default value of @var{width} and @var{height} is 0.
15826 Specify the offsets to place the input image at within the padded area,
15827 with respect to the top/left border of the output image.
15829 The @var{x} expression can reference the value set by the @var{y}
15830 expression, and vice versa.
15832 The default value of @var{x} and @var{y} is 0.
15834 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15835 so the input image is centered on the padded area.
15838 Specify the color of the padded area. For the syntax of this option,
15839 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15840 manual,ffmpeg-utils}.
15842 The default value of @var{color} is "black".
15845 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15847 It accepts the following values:
15851 Only evaluate expressions once during the filter initialization or when
15852 a command is processed.
15855 Evaluate expressions for each incoming frame.
15859 Default value is @samp{init}.
15862 Pad to aspect instead to a resolution.
15866 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15867 options are expressions containing the following constants:
15872 The input video width and height.
15876 These are the same as @var{in_w} and @var{in_h}.
15880 The output width and height (the size of the padded area), as
15881 specified by the @var{width} and @var{height} expressions.
15885 These are the same as @var{out_w} and @var{out_h}.
15889 The x and y offsets as specified by the @var{x} and @var{y}
15890 expressions, or NAN if not yet specified.
15893 same as @var{iw} / @var{ih}
15896 input sample aspect ratio
15899 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15903 The horizontal and vertical chroma subsample values. For example for the
15904 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15907 @subsection Examples
15911 Add paddings with the color "violet" to the input video. The output video
15912 size is 640x480, and the top-left corner of the input video is placed at
15915 pad=640:480:0:40:violet
15918 The example above is equivalent to the following command:
15920 pad=width=640:height=480:x=0:y=40:color=violet
15924 Pad the input to get an output with dimensions increased by 3/2,
15925 and put the input video at the center of the padded area:
15927 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15931 Pad the input to get a squared output with size equal to the maximum
15932 value between the input width and height, and put the input video at
15933 the center of the padded area:
15935 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15939 Pad the input to get a final w/h ratio of 16:9:
15941 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15945 In case of anamorphic video, in order to set the output display aspect
15946 correctly, it is necessary to use @var{sar} in the expression,
15947 according to the relation:
15949 (ih * X / ih) * sar = output_dar
15950 X = output_dar / sar
15953 Thus the previous example needs to be modified to:
15955 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15959 Double the output size and put the input video in the bottom-right
15960 corner of the output padded area:
15962 pad="2*iw:2*ih:ow-iw:oh-ih"
15966 @anchor{palettegen}
15967 @section palettegen
15969 Generate one palette for a whole video stream.
15971 It accepts the following options:
15975 Set the maximum number of colors to quantize in the palette.
15976 Note: the palette will still contain 256 colors; the unused palette entries
15979 @item reserve_transparent
15980 Create a palette of 255 colors maximum and reserve the last one for
15981 transparency. Reserving the transparency color is useful for GIF optimization.
15982 If not set, the maximum of colors in the palette will be 256. You probably want
15983 to disable this option for a standalone image.
15986 @item transparency_color
15987 Set the color that will be used as background for transparency.
15990 Set statistics mode.
15992 It accepts the following values:
15995 Compute full frame histograms.
15997 Compute histograms only for the part that differs from previous frame. This
15998 might be relevant to give more importance to the moving part of your input if
15999 the background is static.
16001 Compute new histogram for each frame.
16004 Default value is @var{full}.
16007 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
16008 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
16009 color quantization of the palette. This information is also visible at
16010 @var{info} logging level.
16012 @subsection Examples
16016 Generate a representative palette of a given video using @command{ffmpeg}:
16018 ffmpeg -i input.mkv -vf palettegen palette.png
16022 @section paletteuse
16024 Use a palette to downsample an input video stream.
16026 The filter takes two inputs: one video stream and a palette. The palette must
16027 be a 256 pixels image.
16029 It accepts the following options:
16033 Select dithering mode. Available algorithms are:
16036 Ordered 8x8 bayer dithering (deterministic)
16038 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
16039 Note: this dithering is sometimes considered "wrong" and is included as a
16041 @item floyd_steinberg
16042 Floyd and Steingberg dithering (error diffusion)
16044 Frankie Sierra dithering v2 (error diffusion)
16046 Frankie Sierra dithering v2 "Lite" (error diffusion)
16049 Default is @var{sierra2_4a}.
16052 When @var{bayer} dithering is selected, this option defines the scale of the
16053 pattern (how much the crosshatch pattern is visible). A low value means more
16054 visible pattern for less banding, and higher value means less visible pattern
16055 at the cost of more banding.
16057 The option must be an integer value in the range [0,5]. Default is @var{2}.
16060 If set, define the zone to process
16064 Only the changing rectangle will be reprocessed. This is similar to GIF
16065 cropping/offsetting compression mechanism. This option can be useful for speed
16066 if only a part of the image is changing, and has use cases such as limiting the
16067 scope of the error diffusal @option{dither} to the rectangle that bounds the
16068 moving scene (it leads to more deterministic output if the scene doesn't change
16069 much, and as a result less moving noise and better GIF compression).
16072 Default is @var{none}.
16075 Take new palette for each output frame.
16077 @item alpha_threshold
16078 Sets the alpha threshold for transparency. Alpha values above this threshold
16079 will be treated as completely opaque, and values below this threshold will be
16080 treated as completely transparent.
16082 The option must be an integer value in the range [0,255]. Default is @var{128}.
16085 @subsection Examples
16089 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
16090 using @command{ffmpeg}:
16092 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
16096 @section perspective
16098 Correct perspective of video not recorded perpendicular to the screen.
16100 A description of the accepted parameters follows.
16111 Set coordinates expression for top left, top right, bottom left and bottom right corners.
16112 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
16113 If the @code{sense} option is set to @code{source}, then the specified points will be sent
16114 to the corners of the destination. If the @code{sense} option is set to @code{destination},
16115 then the corners of the source will be sent to the specified coordinates.
16117 The expressions can use the following variables:
16122 the width and height of video frame.
16126 Output frame count.
16129 @item interpolation
16130 Set interpolation for perspective correction.
16132 It accepts the following values:
16138 Default value is @samp{linear}.
16141 Set interpretation of coordinate options.
16143 It accepts the following values:
16147 Send point in the source specified by the given coordinates to
16148 the corners of the destination.
16150 @item 1, destination
16152 Send the corners of the source to the point in the destination specified
16153 by the given coordinates.
16155 Default value is @samp{source}.
16159 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
16161 It accepts the following values:
16164 only evaluate expressions once during the filter initialization or
16165 when a command is processed
16168 evaluate expressions for each incoming frame
16171 Default value is @samp{init}.
16176 Delay interlaced video by one field time so that the field order changes.
16178 The intended use is to fix PAL movies that have been captured with the
16179 opposite field order to the film-to-video transfer.
16181 A description of the accepted parameters follows.
16187 It accepts the following values:
16190 Capture field order top-first, transfer bottom-first.
16191 Filter will delay the bottom field.
16194 Capture field order bottom-first, transfer top-first.
16195 Filter will delay the top field.
16198 Capture and transfer with the same field order. This mode only exists
16199 for the documentation of the other options to refer to, but if you
16200 actually select it, the filter will faithfully do nothing.
16203 Capture field order determined automatically by field flags, transfer
16205 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
16206 basis using field flags. If no field information is available,
16207 then this works just like @samp{u}.
16210 Capture unknown or varying, transfer opposite.
16211 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
16212 analyzing the images and selecting the alternative that produces best
16213 match between the fields.
16216 Capture top-first, transfer unknown or varying.
16217 Filter selects among @samp{t} and @samp{p} using image analysis.
16220 Capture bottom-first, transfer unknown or varying.
16221 Filter selects among @samp{b} and @samp{p} using image analysis.
16224 Capture determined by field flags, transfer unknown or varying.
16225 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
16226 image analysis. If no field information is available, then this works just
16227 like @samp{U}. This is the default mode.
16230 Both capture and transfer unknown or varying.
16231 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
16235 @subsection Commands
16237 This filter supports the all above options as @ref{commands}.
16239 @section photosensitivity
16240 Reduce various flashes in video, so to help users with epilepsy.
16242 It accepts the following options:
16245 Set how many frames to use when filtering. Default is 30.
16248 Set detection threshold factor. Default is 1.
16252 Set how many pixels to skip when sampling frames. Default is 1.
16253 Allowed range is from 1 to 1024.
16256 Leave frames unchanged. Default is disabled.
16259 @section pixdesctest
16261 Pixel format descriptor test filter, mainly useful for internal
16262 testing. The output video should be equal to the input video.
16266 format=monow, pixdesctest
16269 can be used to test the monowhite pixel format descriptor definition.
16273 Display sample values of color channels. Mainly useful for checking color
16274 and levels. Minimum supported resolution is 640x480.
16276 The filters accept the following options:
16280 Set scope X position, relative offset on X axis.
16283 Set scope Y position, relative offset on Y axis.
16292 Set window opacity. This window also holds statistics about pixel area.
16295 Set window X position, relative offset on X axis.
16298 Set window Y position, relative offset on Y axis.
16301 @subsection Commands
16303 This filter supports same @ref{commands} as options.
16307 Enable the specified chain of postprocessing subfilters using libpostproc. This
16308 library should be automatically selected with a GPL build (@code{--enable-gpl}).
16309 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
16310 Each subfilter and some options have a short and a long name that can be used
16311 interchangeably, i.e. dr/dering are the same.
16313 The filters accept the following options:
16317 Set postprocessing subfilters string.
16320 All subfilters share common options to determine their scope:
16324 Honor the quality commands for this subfilter.
16327 Do chrominance filtering, too (default).
16330 Do luminance filtering only (no chrominance).
16333 Do chrominance filtering only (no luminance).
16336 These options can be appended after the subfilter name, separated by a '|'.
16338 Available subfilters are:
16341 @item hb/hdeblock[|difference[|flatness]]
16342 Horizontal deblocking filter
16345 Difference factor where higher values mean more deblocking (default: @code{32}).
16347 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16350 @item vb/vdeblock[|difference[|flatness]]
16351 Vertical deblocking filter
16354 Difference factor where higher values mean more deblocking (default: @code{32}).
16356 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16359 @item ha/hadeblock[|difference[|flatness]]
16360 Accurate horizontal deblocking filter
16363 Difference factor where higher values mean more deblocking (default: @code{32}).
16365 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16368 @item va/vadeblock[|difference[|flatness]]
16369 Accurate vertical deblocking filter
16372 Difference factor where higher values mean more deblocking (default: @code{32}).
16374 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16378 The horizontal and vertical deblocking filters share the difference and
16379 flatness values so you cannot set different horizontal and vertical
16383 @item h1/x1hdeblock
16384 Experimental horizontal deblocking filter
16386 @item v1/x1vdeblock
16387 Experimental vertical deblocking filter
16392 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
16395 larger -> stronger filtering
16397 larger -> stronger filtering
16399 larger -> stronger filtering
16402 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
16405 Stretch luminance to @code{0-255}.
16408 @item lb/linblenddeint
16409 Linear blend deinterlacing filter that deinterlaces the given block by
16410 filtering all lines with a @code{(1 2 1)} filter.
16412 @item li/linipoldeint
16413 Linear interpolating deinterlacing filter that deinterlaces the given block by
16414 linearly interpolating every second line.
16416 @item ci/cubicipoldeint
16417 Cubic interpolating deinterlacing filter deinterlaces the given block by
16418 cubically interpolating every second line.
16420 @item md/mediandeint
16421 Median deinterlacing filter that deinterlaces the given block by applying a
16422 median filter to every second line.
16424 @item fd/ffmpegdeint
16425 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
16426 second line with a @code{(-1 4 2 4 -1)} filter.
16429 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
16430 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
16432 @item fq/forceQuant[|quantizer]
16433 Overrides the quantizer table from the input with the constant quantizer you
16441 Default pp filter combination (@code{hb|a,vb|a,dr|a})
16444 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
16447 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
16450 @subsection Examples
16454 Apply horizontal and vertical deblocking, deringing and automatic
16455 brightness/contrast:
16461 Apply default filters without brightness/contrast correction:
16467 Apply default filters and temporal denoiser:
16469 pp=default/tmpnoise|1|2|3
16473 Apply deblocking on luminance only, and switch vertical deblocking on or off
16474 automatically depending on available CPU time:
16481 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
16482 similar to spp = 6 with 7 point DCT, where only the center sample is
16485 The filter accepts the following options:
16489 Force a constant quantization parameter. It accepts an integer in range
16490 0 to 63. If not set, the filter will use the QP from the video stream
16494 Set thresholding mode. Available modes are:
16498 Set hard thresholding.
16500 Set soft thresholding (better de-ringing effect, but likely blurrier).
16502 Set medium thresholding (good results, default).
16506 @section premultiply
16507 Apply alpha premultiply effect to input video stream using first plane
16508 of second stream as alpha.
16510 Both streams must have same dimensions and same pixel format.
16512 The filter accepts the following option:
16516 Set which planes will be processed, unprocessed planes will be copied.
16517 By default value 0xf, all planes will be processed.
16520 Do not require 2nd input for processing, instead use alpha plane from input stream.
16524 Apply prewitt operator to input video stream.
16526 The filter accepts the following option:
16530 Set which planes will be processed, unprocessed planes will be copied.
16531 By default value 0xf, all planes will be processed.
16534 Set value which will be multiplied with filtered result.
16537 Set value which will be added to filtered result.
16540 @subsection Commands
16542 This filter supports the all above options as @ref{commands}.
16544 @section pseudocolor
16546 Alter frame colors in video with pseudocolors.
16548 This filter accepts the following options:
16552 set pixel first component expression
16555 set pixel second component expression
16558 set pixel third component expression
16561 set pixel fourth component expression, corresponds to the alpha component
16564 set component to use as base for altering colors
16567 Pick one of built-in LUTs. By default is set to none.
16584 Set opacity of output colors. Allowed range is from 0 to 1.
16585 Default value is set to 1.
16588 Each of the expression options specifies the expression to use for computing
16589 the lookup table for the corresponding pixel component values.
16591 The expressions can contain the following constants and functions:
16596 The input width and height.
16599 The input value for the pixel component.
16601 @item ymin, umin, vmin, amin
16602 The minimum allowed component value.
16604 @item ymax, umax, vmax, amax
16605 The maximum allowed component value.
16608 All expressions default to "val".
16610 @subsection Commands
16612 This filter supports the all above options as @ref{commands}.
16614 @subsection Examples
16618 Change too high luma values to gradient:
16620 pseudocolor="'if(between(val,ymax,amax),lerp(ymin,ymax,(val-ymax)/(amax-ymax)),-1):if(between(val,ymax,amax),lerp(umax,umin,(val-ymax)/(amax-ymax)),-1):if(between(val,ymax,amax),lerp(vmin,vmax,(val-ymax)/(amax-ymax)),-1):-1'"
16626 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16627 Ratio) between two input videos.
16629 This filter takes in input two input videos, the first input is
16630 considered the "main" source and is passed unchanged to the
16631 output. The second input is used as a "reference" video for computing
16634 Both video inputs must have the same resolution and pixel format for
16635 this filter to work correctly. Also it assumes that both inputs
16636 have the same number of frames, which are compared one by one.
16638 The obtained average PSNR is printed through the logging system.
16640 The filter stores the accumulated MSE (mean squared error) of each
16641 frame, and at the end of the processing it is averaged across all frames
16642 equally, and the following formula is applied to obtain the PSNR:
16645 PSNR = 10*log10(MAX^2/MSE)
16648 Where MAX is the average of the maximum values of each component of the
16651 The description of the accepted parameters follows.
16654 @item stats_file, f
16655 If specified the filter will use the named file to save the PSNR of
16656 each individual frame. When filename equals "-" the data is sent to
16659 @item stats_version
16660 Specifies which version of the stats file format to use. Details of
16661 each format are written below.
16662 Default value is 1.
16664 @item stats_add_max
16665 Determines whether the max value is output to the stats log.
16666 Default value is 0.
16667 Requires stats_version >= 2. If this is set and stats_version < 2,
16668 the filter will return an error.
16671 This filter also supports the @ref{framesync} options.
16673 The file printed if @var{stats_file} is selected, contains a sequence of
16674 key/value pairs of the form @var{key}:@var{value} for each compared
16677 If a @var{stats_version} greater than 1 is specified, a header line precedes
16678 the list of per-frame-pair stats, with key value pairs following the frame
16679 format with the following parameters:
16682 @item psnr_log_version
16683 The version of the log file format. Will match @var{stats_version}.
16686 A comma separated list of the per-frame-pair parameters included in
16690 A description of each shown per-frame-pair parameter follows:
16694 sequential number of the input frame, starting from 1
16697 Mean Square Error pixel-by-pixel average difference of the compared
16698 frames, averaged over all the image components.
16700 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16701 Mean Square Error pixel-by-pixel average difference of the compared
16702 frames for the component specified by the suffix.
16704 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16705 Peak Signal to Noise ratio of the compared frames for the component
16706 specified by the suffix.
16708 @item max_avg, max_y, max_u, max_v
16709 Maximum allowed value for each channel, and average over all
16713 @subsection Examples
16718 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16719 [main][ref] psnr="stats_file=stats.log" [out]
16722 On this example the input file being processed is compared with the
16723 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16724 is stored in @file{stats.log}.
16727 Another example with different containers:
16729 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]psnr" -f null -
16736 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16737 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16740 The pullup filter is designed to take advantage of future context in making
16741 its decisions. This filter is stateless in the sense that it does not lock
16742 onto a pattern to follow, but it instead looks forward to the following
16743 fields in order to identify matches and rebuild progressive frames.
16745 To produce content with an even framerate, insert the fps filter after
16746 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16747 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16749 The filter accepts the following options:
16756 These options set the amount of "junk" to ignore at the left, right, top, and
16757 bottom of the image, respectively. Left and right are in units of 8 pixels,
16758 while top and bottom are in units of 2 lines.
16759 The default is 8 pixels on each side.
16762 Set the strict breaks. Setting this option to 1 will reduce the chances of
16763 filter generating an occasional mismatched frame, but it may also cause an
16764 excessive number of frames to be dropped during high motion sequences.
16765 Conversely, setting it to -1 will make filter match fields more easily.
16766 This may help processing of video where there is slight blurring between
16767 the fields, but may also cause there to be interlaced frames in the output.
16768 Default value is @code{0}.
16771 Set the metric plane to use. It accepts the following values:
16777 Use chroma blue plane.
16780 Use chroma red plane.
16783 This option may be set to use chroma plane instead of the default luma plane
16784 for doing filter's computations. This may improve accuracy on very clean
16785 source material, but more likely will decrease accuracy, especially if there
16786 is chroma noise (rainbow effect) or any grayscale video.
16787 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16788 load and make pullup usable in realtime on slow machines.
16791 For best results (without duplicated frames in the output file) it is
16792 necessary to change the output frame rate. For example, to inverse
16793 telecine NTSC input:
16795 ffmpeg -i input -vf pullup -r 24000/1001 ...
16800 Change video quantization parameters (QP).
16802 The filter accepts the following option:
16806 Set expression for quantization parameter.
16809 The expression is evaluated through the eval API and can contain, among others,
16810 the following constants:
16814 1 if index is not 129, 0 otherwise.
16817 Sequential index starting from -129 to 128.
16820 @subsection Examples
16824 Some equation like:
16832 Flush video frames from internal cache of frames into a random order.
16833 No frame is discarded.
16834 Inspired by @ref{frei0r} nervous filter.
16838 Set size in number of frames of internal cache, in range from @code{2} to
16839 @code{512}. Default is @code{30}.
16842 Set seed for random number generator, must be an integer included between
16843 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16844 less than @code{0}, the filter will try to use a good random seed on a
16848 @section readeia608
16850 Read closed captioning (EIA-608) information from the top lines of a video frame.
16852 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16853 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16854 with EIA-608 data (starting from 0). A description of each metadata value follows:
16857 @item lavfi.readeia608.X.cc
16858 The two bytes stored as EIA-608 data (printed in hexadecimal).
16860 @item lavfi.readeia608.X.line
16861 The number of the line on which the EIA-608 data was identified and read.
16864 This filter accepts the following options:
16868 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16871 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16874 Set the ratio of width reserved for sync code detection.
16875 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16878 Enable checking the parity bit. In the event of a parity error, the filter will output
16879 @code{0x00} for that character. Default is false.
16882 Lowpass lines prior to further processing. Default is enabled.
16885 @subsection Commands
16887 This filter supports the all above options as @ref{commands}.
16889 @subsection Examples
16893 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16895 ffprobe -f lavfi -i movie=captioned_video.mov,readeia608 -show_entries frame=pkt_pts_time:frame_tags=lavfi.readeia608.0.cc,lavfi.readeia608.1.cc -of csv
16901 Read vertical interval timecode (VITC) information from the top lines of a
16904 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16905 timecode value, if a valid timecode has been detected. Further metadata key
16906 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16907 timecode data has been found or not.
16909 This filter accepts the following options:
16913 Set the maximum number of lines to scan for VITC data. If the value is set to
16914 @code{-1} the full video frame is scanned. Default is @code{45}.
16917 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16918 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16921 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16922 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16925 @subsection Examples
16929 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16930 draw @code{--:--:--:--} as a placeholder:
16932 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16938 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16940 Destination pixel at position (X, Y) will be picked from source (x, y) position
16941 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16942 value for pixel will be used for destination pixel.
16944 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16945 will have Xmap/Ymap video stream dimensions.
16946 Xmap and Ymap input video streams are 16bit depth, single channel.
16950 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16951 Default is @code{color}.
16954 Specify the color of the unmapped pixels. For the syntax of this option,
16955 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16956 manual,ffmpeg-utils}. Default color is @code{black}.
16959 @section removegrain
16961 The removegrain filter is a spatial denoiser for progressive video.
16965 Set mode for the first plane.
16968 Set mode for the second plane.
16971 Set mode for the third plane.
16974 Set mode for the fourth plane.
16977 Range of mode is from 0 to 24. Description of each mode follows:
16981 Leave input plane unchanged. Default.
16984 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16987 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16990 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16993 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16994 This is equivalent to a median filter.
16997 Line-sensitive clipping giving the minimal change.
17000 Line-sensitive clipping, intermediate.
17003 Line-sensitive clipping, intermediate.
17006 Line-sensitive clipping, intermediate.
17009 Line-sensitive clipping on a line where the neighbours pixels are the closest.
17012 Replaces the target pixel with the closest neighbour.
17015 [1 2 1] horizontal and vertical kernel blur.
17021 Bob mode, interpolates top field from the line where the neighbours
17022 pixels are the closest.
17025 Bob mode, interpolates bottom field from the line where the neighbours
17026 pixels are the closest.
17029 Bob mode, interpolates top field. Same as 13 but with a more complicated
17030 interpolation formula.
17033 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
17034 interpolation formula.
17037 Clips the pixel with the minimum and maximum of respectively the maximum and
17038 minimum of each pair of opposite neighbour pixels.
17041 Line-sensitive clipping using opposite neighbours whose greatest distance from
17042 the current pixel is minimal.
17045 Replaces the pixel with the average of its 8 neighbours.
17048 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
17051 Clips pixels using the averages of opposite neighbour.
17054 Same as mode 21 but simpler and faster.
17057 Small edge and halo removal, but reputed useless.
17063 @section removelogo
17065 Suppress a TV station logo, using an image file to determine which
17066 pixels comprise the logo. It works by filling in the pixels that
17067 comprise the logo with neighboring pixels.
17069 The filter accepts the following options:
17073 Set the filter bitmap file, which can be any image format supported by
17074 libavformat. The width and height of the image file must match those of the
17075 video stream being processed.
17078 Pixels in the provided bitmap image with a value of zero are not
17079 considered part of the logo, non-zero pixels are considered part of
17080 the logo. If you use white (255) for the logo and black (0) for the
17081 rest, you will be safe. For making the filter bitmap, it is
17082 recommended to take a screen capture of a black frame with the logo
17083 visible, and then using a threshold filter followed by the erode
17084 filter once or twice.
17086 If needed, little splotches can be fixed manually. Remember that if
17087 logo pixels are not covered, the filter quality will be much
17088 reduced. Marking too many pixels as part of the logo does not hurt as
17089 much, but it will increase the amount of blurring needed to cover over
17090 the image and will destroy more information than necessary, and extra
17091 pixels will slow things down on a large logo.
17093 @section repeatfields
17095 This filter uses the repeat_field flag from the Video ES headers and hard repeats
17096 fields based on its value.
17100 Reverse a video clip.
17102 Warning: This filter requires memory to buffer the entire clip, so trimming
17105 @subsection Examples
17109 Take the first 5 seconds of a clip, and reverse it.
17116 Shift R/G/B/A pixels horizontally and/or vertically.
17118 The filter accepts the following options:
17121 Set amount to shift red horizontally.
17123 Set amount to shift red vertically.
17125 Set amount to shift green horizontally.
17127 Set amount to shift green vertically.
17129 Set amount to shift blue horizontally.
17131 Set amount to shift blue vertically.
17133 Set amount to shift alpha horizontally.
17135 Set amount to shift alpha vertically.
17137 Set edge mode, can be @var{smear}, default, or @var{warp}.
17140 @subsection Commands
17142 This filter supports the all above options as @ref{commands}.
17145 Apply roberts cross operator to input video stream.
17147 The filter accepts the following option:
17151 Set which planes will be processed, unprocessed planes will be copied.
17152 By default value 0xf, all planes will be processed.
17155 Set value which will be multiplied with filtered result.
17158 Set value which will be added to filtered result.
17161 @subsection Commands
17163 This filter supports the all above options as @ref{commands}.
17167 Rotate video by an arbitrary angle expressed in radians.
17169 The filter accepts the following options:
17171 A description of the optional parameters follows.
17174 Set an expression for the angle by which to rotate the input video
17175 clockwise, expressed as a number of radians. A negative value will
17176 result in a counter-clockwise rotation. By default it is set to "0".
17178 This expression is evaluated for each frame.
17181 Set the output width expression, default value is "iw".
17182 This expression is evaluated just once during configuration.
17185 Set the output height expression, default value is "ih".
17186 This expression is evaluated just once during configuration.
17189 Enable bilinear interpolation if set to 1, a value of 0 disables
17190 it. Default value is 1.
17193 Set the color used to fill the output area not covered by the rotated
17194 image. For the general syntax of this option, check the
17195 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
17196 If the special value "none" is selected then no
17197 background is printed (useful for example if the background is never shown).
17199 Default value is "black".
17202 The expressions for the angle and the output size can contain the
17203 following constants and functions:
17207 sequential number of the input frame, starting from 0. It is always NAN
17208 before the first frame is filtered.
17211 time in seconds of the input frame, it is set to 0 when the filter is
17212 configured. It is always NAN before the first frame is filtered.
17216 horizontal and vertical chroma subsample values. For example for the
17217 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17221 the input video width and height
17225 the output width and height, that is the size of the padded area as
17226 specified by the @var{width} and @var{height} expressions
17230 the minimal width/height required for completely containing the input
17231 video rotated by @var{a} radians.
17233 These are only available when computing the @option{out_w} and
17234 @option{out_h} expressions.
17237 @subsection Examples
17241 Rotate the input by PI/6 radians clockwise:
17247 Rotate the input by PI/6 radians counter-clockwise:
17253 Rotate the input by 45 degrees clockwise:
17259 Apply a constant rotation with period T, starting from an angle of PI/3:
17261 rotate=PI/3+2*PI*t/T
17265 Make the input video rotation oscillating with a period of T
17266 seconds and an amplitude of A radians:
17268 rotate=A*sin(2*PI/T*t)
17272 Rotate the video, output size is chosen so that the whole rotating
17273 input video is always completely contained in the output:
17275 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
17279 Rotate the video, reduce the output size so that no background is ever
17282 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
17286 @subsection Commands
17288 The filter supports the following commands:
17292 Set the angle expression.
17293 The command accepts the same syntax of the corresponding option.
17295 If the specified expression is not valid, it is kept at its current
17301 Apply Shape Adaptive Blur.
17303 The filter accepts the following options:
17306 @item luma_radius, lr
17307 Set luma blur filter strength, must be a value in range 0.1-4.0, default
17308 value is 1.0. A greater value will result in a more blurred image, and
17309 in slower processing.
17311 @item luma_pre_filter_radius, lpfr
17312 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
17315 @item luma_strength, ls
17316 Set luma maximum difference between pixels to still be considered, must
17317 be a value in the 0.1-100.0 range, default value is 1.0.
17319 @item chroma_radius, cr
17320 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
17321 greater value will result in a more blurred image, and in slower
17324 @item chroma_pre_filter_radius, cpfr
17325 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
17327 @item chroma_strength, cs
17328 Set chroma maximum difference between pixels to still be considered,
17329 must be a value in the -0.9-100.0 range.
17332 Each chroma option value, if not explicitly specified, is set to the
17333 corresponding luma option value.
17338 Scale (resize) the input video, using the libswscale library.
17340 The scale filter forces the output display aspect ratio to be the same
17341 of the input, by changing the output sample aspect ratio.
17343 If the input image format is different from the format requested by
17344 the next filter, the scale filter will convert the input to the
17347 @subsection Options
17348 The filter accepts the following options, or any of the options
17349 supported by the libswscale scaler.
17351 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
17352 the complete list of scaler options.
17357 Set the output video dimension expression. Default value is the input
17360 If the @var{width} or @var{w} value is 0, the input width is used for
17361 the output. If the @var{height} or @var{h} value is 0, the input height
17362 is used for the output.
17364 If one and only one of the values is -n with n >= 1, the scale filter
17365 will use a value that maintains the aspect ratio of the input image,
17366 calculated from the other specified dimension. After that it will,
17367 however, make sure that the calculated dimension is divisible by n and
17368 adjust the value if necessary.
17370 If both values are -n with n >= 1, the behavior will be identical to
17371 both values being set to 0 as previously detailed.
17373 See below for the list of accepted constants for use in the dimension
17377 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
17381 Only evaluate expressions once during the filter initialization or when a command is processed.
17384 Evaluate expressions for each incoming frame.
17388 Default value is @samp{init}.
17392 Set the interlacing mode. It accepts the following values:
17396 Force interlaced aware scaling.
17399 Do not apply interlaced scaling.
17402 Select interlaced aware scaling depending on whether the source frames
17403 are flagged as interlaced or not.
17406 Default value is @samp{0}.
17409 Set libswscale scaling flags. See
17410 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
17411 complete list of values. If not explicitly specified the filter applies
17415 @item param0, param1
17416 Set libswscale input parameters for scaling algorithms that need them. See
17417 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
17418 complete documentation. If not explicitly specified the filter applies
17424 Set the video size. For the syntax of this option, check the
17425 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17427 @item in_color_matrix
17428 @item out_color_matrix
17429 Set in/output YCbCr color space type.
17431 This allows the autodetected value to be overridden as well as allows forcing
17432 a specific value used for the output and encoder.
17434 If not specified, the color space type depends on the pixel format.
17440 Choose automatically.
17443 Format conforming to International Telecommunication Union (ITU)
17444 Recommendation BT.709.
17447 Set color space conforming to the United States Federal Communications
17448 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
17453 Set color space conforming to:
17457 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
17460 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
17463 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
17468 Set color space conforming to SMPTE ST 240:1999.
17471 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
17476 Set in/output YCbCr sample range.
17478 This allows the autodetected value to be overridden as well as allows forcing
17479 a specific value used for the output and encoder. If not specified, the
17480 range depends on the pixel format. Possible values:
17484 Choose automatically.
17487 Set full range (0-255 in case of 8-bit luma).
17489 @item mpeg/limited/tv
17490 Set "MPEG" range (16-235 in case of 8-bit luma).
17493 @item force_original_aspect_ratio
17494 Enable decreasing or increasing output video width or height if necessary to
17495 keep the original aspect ratio. Possible values:
17499 Scale the video as specified and disable this feature.
17502 The output video dimensions will automatically be decreased if needed.
17505 The output video dimensions will automatically be increased if needed.
17509 One useful instance of this option is that when you know a specific device's
17510 maximum allowed resolution, you can use this to limit the output video to
17511 that, while retaining the aspect ratio. For example, device A allows
17512 1280x720 playback, and your video is 1920x800. Using this option (set it to
17513 decrease) and specifying 1280x720 to the command line makes the output
17516 Please note that this is a different thing than specifying -1 for @option{w}
17517 or @option{h}, you still need to specify the output resolution for this option
17520 @item force_divisible_by
17521 Ensures that both the output dimensions, width and height, are divisible by the
17522 given integer when used together with @option{force_original_aspect_ratio}. This
17523 works similar to using @code{-n} in the @option{w} and @option{h} options.
17525 This option respects the value set for @option{force_original_aspect_ratio},
17526 increasing or decreasing the resolution accordingly. The video's aspect ratio
17527 may be slightly modified.
17529 This option can be handy if you need to have a video fit within or exceed
17530 a defined resolution using @option{force_original_aspect_ratio} but also have
17531 encoder restrictions on width or height divisibility.
17535 The values of the @option{w} and @option{h} options are expressions
17536 containing the following constants:
17541 The input width and height
17545 These are the same as @var{in_w} and @var{in_h}.
17549 The output (scaled) width and height
17553 These are the same as @var{out_w} and @var{out_h}
17556 The same as @var{iw} / @var{ih}
17559 input sample aspect ratio
17562 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17566 horizontal and vertical input chroma subsample values. For example for the
17567 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17571 horizontal and vertical output chroma subsample values. For example for the
17572 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17575 The (sequential) number of the input frame, starting from 0.
17576 Only available with @code{eval=frame}.
17579 The presentation timestamp of the input frame, expressed as a number of
17580 seconds. Only available with @code{eval=frame}.
17583 The position (byte offset) of the frame in the input stream, or NaN if
17584 this information is unavailable and/or meaningless (for example in case of synthetic video).
17585 Only available with @code{eval=frame}.
17588 @subsection Examples
17592 Scale the input video to a size of 200x100
17597 This is equivalent to:
17608 Specify a size abbreviation for the output size:
17613 which can also be written as:
17619 Scale the input to 2x:
17621 scale=w=2*iw:h=2*ih
17625 The above is the same as:
17627 scale=2*in_w:2*in_h
17631 Scale the input to 2x with forced interlaced scaling:
17633 scale=2*iw:2*ih:interl=1
17637 Scale the input to half size:
17639 scale=w=iw/2:h=ih/2
17643 Increase the width, and set the height to the same size:
17649 Seek Greek harmony:
17656 Increase the height, and set the width to 3/2 of the height:
17658 scale=w=3/2*oh:h=3/5*ih
17662 Increase the size, making the size a multiple of the chroma
17665 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17669 Increase the width to a maximum of 500 pixels,
17670 keeping the same aspect ratio as the input:
17672 scale=w='min(500\, iw*3/2):h=-1'
17676 Make pixels square by combining scale and setsar:
17678 scale='trunc(ih*dar):ih',setsar=1/1
17682 Make pixels square by combining scale and setsar,
17683 making sure the resulting resolution is even (required by some codecs):
17685 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17689 @subsection Commands
17691 This filter supports the following commands:
17695 Set the output video dimension expression.
17696 The command accepts the same syntax of the corresponding option.
17698 If the specified expression is not valid, it is kept at its current
17704 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17705 format conversion on CUDA video frames. Setting the output width and height
17706 works in the same way as for the @var{scale} filter.
17708 The following additional options are accepted:
17711 The pixel format of the output CUDA frames. If set to the string "same" (the
17712 default), the input format will be kept. Note that automatic format negotiation
17713 and conversion is not yet supported for hardware frames
17716 The interpolation algorithm used for resizing. One of the following:
17723 @item cubic2p_bspline
17724 2-parameter cubic (B=1, C=0)
17726 @item cubic2p_catmullrom
17727 2-parameter cubic (B=0, C=1/2)
17729 @item cubic2p_b05c03
17730 2-parameter cubic (B=1/2, C=3/10)
17738 @item force_original_aspect_ratio
17739 Enable decreasing or increasing output video width or height if necessary to
17740 keep the original aspect ratio. Possible values:
17744 Scale the video as specified and disable this feature.
17747 The output video dimensions will automatically be decreased if needed.
17750 The output video dimensions will automatically be increased if needed.
17754 One useful instance of this option is that when you know a specific device's
17755 maximum allowed resolution, you can use this to limit the output video to
17756 that, while retaining the aspect ratio. For example, device A allows
17757 1280x720 playback, and your video is 1920x800. Using this option (set it to
17758 decrease) and specifying 1280x720 to the command line makes the output
17761 Please note that this is a different thing than specifying -1 for @option{w}
17762 or @option{h}, you still need to specify the output resolution for this option
17765 @item force_divisible_by
17766 Ensures that both the output dimensions, width and height, are divisible by the
17767 given integer when used together with @option{force_original_aspect_ratio}. This
17768 works similar to using @code{-n} in the @option{w} and @option{h} options.
17770 This option respects the value set for @option{force_original_aspect_ratio},
17771 increasing or decreasing the resolution accordingly. The video's aspect ratio
17772 may be slightly modified.
17774 This option can be handy if you need to have a video fit within or exceed
17775 a defined resolution using @option{force_original_aspect_ratio} but also have
17776 encoder restrictions on width or height divisibility.
17782 Scale (resize) the input video, based on a reference video.
17784 See the scale filter for available options, scale2ref supports the same but
17785 uses the reference video instead of the main input as basis. scale2ref also
17786 supports the following additional constants for the @option{w} and
17787 @option{h} options:
17792 The main input video's width and height
17795 The same as @var{main_w} / @var{main_h}
17798 The main input video's sample aspect ratio
17800 @item main_dar, mdar
17801 The main input video's display aspect ratio. Calculated from
17802 @code{(main_w / main_h) * main_sar}.
17806 The main input video's horizontal and vertical chroma subsample values.
17807 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17811 The (sequential) number of the main input frame, starting from 0.
17812 Only available with @code{eval=frame}.
17815 The presentation timestamp of the main input frame, expressed as a number of
17816 seconds. Only available with @code{eval=frame}.
17819 The position (byte offset) of the frame in the main input stream, or NaN if
17820 this information is unavailable and/or meaningless (for example in case of synthetic video).
17821 Only available with @code{eval=frame}.
17824 @subsection Examples
17828 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17830 'scale2ref[b][a];[a][b]overlay'
17834 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17836 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17840 @subsection Commands
17842 This filter supports the following commands:
17846 Set the output video dimension expression.
17847 The command accepts the same syntax of the corresponding option.
17849 If the specified expression is not valid, it is kept at its current
17854 Scroll input video horizontally and/or vertically by constant speed.
17856 The filter accepts the following options:
17858 @item horizontal, h
17859 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17860 Negative values changes scrolling direction.
17863 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17864 Negative values changes scrolling direction.
17867 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17870 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17873 @subsection Commands
17875 This filter supports the following @ref{commands}:
17877 @item horizontal, h
17878 Set the horizontal scrolling speed.
17880 Set the vertical scrolling speed.
17886 Detect video scene change.
17888 This filter sets frame metadata with mafd between frame, the scene score, and
17889 forward the frame to the next filter, so they can use these metadata to detect
17890 scene change or others.
17892 In addition, this filter logs a message and sets frame metadata when it detects
17893 a scene change by @option{threshold}.
17895 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17897 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17898 to detect scene change.
17900 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17901 detect scene change with @option{threshold}.
17903 The filter accepts the following options:
17907 Set the scene change detection threshold as a percentage of maximum change. Good
17908 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17911 Default value is @code{10.}.
17914 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17915 You can enable it if you want to get snapshot of scene change frames only.
17918 @anchor{selectivecolor}
17919 @section selectivecolor
17921 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17922 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17923 by the "purity" of the color (that is, how saturated it already is).
17925 This filter is similar to the Adobe Photoshop Selective Color tool.
17927 The filter accepts the following options:
17930 @item correction_method
17931 Select color correction method.
17933 Available values are:
17936 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17939 Specified adjustments are relative to the original component value.
17941 Default is @code{absolute}.
17943 Adjustments for red pixels (pixels where the red component is the maximum)
17945 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17947 Adjustments for green pixels (pixels where the green component is the maximum)
17949 Adjustments for cyan pixels (pixels where the red component is the minimum)
17951 Adjustments for blue pixels (pixels where the blue component is the maximum)
17953 Adjustments for magenta pixels (pixels where the green component is the minimum)
17955 Adjustments for white pixels (pixels where all components are greater than 128)
17957 Adjustments for all pixels except pure black and pure white
17959 Adjustments for black pixels (pixels where all components are lesser than 128)
17961 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17964 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17965 4 space separated floating point adjustment values in the [-1,1] range,
17966 respectively to adjust the amount of cyan, magenta, yellow and black for the
17967 pixels of its range.
17969 @subsection Examples
17973 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17974 increase magenta by 27% in blue areas:
17976 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17980 Use a Photoshop selective color preset:
17982 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17986 @anchor{separatefields}
17987 @section separatefields
17989 The @code{separatefields} takes a frame-based video input and splits
17990 each frame into its components fields, producing a new half height clip
17991 with twice the frame rate and twice the frame count.
17993 This filter use field-dominance information in frame to decide which
17994 of each pair of fields to place first in the output.
17995 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17997 @section setdar, setsar
17999 The @code{setdar} filter sets the Display Aspect Ratio for the filter
18002 This is done by changing the specified Sample (aka Pixel) Aspect
18003 Ratio, according to the following equation:
18005 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
18008 Keep in mind that the @code{setdar} filter does not modify the pixel
18009 dimensions of the video frame. Also, the display aspect ratio set by
18010 this filter may be changed by later filters in the filterchain,
18011 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
18014 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
18015 the filter output video.
18017 Note that as a consequence of the application of this filter, the
18018 output display aspect ratio will change according to the equation
18021 Keep in mind that the sample aspect ratio set by the @code{setsar}
18022 filter may be changed by later filters in the filterchain, e.g. if
18023 another "setsar" or a "setdar" filter is applied.
18025 It accepts the following parameters:
18028 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
18029 Set the aspect ratio used by the filter.
18031 The parameter can be a floating point number string, an expression, or
18032 a string of the form @var{num}:@var{den}, where @var{num} and
18033 @var{den} are the numerator and denominator of the aspect ratio. If
18034 the parameter is not specified, it is assumed the value "0".
18035 In case the form "@var{num}:@var{den}" is used, the @code{:} character
18039 Set the maximum integer value to use for expressing numerator and
18040 denominator when reducing the expressed aspect ratio to a rational.
18041 Default value is @code{100}.
18045 The parameter @var{sar} is an expression containing
18046 the following constants:
18050 These are approximated values for the mathematical constants e
18051 (Euler's number), pi (Greek pi), and phi (the golden ratio).
18054 The input width and height.
18057 These are the same as @var{w} / @var{h}.
18060 The input sample aspect ratio.
18063 The input display aspect ratio. It is the same as
18064 (@var{w} / @var{h}) * @var{sar}.
18067 Horizontal and vertical chroma subsample values. For example, for the
18068 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
18071 @subsection Examples
18076 To change the display aspect ratio to 16:9, specify one of the following:
18083 To change the sample aspect ratio to 10:11, specify:
18089 To set a display aspect ratio of 16:9, and specify a maximum integer value of
18090 1000 in the aspect ratio reduction, use the command:
18092 setdar=ratio=16/9:max=1000
18100 Force field for the output video frame.
18102 The @code{setfield} filter marks the interlace type field for the
18103 output frames. It does not change the input frame, but only sets the
18104 corresponding property, which affects how the frame is treated by
18105 following filters (e.g. @code{fieldorder} or @code{yadif}).
18107 The filter accepts the following options:
18112 Available values are:
18116 Keep the same field property.
18119 Mark the frame as bottom-field-first.
18122 Mark the frame as top-field-first.
18125 Mark the frame as progressive.
18132 Force frame parameter for the output video frame.
18134 The @code{setparams} filter marks interlace and color range for the
18135 output frames. It does not change the input frame, but only sets the
18136 corresponding property, which affects how the frame is treated by
18141 Available values are:
18145 Keep the same field property (default).
18148 Mark the frame as bottom-field-first.
18151 Mark the frame as top-field-first.
18154 Mark the frame as progressive.
18158 Available values are:
18162 Keep the same color range property (default).
18164 @item unspecified, unknown
18165 Mark the frame as unspecified color range.
18167 @item limited, tv, mpeg
18168 Mark the frame as limited range.
18170 @item full, pc, jpeg
18171 Mark the frame as full range.
18174 @item color_primaries
18175 Set the color primaries.
18176 Available values are:
18180 Keep the same color primaries property (default).
18197 Set the color transfer.
18198 Available values are:
18202 Keep the same color trc property (default).
18224 Set the colorspace.
18225 Available values are:
18229 Keep the same colorspace property (default).
18242 @item chroma-derived-nc
18243 @item chroma-derived-c
18249 Apply shear transform to input video.
18251 This filter supports the following options:
18255 Shear factor in X-direction. Default value is 0.
18256 Allowed range is from -2 to 2.
18259 Shear factor in Y-direction. Default value is 0.
18260 Allowed range is from -2 to 2.
18263 Set the color used to fill the output area not covered by the transformed
18264 video. For the general syntax of this option, check the
18265 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18266 If the special value "none" is selected then no
18267 background is printed (useful for example if the background is never shown).
18269 Default value is "black".
18272 Set interpolation type. Can be @code{bilinear} or @code{nearest}. Default is @code{bilinear}.
18275 @subsection Commands
18277 This filter supports the all above options as @ref{commands}.
18281 Show a line containing various information for each input video frame.
18282 The input video is not modified.
18284 This filter supports the following options:
18288 Calculate checksums of each plane. By default enabled.
18291 The shown line contains a sequence of key/value pairs of the form
18292 @var{key}:@var{value}.
18294 The following values are shown in the output:
18298 The (sequential) number of the input frame, starting from 0.
18301 The Presentation TimeStamp of the input frame, expressed as a number of
18302 time base units. The time base unit depends on the filter input pad.
18305 The Presentation TimeStamp of the input frame, expressed as a number of
18309 The position of the frame in the input stream, or -1 if this information is
18310 unavailable and/or meaningless (for example in case of synthetic video).
18313 The pixel format name.
18316 The sample aspect ratio of the input frame, expressed in the form
18317 @var{num}/@var{den}.
18320 The size of the input frame. For the syntax of this option, check the
18321 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18324 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
18325 for bottom field first).
18328 This is 1 if the frame is a key frame, 0 otherwise.
18331 The picture type of the input frame ("I" for an I-frame, "P" for a
18332 P-frame, "B" for a B-frame, or "?" for an unknown type).
18333 Also refer to the documentation of the @code{AVPictureType} enum and of
18334 the @code{av_get_picture_type_char} function defined in
18335 @file{libavutil/avutil.h}.
18338 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
18340 @item plane_checksum
18341 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
18342 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
18345 The mean value of pixels in each plane of the input frame, expressed in the form
18346 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
18349 The standard deviation of pixel values in each plane of the input frame, expressed
18350 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
18354 @section showpalette
18356 Displays the 256 colors palette of each frame. This filter is only relevant for
18357 @var{pal8} pixel format frames.
18359 It accepts the following option:
18363 Set the size of the box used to represent one palette color entry. Default is
18364 @code{30} (for a @code{30x30} pixel box).
18367 @section shuffleframes
18369 Reorder and/or duplicate and/or drop video frames.
18371 It accepts the following parameters:
18375 Set the destination indexes of input frames.
18376 This is space or '|' separated list of indexes that maps input frames to output
18377 frames. Number of indexes also sets maximal value that each index may have.
18378 '-1' index have special meaning and that is to drop frame.
18381 The first frame has the index 0. The default is to keep the input unchanged.
18383 @subsection Examples
18387 Swap second and third frame of every three frames of the input:
18389 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
18393 Swap 10th and 1st frame of every ten frames of the input:
18395 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
18399 @section shufflepixels
18401 Reorder pixels in video frames.
18403 This filter accepts the following options:
18407 Set shuffle direction. Can be forward or inverse direction.
18408 Default direction is forward.
18411 Set shuffle mode. Can be horizontal, vertical or block mode.
18415 Set shuffle block_size. In case of horizontal shuffle mode only width
18416 part of size is used, and in case of vertical shuffle mode only height
18417 part of size is used.
18420 Set random seed used with shuffling pixels. Mainly useful to set to be able
18421 to reverse filtering process to get original input.
18422 For example, to reverse forward shuffle you need to use same parameters
18423 and exact same seed and to set direction to inverse.
18426 @section shuffleplanes
18428 Reorder and/or duplicate video planes.
18430 It accepts the following parameters:
18435 The index of the input plane to be used as the first output plane.
18438 The index of the input plane to be used as the second output plane.
18441 The index of the input plane to be used as the third output plane.
18444 The index of the input plane to be used as the fourth output plane.
18448 The first plane has the index 0. The default is to keep the input unchanged.
18450 @subsection Examples
18454 Swap the second and third planes of the input:
18456 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
18460 @anchor{signalstats}
18461 @section signalstats
18462 Evaluate various visual metrics that assist in determining issues associated
18463 with the digitization of analog video media.
18465 By default the filter will log these metadata values:
18469 Display the minimal Y value contained within the input frame. Expressed in
18473 Display the Y value at the 10% percentile within the input frame. Expressed in
18477 Display the average Y value within the input frame. Expressed in range of
18481 Display the Y value at the 90% percentile within the input frame. Expressed in
18485 Display the maximum Y value contained within the input frame. Expressed in
18489 Display the minimal U value contained within the input frame. Expressed in
18493 Display the U value at the 10% percentile within the input frame. Expressed in
18497 Display the average U value within the input frame. Expressed in range of
18501 Display the U value at the 90% percentile within the input frame. Expressed in
18505 Display the maximum U value contained within the input frame. Expressed in
18509 Display the minimal V value contained within the input frame. Expressed in
18513 Display the V value at the 10% percentile within the input frame. Expressed in
18517 Display the average V value within the input frame. Expressed in range of
18521 Display the V value at the 90% percentile within the input frame. Expressed in
18525 Display the maximum V value contained within the input frame. Expressed in
18529 Display the minimal saturation value contained within the input frame.
18530 Expressed in range of [0-~181.02].
18533 Display the saturation value at the 10% percentile within the input frame.
18534 Expressed in range of [0-~181.02].
18537 Display the average saturation value within the input frame. Expressed in range
18541 Display the saturation value at the 90% percentile within the input frame.
18542 Expressed in range of [0-~181.02].
18545 Display the maximum saturation value contained within the input frame.
18546 Expressed in range of [0-~181.02].
18549 Display the median value for hue within the input frame. Expressed in range of
18553 Display the average value for hue within the input frame. Expressed in range of
18557 Display the average of sample value difference between all values of the Y
18558 plane in the current frame and corresponding values of the previous input frame.
18559 Expressed in range of [0-255].
18562 Display the average of sample value difference between all values of the U
18563 plane in the current frame and corresponding values of the previous input frame.
18564 Expressed in range of [0-255].
18567 Display the average of sample value difference between all values of the V
18568 plane in the current frame and corresponding values of the previous input frame.
18569 Expressed in range of [0-255].
18572 Display bit depth of Y plane in current frame.
18573 Expressed in range of [0-16].
18576 Display bit depth of U plane in current frame.
18577 Expressed in range of [0-16].
18580 Display bit depth of V plane in current frame.
18581 Expressed in range of [0-16].
18584 The filter accepts the following options:
18590 @option{stat} specify an additional form of image analysis.
18591 @option{out} output video with the specified type of pixel highlighted.
18593 Both options accept the following values:
18597 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18598 unlike the neighboring pixels of the same field. Examples of temporal outliers
18599 include the results of video dropouts, head clogs, or tape tracking issues.
18602 Identify @var{vertical line repetition}. Vertical line repetition includes
18603 similar rows of pixels within a frame. In born-digital video vertical line
18604 repetition is common, but this pattern is uncommon in video digitized from an
18605 analog source. When it occurs in video that results from the digitization of an
18606 analog source it can indicate concealment from a dropout compensator.
18609 Identify pixels that fall outside of legal broadcast range.
18613 Set the highlight color for the @option{out} option. The default color is
18617 @subsection Examples
18621 Output data of various video metrics:
18623 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18627 Output specific data about the minimum and maximum values of the Y plane per frame:
18629 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18633 Playback video while highlighting pixels that are outside of broadcast range in red.
18635 ffplay example.mov -vf signalstats="out=brng:color=red"
18639 Playback video with signalstats metadata drawn over the frame.
18641 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18644 The contents of signalstat_drawtext.txt used in the command are:
18647 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18648 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18649 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18650 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18658 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18659 input. In this case the matching between the inputs can be calculated additionally.
18660 The filter always passes through the first input. The signature of each stream can
18661 be written into a file.
18663 It accepts the following options:
18667 Enable or disable the matching process.
18669 Available values are:
18673 Disable the calculation of a matching (default).
18675 Calculate the matching for the whole video and output whether the whole video
18676 matches or only parts.
18678 Calculate only until a matching is found or the video ends. Should be faster in
18683 Set the number of inputs. The option value must be a non negative integer.
18684 Default value is 1.
18687 Set the path to which the output is written. If there is more than one input,
18688 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18689 integer), that will be replaced with the input number. If no filename is
18690 specified, no output will be written. This is the default.
18693 Choose the output format.
18695 Available values are:
18699 Use the specified binary representation (default).
18701 Use the specified xml representation.
18705 Set threshold to detect one word as similar. The option value must be an integer
18706 greater than zero. The default value is 9000.
18709 Set threshold to detect all words as similar. The option value must be an integer
18710 greater than zero. The default value is 60000.
18713 Set threshold to detect frames as similar. The option value must be an integer
18714 greater than zero. The default value is 116.
18717 Set the minimum length of a sequence in frames to recognize it as matching
18718 sequence. The option value must be a non negative integer value.
18719 The default value is 0.
18722 Set the minimum relation, that matching frames to all frames must have.
18723 The option value must be a double value between 0 and 1. The default value is 0.5.
18726 @subsection Examples
18730 To calculate the signature of an input video and store it in signature.bin:
18732 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18736 To detect whether two videos match and store the signatures in XML format in
18737 signature0.xml and signature1.xml:
18739 ffmpeg -i input1.mkv -i input2.mkv -filter_complex "[0:v][1:v] signature=nb_inputs=2:detectmode=full:format=xml:filename=signature%d.xml" -map :v -f null -
18747 Blur the input video without impacting the outlines.
18749 It accepts the following options:
18752 @item luma_radius, lr
18753 Set the luma radius. The option value must be a float number in
18754 the range [0.1,5.0] that specifies the variance of the gaussian filter
18755 used to blur the image (slower if larger). Default value is 1.0.
18757 @item luma_strength, ls
18758 Set the luma strength. The option value must be a float number
18759 in the range [-1.0,1.0] that configures the blurring. A value included
18760 in [0.0,1.0] will blur the image whereas a value included in
18761 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18763 @item luma_threshold, lt
18764 Set the luma threshold used as a coefficient to determine
18765 whether a pixel should be blurred or not. The option value must be an
18766 integer in the range [-30,30]. A value of 0 will filter all the image,
18767 a value included in [0,30] will filter flat areas and a value included
18768 in [-30,0] will filter edges. Default value is 0.
18770 @item chroma_radius, cr
18771 Set the chroma radius. The option value must be a float number in
18772 the range [0.1,5.0] that specifies the variance of the gaussian filter
18773 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18775 @item chroma_strength, cs
18776 Set the chroma strength. The option value must be a float number
18777 in the range [-1.0,1.0] that configures the blurring. A value included
18778 in [0.0,1.0] will blur the image whereas a value included in
18779 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18781 @item chroma_threshold, ct
18782 Set the chroma threshold used as a coefficient to determine
18783 whether a pixel should be blurred or not. The option value must be an
18784 integer in the range [-30,30]. A value of 0 will filter all the image,
18785 a value included in [0,30] will filter flat areas and a value included
18786 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18789 If a chroma option is not explicitly set, the corresponding luma value
18793 Apply sobel operator to input video stream.
18795 The filter accepts the following option:
18799 Set which planes will be processed, unprocessed planes will be copied.
18800 By default value 0xf, all planes will be processed.
18803 Set value which will be multiplied with filtered result.
18806 Set value which will be added to filtered result.
18809 @subsection Commands
18811 This filter supports the all above options as @ref{commands}.
18816 Apply a simple postprocessing filter that compresses and decompresses the image
18817 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18818 and average the results.
18820 The filter accepts the following options:
18824 Set quality. This option defines the number of levels for averaging. It accepts
18825 an integer in the range 0-6. If set to @code{0}, the filter will have no
18826 effect. A value of @code{6} means the higher quality. For each increment of
18827 that value the speed drops by a factor of approximately 2. Default value is
18831 Force a constant quantization parameter. If not set, the filter will use the QP
18832 from the video stream (if available).
18835 Set thresholding mode. Available modes are:
18839 Set hard thresholding (default).
18841 Set soft thresholding (better de-ringing effect, but likely blurrier).
18844 @item use_bframe_qp
18845 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18846 option may cause flicker since the B-Frames have often larger QP. Default is
18847 @code{0} (not enabled).
18850 @subsection Commands
18852 This filter supports the following commands:
18854 @item quality, level
18855 Set quality level. The value @code{max} can be used to set the maximum level,
18856 currently @code{6}.
18862 Scale the input by applying one of the super-resolution methods based on
18863 convolutional neural networks. Supported models:
18867 Super-Resolution Convolutional Neural Network model (SRCNN).
18868 See @url{https://arxiv.org/abs/1501.00092}.
18871 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18872 See @url{https://arxiv.org/abs/1609.05158}.
18875 Training scripts as well as scripts for model file (.pb) saving can be found at
18876 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18877 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18879 Native model files (.model) can be generated from TensorFlow model
18880 files (.pb) by using tools/python/convert.py
18882 The filter accepts the following options:
18886 Specify which DNN backend to use for model loading and execution. This option accepts
18887 the following values:
18891 Native implementation of DNN loading and execution.
18894 TensorFlow backend. To enable this backend you
18895 need to install the TensorFlow for C library (see
18896 @url{https://www.tensorflow.org/install/lang_c}) and configure FFmpeg with
18897 @code{--enable-libtensorflow}
18900 Default value is @samp{native}.
18903 Set path to model file specifying network architecture and its parameters.
18904 Note that different backends use different file formats. TensorFlow backend
18905 can load files for both formats, while native backend can load files for only
18909 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18910 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18911 input upscaled using bicubic upscaling with proper scale factor.
18914 This feature can also be finished with @ref{dnn_processing} filter.
18918 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18920 This filter takes in input two input videos, the first input is
18921 considered the "main" source and is passed unchanged to the
18922 output. The second input is used as a "reference" video for computing
18925 Both video inputs must have the same resolution and pixel format for
18926 this filter to work correctly. Also it assumes that both inputs
18927 have the same number of frames, which are compared one by one.
18929 The filter stores the calculated SSIM of each frame.
18931 The description of the accepted parameters follows.
18934 @item stats_file, f
18935 If specified the filter will use the named file to save the SSIM of
18936 each individual frame. When filename equals "-" the data is sent to
18940 The file printed if @var{stats_file} is selected, contains a sequence of
18941 key/value pairs of the form @var{key}:@var{value} for each compared
18944 A description of each shown parameter follows:
18948 sequential number of the input frame, starting from 1
18950 @item Y, U, V, R, G, B
18951 SSIM of the compared frames for the component specified by the suffix.
18954 SSIM of the compared frames for the whole frame.
18957 Same as above but in dB representation.
18960 This filter also supports the @ref{framesync} options.
18962 @subsection Examples
18967 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18968 [main][ref] ssim="stats_file=stats.log" [out]
18971 On this example the input file being processed is compared with the
18972 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18973 is stored in @file{stats.log}.
18976 Another example with both psnr and ssim at same time:
18978 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18982 Another example with different containers:
18984 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]ssim" -f null -
18990 Convert between different stereoscopic image formats.
18992 The filters accept the following options:
18996 Set stereoscopic image format of input.
18998 Available values for input image formats are:
19001 side by side parallel (left eye left, right eye right)
19004 side by side crosseye (right eye left, left eye right)
19007 side by side parallel with half width resolution
19008 (left eye left, right eye right)
19011 side by side crosseye with half width resolution
19012 (right eye left, left eye right)
19016 above-below (left eye above, right eye below)
19020 above-below (right eye above, left eye below)
19024 above-below with half height resolution
19025 (left eye above, right eye below)
19029 above-below with half height resolution
19030 (right eye above, left eye below)
19033 alternating frames (left eye first, right eye second)
19036 alternating frames (right eye first, left eye second)
19039 interleaved rows (left eye has top row, right eye starts on next row)
19042 interleaved rows (right eye has top row, left eye starts on next row)
19045 interleaved columns, left eye first
19048 interleaved columns, right eye first
19050 Default value is @samp{sbsl}.
19054 Set stereoscopic image format of output.
19058 side by side parallel (left eye left, right eye right)
19061 side by side crosseye (right eye left, left eye right)
19064 side by side parallel with half width resolution
19065 (left eye left, right eye right)
19068 side by side crosseye with half width resolution
19069 (right eye left, left eye right)
19073 above-below (left eye above, right eye below)
19077 above-below (right eye above, left eye below)
19081 above-below with half height resolution
19082 (left eye above, right eye below)
19086 above-below with half height resolution
19087 (right eye above, left eye below)
19090 alternating frames (left eye first, right eye second)
19093 alternating frames (right eye first, left eye second)
19096 interleaved rows (left eye has top row, right eye starts on next row)
19099 interleaved rows (right eye has top row, left eye starts on next row)
19102 anaglyph red/blue gray
19103 (red filter on left eye, blue filter on right eye)
19106 anaglyph red/green gray
19107 (red filter on left eye, green filter on right eye)
19110 anaglyph red/cyan gray
19111 (red filter on left eye, cyan filter on right eye)
19114 anaglyph red/cyan half colored
19115 (red filter on left eye, cyan filter on right eye)
19118 anaglyph red/cyan color
19119 (red filter on left eye, cyan filter on right eye)
19122 anaglyph red/cyan color optimized with the least squares projection of dubois
19123 (red filter on left eye, cyan filter on right eye)
19126 anaglyph green/magenta gray
19127 (green filter on left eye, magenta filter on right eye)
19130 anaglyph green/magenta half colored
19131 (green filter on left eye, magenta filter on right eye)
19134 anaglyph green/magenta colored
19135 (green filter on left eye, magenta filter on right eye)
19138 anaglyph green/magenta color optimized with the least squares projection of dubois
19139 (green filter on left eye, magenta filter on right eye)
19142 anaglyph yellow/blue gray
19143 (yellow filter on left eye, blue filter on right eye)
19146 anaglyph yellow/blue half colored
19147 (yellow filter on left eye, blue filter on right eye)
19150 anaglyph yellow/blue colored
19151 (yellow filter on left eye, blue filter on right eye)
19154 anaglyph yellow/blue color optimized with the least squares projection of dubois
19155 (yellow filter on left eye, blue filter on right eye)
19158 mono output (left eye only)
19161 mono output (right eye only)
19164 checkerboard, left eye first
19167 checkerboard, right eye first
19170 interleaved columns, left eye first
19173 interleaved columns, right eye first
19179 Default value is @samp{arcd}.
19182 @subsection Examples
19186 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
19192 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
19198 @section streamselect, astreamselect
19199 Select video or audio streams.
19201 The filter accepts the following options:
19205 Set number of inputs. Default is 2.
19208 Set input indexes to remap to outputs.
19211 @subsection Commands
19213 The @code{streamselect} and @code{astreamselect} filter supports the following
19218 Set input indexes to remap to outputs.
19221 @subsection Examples
19225 Select first 5 seconds 1st stream and rest of time 2nd stream:
19227 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
19231 Same as above, but for audio:
19233 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
19240 Draw subtitles on top of input video using the libass library.
19242 To enable compilation of this filter you need to configure FFmpeg with
19243 @code{--enable-libass}. This filter also requires a build with libavcodec and
19244 libavformat to convert the passed subtitles file to ASS (Advanced Substation
19245 Alpha) subtitles format.
19247 The filter accepts the following options:
19251 Set the filename of the subtitle file to read. It must be specified.
19253 @item original_size
19254 Specify the size of the original video, the video for which the ASS file
19255 was composed. For the syntax of this option, check the
19256 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19257 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
19258 correctly scale the fonts if the aspect ratio has been changed.
19261 Set a directory path containing fonts that can be used by the filter.
19262 These fonts will be used in addition to whatever the font provider uses.
19265 Process alpha channel, by default alpha channel is untouched.
19268 Set subtitles input character encoding. @code{subtitles} filter only. Only
19269 useful if not UTF-8.
19271 @item stream_index, si
19272 Set subtitles stream index. @code{subtitles} filter only.
19275 Override default style or script info parameters of the subtitles. It accepts a
19276 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
19279 If the first key is not specified, it is assumed that the first value
19280 specifies the @option{filename}.
19282 For example, to render the file @file{sub.srt} on top of the input
19283 video, use the command:
19288 which is equivalent to:
19290 subtitles=filename=sub.srt
19293 To render the default subtitles stream from file @file{video.mkv}, use:
19295 subtitles=video.mkv
19298 To render the second subtitles stream from that file, use:
19300 subtitles=video.mkv:si=1
19303 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
19304 @code{DejaVu Serif}, use:
19306 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
19309 @section super2xsai
19311 Scale the input by 2x and smooth using the Super2xSaI (Scale and
19312 Interpolate) pixel art scaling algorithm.
19314 Useful for enlarging pixel art images without reducing sharpness.
19318 Swap two rectangular objects in video.
19320 This filter accepts the following options:
19330 Set 1st rect x coordinate.
19333 Set 1st rect y coordinate.
19336 Set 2nd rect x coordinate.
19339 Set 2nd rect y coordinate.
19341 All expressions are evaluated once for each frame.
19344 The all options are expressions containing the following constants:
19349 The input width and height.
19352 same as @var{w} / @var{h}
19355 input sample aspect ratio
19358 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
19361 The number of the input frame, starting from 0.
19364 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
19367 the position in the file of the input frame, NAN if unknown
19370 @subsection Commands
19372 This filter supports the all above options as @ref{commands}.
19378 Blend successive video frames.
19384 Apply telecine process to the video.
19386 This filter accepts the following options:
19395 The default value is @code{top}.
19399 A string of numbers representing the pulldown pattern you wish to apply.
19400 The default value is @code{23}.
19404 Some typical patterns:
19409 24p: 2332 (preferred)
19416 24p: 222222222223 ("Euro pulldown")
19421 @section thistogram
19423 Compute and draw a color distribution histogram for the input video across time.
19425 Unlike @ref{histogram} video filter which only shows histogram of single input frame
19426 at certain time, this filter shows also past histograms of number of frames defined
19427 by @code{width} option.
19429 The computed histogram is a representation of the color component
19430 distribution in an image.
19432 The filter accepts the following options:
19436 Set width of single color component output. Default value is @code{0}.
19437 Value of @code{0} means width will be picked from input video.
19438 This also set number of passed histograms to keep.
19439 Allowed range is [0, 8192].
19441 @item display_mode, d
19443 It accepts the following values:
19446 Per color component graphs are placed below each other.
19449 Per color component graphs are placed side by side.
19452 Presents information identical to that in the @code{parade}, except
19453 that the graphs representing color components are superimposed directly
19456 Default is @code{stack}.
19458 @item levels_mode, m
19459 Set mode. Can be either @code{linear}, or @code{logarithmic}.
19460 Default is @code{linear}.
19462 @item components, c
19463 Set what color components to display.
19464 Default is @code{7}.
19467 Set background opacity. Default is @code{0.9}.
19470 Show envelope. Default is disabled.
19473 Set envelope color. Default is @code{gold}.
19478 Available values for slide is:
19481 Draw new frame when right border is reached.
19484 Replace old columns with new ones.
19487 Scroll from right to left.
19490 Scroll from left to right.
19493 Draw single picture.
19496 Default is @code{replace}.
19501 Apply threshold effect to video stream.
19503 This filter needs four video streams to perform thresholding.
19504 First stream is stream we are filtering.
19505 Second stream is holding threshold values, third stream is holding min values,
19506 and last, fourth stream is holding max values.
19508 The filter accepts the following option:
19512 Set which planes will be processed, unprocessed planes will be copied.
19513 By default value 0xf, all planes will be processed.
19516 For example if first stream pixel's component value is less then threshold value
19517 of pixel component from 2nd threshold stream, third stream value will picked,
19518 otherwise fourth stream pixel component value will be picked.
19520 Using color source filter one can perform various types of thresholding:
19522 @subsection Examples
19526 Binary threshold, using gray color as threshold:
19528 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
19532 Inverted binary threshold, using gray color as threshold:
19534 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
19538 Truncate binary threshold, using gray color as threshold:
19540 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
19544 Threshold to zero, using gray color as threshold:
19546 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
19550 Inverted threshold to zero, using gray color as threshold:
19552 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
19557 Select the most representative frame in a given sequence of consecutive frames.
19559 The filter accepts the following options:
19563 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
19564 will pick one of them, and then handle the next batch of @var{n} frames until
19565 the end. Default is @code{100}.
19568 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
19569 value will result in a higher memory usage, so a high value is not recommended.
19571 @subsection Examples
19575 Extract one picture each 50 frames:
19581 Complete example of a thumbnail creation with @command{ffmpeg}:
19583 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
19590 Tile several successive frames together.
19592 The @ref{untile} filter can do the reverse.
19594 The filter accepts the following options:
19599 Set the grid size (i.e. the number of lines and columns). For the syntax of
19600 this option, check the
19601 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19604 Set the maximum number of frames to render in the given area. It must be less
19605 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19606 the area will be used.
19609 Set the outer border margin in pixels.
19612 Set the inner border thickness (i.e. the number of pixels between frames). For
19613 more advanced padding options (such as having different values for the edges),
19614 refer to the pad video filter.
19617 Specify the color of the unused area. For the syntax of this option, check the
19618 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19619 The default value of @var{color} is "black".
19622 Set the number of frames to overlap when tiling several successive frames together.
19623 The value must be between @code{0} and @var{nb_frames - 1}.
19626 Set the number of frames to initially be empty before displaying first output frame.
19627 This controls how soon will one get first output frame.
19628 The value must be between @code{0} and @var{nb_frames - 1}.
19631 @subsection Examples
19635 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19637 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19639 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19640 duplicating each output frame to accommodate the originally detected frame
19644 Display @code{5} pictures in an area of @code{3x2} frames,
19645 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19646 mixed flat and named options:
19648 tile=3x2:nb_frames=5:padding=7:margin=2
19652 @section tinterlace
19654 Perform various types of temporal field interlacing.
19656 Frames are counted starting from 1, so the first input frame is
19659 The filter accepts the following options:
19664 Specify the mode of the interlacing. This option can also be specified
19665 as a value alone. See below for a list of values for this option.
19667 Available values are:
19671 Move odd frames into the upper field, even into the lower field,
19672 generating a double height frame at half frame rate.
19676 Frame 1 Frame 2 Frame 3 Frame 4
19678 11111 22222 33333 44444
19679 11111 22222 33333 44444
19680 11111 22222 33333 44444
19681 11111 22222 33333 44444
19695 Only output odd frames, even frames are dropped, generating a frame with
19696 unchanged height at half frame rate.
19701 Frame 1 Frame 2 Frame 3 Frame 4
19703 11111 22222 33333 44444
19704 11111 22222 33333 44444
19705 11111 22222 33333 44444
19706 11111 22222 33333 44444
19716 Only output even frames, odd frames are dropped, generating a frame with
19717 unchanged height at half frame rate.
19722 Frame 1 Frame 2 Frame 3 Frame 4
19724 11111 22222 33333 44444
19725 11111 22222 33333 44444
19726 11111 22222 33333 44444
19727 11111 22222 33333 44444
19737 Expand each frame to full height, but pad alternate lines with black,
19738 generating a frame with double height at the same input frame rate.
19743 Frame 1 Frame 2 Frame 3 Frame 4
19745 11111 22222 33333 44444
19746 11111 22222 33333 44444
19747 11111 22222 33333 44444
19748 11111 22222 33333 44444
19751 11111 ..... 33333 .....
19752 ..... 22222 ..... 44444
19753 11111 ..... 33333 .....
19754 ..... 22222 ..... 44444
19755 11111 ..... 33333 .....
19756 ..... 22222 ..... 44444
19757 11111 ..... 33333 .....
19758 ..... 22222 ..... 44444
19762 @item interleave_top, 4
19763 Interleave the upper field from odd frames with the lower field from
19764 even frames, generating a frame with unchanged height at half frame rate.
19769 Frame 1 Frame 2 Frame 3 Frame 4
19771 11111<- 22222 33333<- 44444
19772 11111 22222<- 33333 44444<-
19773 11111<- 22222 33333<- 44444
19774 11111 22222<- 33333 44444<-
19784 @item interleave_bottom, 5
19785 Interleave the lower field from odd frames with the upper field from
19786 even frames, generating a frame with unchanged height at half frame rate.
19791 Frame 1 Frame 2 Frame 3 Frame 4
19793 11111 22222<- 33333 44444<-
19794 11111<- 22222 33333<- 44444
19795 11111 22222<- 33333 44444<-
19796 11111<- 22222 33333<- 44444
19806 @item interlacex2, 6
19807 Double frame rate with unchanged height. Frames are inserted each
19808 containing the second temporal field from the previous input frame and
19809 the first temporal field from the next input frame. This mode relies on
19810 the top_field_first flag. Useful for interlaced video displays with no
19811 field synchronisation.
19816 Frame 1 Frame 2 Frame 3 Frame 4
19818 11111 22222 33333 44444
19819 11111 22222 33333 44444
19820 11111 22222 33333 44444
19821 11111 22222 33333 44444
19824 11111 22222 22222 33333 33333 44444 44444
19825 11111 11111 22222 22222 33333 33333 44444
19826 11111 22222 22222 33333 33333 44444 44444
19827 11111 11111 22222 22222 33333 33333 44444
19832 Move odd frames into the upper field, even into the lower field,
19833 generating a double height frame at same frame rate.
19838 Frame 1 Frame 2 Frame 3 Frame 4
19840 11111 22222 33333 44444
19841 11111 22222 33333 44444
19842 11111 22222 33333 44444
19843 11111 22222 33333 44444
19846 11111 33333 33333 55555
19847 22222 22222 44444 44444
19848 11111 33333 33333 55555
19849 22222 22222 44444 44444
19850 11111 33333 33333 55555
19851 22222 22222 44444 44444
19852 11111 33333 33333 55555
19853 22222 22222 44444 44444
19858 Numeric values are deprecated but are accepted for backward
19859 compatibility reasons.
19861 Default mode is @code{merge}.
19864 Specify flags influencing the filter process.
19866 Available value for @var{flags} is:
19869 @item low_pass_filter, vlpf
19870 Enable linear vertical low-pass filtering in the filter.
19871 Vertical low-pass filtering is required when creating an interlaced
19872 destination from a progressive source which contains high-frequency
19873 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19876 @item complex_filter, cvlpf
19877 Enable complex vertical low-pass filtering.
19878 This will slightly less reduce interlace 'twitter' and Moire
19879 patterning but better retain detail and subjective sharpness impression.
19882 Bypass already interlaced frames, only adjust the frame rate.
19885 Vertical low-pass filtering and bypassing already interlaced frames can only be
19886 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19891 Pick median pixels from several successive input video frames.
19893 The filter accepts the following options:
19897 Set radius of median filter.
19898 Default is 1. Allowed range is from 1 to 127.
19901 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19904 Set median percentile. Default value is @code{0.5}.
19905 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19906 minimum values, and @code{1} maximum values.
19909 @subsection Commands
19911 This filter supports all above options as @ref{commands}, excluding option @code{radius}.
19913 @section tmidequalizer
19915 Apply Temporal Midway Video Equalization effect.
19917 Midway Video Equalization adjusts a sequence of video frames to have the same
19918 histograms, while maintaining their dynamics as much as possible. It's
19919 useful for e.g. matching exposures from a video frames sequence.
19921 This filter accepts the following option:
19925 Set filtering radius. Default is @code{5}. Allowed range is from 1 to 127.
19928 Set filtering sigma. Default is @code{0.5}. This controls strength of filtering.
19929 Setting this option to 0 effectively does nothing.
19932 Set which planes to process. Default is @code{15}, which is all available planes.
19937 Mix successive video frames.
19939 A description of the accepted options follows.
19943 The number of successive frames to mix. If unspecified, it defaults to 3.
19946 Specify weight of each input video frame.
19947 Each weight is separated by space. If number of weights is smaller than
19948 number of @var{frames} last specified weight will be used for all remaining
19952 Specify scale, if it is set it will be multiplied with sum
19953 of each weight multiplied with pixel values to give final destination
19954 pixel value. By default @var{scale} is auto scaled to sum of weights.
19957 @subsection Examples
19961 Average 7 successive frames:
19963 tmix=frames=7:weights="1 1 1 1 1 1 1"
19967 Apply simple temporal convolution:
19969 tmix=frames=3:weights="-1 3 -1"
19973 Similar as above but only showing temporal differences:
19975 tmix=frames=3:weights="-1 2 -1":scale=1
19979 @subsection Commands
19981 This filter supports the following commands:
19985 Syntax is same as option with same name.
19990 Tone map colors from different dynamic ranges.
19992 This filter expects data in single precision floating point, as it needs to
19993 operate on (and can output) out-of-range values. Another filter, such as
19994 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19996 The tonemapping algorithms implemented only work on linear light, so input
19997 data should be linearized beforehand (and possibly correctly tagged).
20000 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
20003 @subsection Options
20004 The filter accepts the following options.
20008 Set the tone map algorithm to use.
20010 Possible values are:
20013 Do not apply any tone map, only desaturate overbright pixels.
20016 Hard-clip any out-of-range values. Use it for perfect color accuracy for
20017 in-range values, while distorting out-of-range values.
20020 Stretch the entire reference gamut to a linear multiple of the display.
20023 Fit a logarithmic transfer between the tone curves.
20026 Preserve overall image brightness with a simple curve, using nonlinear
20027 contrast, which results in flattening details and degrading color accuracy.
20030 Preserve both dark and bright details better than @var{reinhard}, at the cost
20031 of slightly darkening everything. Use it when detail preservation is more
20032 important than color and brightness accuracy.
20035 Smoothly map out-of-range values, while retaining contrast and colors for
20036 in-range material as much as possible. Use it when color accuracy is more
20037 important than detail preservation.
20043 Tune the tone mapping algorithm.
20045 This affects the following algorithms:
20051 Specifies the scale factor to use while stretching.
20055 Specifies the exponent of the function.
20059 Specify an extra linear coefficient to multiply into the signal before clipping.
20063 Specify the local contrast coefficient at the display peak.
20064 Default to 0.5, which means that in-gamut values will be about half as bright
20071 Specify the transition point from linear to mobius transform. Every value
20072 below this point is guaranteed to be mapped 1:1. The higher the value, the
20073 more accurate the result will be, at the cost of losing bright details.
20074 Default to 0.3, which due to the steep initial slope still preserves in-range
20075 colors fairly accurately.
20079 Apply desaturation for highlights that exceed this level of brightness. The
20080 higher the parameter, the more color information will be preserved. This
20081 setting helps prevent unnaturally blown-out colors for super-highlights, by
20082 (smoothly) turning into white instead. This makes images feel more natural,
20083 at the cost of reducing information about out-of-range colors.
20085 The default of 2.0 is somewhat conservative and will mostly just apply to
20086 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
20088 This option works only if the input frame has a supported color tag.
20091 Override signal/nominal/reference peak with this value. Useful when the
20092 embedded peak information in display metadata is not reliable or when tone
20093 mapping from a lower range to a higher range.
20098 Temporarily pad video frames.
20100 The filter accepts the following options:
20104 Specify number of delay frames before input video stream. Default is 0.
20107 Specify number of padding frames after input video stream.
20108 Set to -1 to pad indefinitely. Default is 0.
20111 Set kind of frames added to beginning of stream.
20112 Can be either @var{add} or @var{clone}.
20113 With @var{add} frames of solid-color are added.
20114 With @var{clone} frames are clones of first frame.
20115 Default is @var{add}.
20118 Set kind of frames added to end of stream.
20119 Can be either @var{add} or @var{clone}.
20120 With @var{add} frames of solid-color are added.
20121 With @var{clone} frames are clones of last frame.
20122 Default is @var{add}.
20124 @item start_duration, stop_duration
20125 Specify the duration of the start/stop delay. See
20126 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20127 for the accepted syntax.
20128 These options override @var{start} and @var{stop}. Default is 0.
20131 Specify the color of the padded area. For the syntax of this option,
20132 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
20133 manual,ffmpeg-utils}.
20135 The default value of @var{color} is "black".
20141 Transpose rows with columns in the input video and optionally flip it.
20143 It accepts the following parameters:
20148 Specify the transposition direction.
20150 Can assume the following values:
20152 @item 0, 4, cclock_flip
20153 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
20161 Rotate by 90 degrees clockwise, that is:
20169 Rotate by 90 degrees counterclockwise, that is:
20176 @item 3, 7, clock_flip
20177 Rotate by 90 degrees clockwise and vertically flip, that is:
20185 For values between 4-7, the transposition is only done if the input
20186 video geometry is portrait and not landscape. These values are
20187 deprecated, the @code{passthrough} option should be used instead.
20189 Numerical values are deprecated, and should be dropped in favor of
20190 symbolic constants.
20193 Do not apply the transposition if the input geometry matches the one
20194 specified by the specified value. It accepts the following values:
20197 Always apply transposition.
20199 Preserve portrait geometry (when @var{height} >= @var{width}).
20201 Preserve landscape geometry (when @var{width} >= @var{height}).
20204 Default value is @code{none}.
20207 For example to rotate by 90 degrees clockwise and preserve portrait
20210 transpose=dir=1:passthrough=portrait
20213 The command above can also be specified as:
20215 transpose=1:portrait
20218 @section transpose_npp
20220 Transpose rows with columns in the input video and optionally flip it.
20221 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
20223 It accepts the following parameters:
20228 Specify the transposition direction.
20230 Can assume the following values:
20233 Rotate by 90 degrees counterclockwise and vertically flip. (default)
20236 Rotate by 90 degrees clockwise.
20239 Rotate by 90 degrees counterclockwise.
20242 Rotate by 90 degrees clockwise and vertically flip.
20246 Do not apply the transposition if the input geometry matches the one
20247 specified by the specified value. It accepts the following values:
20250 Always apply transposition. (default)
20252 Preserve portrait geometry (when @var{height} >= @var{width}).
20254 Preserve landscape geometry (when @var{width} >= @var{height}).
20260 Trim the input so that the output contains one continuous subpart of the input.
20262 It accepts the following parameters:
20265 Specify the time of the start of the kept section, i.e. the frame with the
20266 timestamp @var{start} will be the first frame in the output.
20269 Specify the time of the first frame that will be dropped, i.e. the frame
20270 immediately preceding the one with the timestamp @var{end} will be the last
20271 frame in the output.
20274 This is the same as @var{start}, except this option sets the start timestamp
20275 in timebase units instead of seconds.
20278 This is the same as @var{end}, except this option sets the end timestamp
20279 in timebase units instead of seconds.
20282 The maximum duration of the output in seconds.
20285 The number of the first frame that should be passed to the output.
20288 The number of the first frame that should be dropped.
20291 @option{start}, @option{end}, and @option{duration} are expressed as time
20292 duration specifications; see
20293 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20294 for the accepted syntax.
20296 Note that the first two sets of the start/end options and the @option{duration}
20297 option look at the frame timestamp, while the _frame variants simply count the
20298 frames that pass through the filter. Also note that this filter does not modify
20299 the timestamps. If you wish for the output timestamps to start at zero, insert a
20300 setpts filter after the trim filter.
20302 If multiple start or end options are set, this filter tries to be greedy and
20303 keep all the frames that match at least one of the specified constraints. To keep
20304 only the part that matches all the constraints at once, chain multiple trim
20307 The defaults are such that all the input is kept. So it is possible to set e.g.
20308 just the end values to keep everything before the specified time.
20313 Drop everything except the second minute of input:
20315 ffmpeg -i INPUT -vf trim=60:120
20319 Keep only the first second:
20321 ffmpeg -i INPUT -vf trim=duration=1
20326 @section unpremultiply
20327 Apply alpha unpremultiply effect to input video stream using first plane
20328 of second stream as alpha.
20330 Both streams must have same dimensions and same pixel format.
20332 The filter accepts the following option:
20336 Set which planes will be processed, unprocessed planes will be copied.
20337 By default value 0xf, all planes will be processed.
20339 If the format has 1 or 2 components, then luma is bit 0.
20340 If the format has 3 or 4 components:
20341 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
20342 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
20343 If present, the alpha channel is always the last bit.
20346 Do not require 2nd input for processing, instead use alpha plane from input stream.
20352 Sharpen or blur the input video.
20354 It accepts the following parameters:
20357 @item luma_msize_x, lx
20358 Set the luma matrix horizontal size. It must be an odd integer between
20359 3 and 23. The default value is 5.
20361 @item luma_msize_y, ly
20362 Set the luma matrix vertical size. It must be an odd integer between 3
20363 and 23. The default value is 5.
20365 @item luma_amount, la
20366 Set the luma effect strength. It must be a floating point number, reasonable
20367 values lay between -1.5 and 1.5.
20369 Negative values will blur the input video, while positive values will
20370 sharpen it, a value of zero will disable the effect.
20372 Default value is 1.0.
20374 @item chroma_msize_x, cx
20375 Set the chroma matrix horizontal size. It must be an odd integer
20376 between 3 and 23. The default value is 5.
20378 @item chroma_msize_y, cy
20379 Set the chroma matrix vertical size. It must be an odd integer
20380 between 3 and 23. The default value is 5.
20382 @item chroma_amount, ca
20383 Set the chroma effect strength. It must be a floating point number, reasonable
20384 values lay between -1.5 and 1.5.
20386 Negative values will blur the input video, while positive values will
20387 sharpen it, a value of zero will disable the effect.
20389 Default value is 0.0.
20393 All parameters are optional and default to the equivalent of the
20394 string '5:5:1.0:5:5:0.0'.
20396 @subsection Examples
20400 Apply strong luma sharpen effect:
20402 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
20406 Apply a strong blur of both luma and chroma parameters:
20408 unsharp=7:7:-2:7:7:-2
20415 Decompose a video made of tiled images into the individual images.
20417 The frame rate of the output video is the frame rate of the input video
20418 multiplied by the number of tiles.
20420 This filter does the reverse of @ref{tile}.
20422 The filter accepts the following options:
20427 Set the grid size (i.e. the number of lines and columns). For the syntax of
20428 this option, check the
20429 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20432 @subsection Examples
20436 Produce a 1-second video from a still image file made of 25 frames stacked
20437 vertically, like an analogic film reel:
20439 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
20445 Apply ultra slow/simple postprocessing filter that compresses and decompresses
20446 the image at several (or - in the case of @option{quality} level @code{8} - all)
20447 shifts and average the results.
20449 The way this differs from the behavior of spp is that uspp actually encodes &
20450 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
20451 DCT similar to MJPEG.
20453 The filter accepts the following options:
20457 Set quality. This option defines the number of levels for averaging. It accepts
20458 an integer in the range 0-8. If set to @code{0}, the filter will have no
20459 effect. A value of @code{8} means the higher quality. For each increment of
20460 that value the speed drops by a factor of approximately 2. Default value is
20464 Force a constant quantization parameter. If not set, the filter will use the QP
20465 from the video stream (if available).
20470 Convert 360 videos between various formats.
20472 The filter accepts the following options:
20478 Set format of the input/output video.
20486 Equirectangular projection.
20491 Cubemap with 3x2/6x1/1x6 layout.
20493 Format specific options:
20498 Set padding proportion for the input/output cubemap. Values in decimals.
20505 1% of face is padding. For example, with 1920x1280 resolution face size would be 640x640 and padding would be 3 pixels from each side. (640 * 0.01 = 6 pixels)
20508 Default value is @b{@samp{0}}.
20509 Maximum value is @b{@samp{0.1}}.
20513 Set fixed padding for the input/output cubemap. Values in pixels.
20515 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
20519 Set order of faces for the input/output cubemap. Choose one direction for each position.
20521 Designation of directions:
20537 Default value is @b{@samp{rludfb}}.
20541 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
20543 Designation of angles:
20546 0 degrees clockwise
20548 90 degrees clockwise
20550 180 degrees clockwise
20552 270 degrees clockwise
20555 Default value is @b{@samp{000000}}.
20559 Equi-Angular Cubemap.
20566 Format specific options:
20571 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20573 If diagonal field of view is set it overrides horizontal and vertical field of view.
20578 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20580 If diagonal field of view is set it overrides horizontal and vertical field of view.
20586 Format specific options:
20591 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20593 If diagonal field of view is set it overrides horizontal and vertical field of view.
20598 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20600 If diagonal field of view is set it overrides horizontal and vertical field of view.
20606 Facebook's 360 formats.
20609 Stereographic format.
20611 Format specific options:
20616 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20618 If diagonal field of view is set it overrides horizontal and vertical field of view.
20623 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20625 If diagonal field of view is set it overrides horizontal and vertical field of view.
20632 Ball format, gives significant distortion toward the back.
20635 Hammer-Aitoff map projection format.
20638 Sinusoidal map projection format.
20641 Fisheye projection.
20643 Format specific options:
20648 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20650 If diagonal field of view is set it overrides horizontal and vertical field of view.
20655 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20657 If diagonal field of view is set it overrides horizontal and vertical field of view.
20661 Pannini projection.
20663 Format specific options:
20666 Set output pannini parameter.
20669 Set input pannini parameter.
20673 Cylindrical projection.
20675 Format specific options:
20680 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20682 If diagonal field of view is set it overrides horizontal and vertical field of view.
20687 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20689 If diagonal field of view is set it overrides horizontal and vertical field of view.
20693 Perspective projection. @i{(output only)}
20695 Format specific options:
20698 Set perspective parameter.
20702 Tetrahedron projection.
20705 Truncated square pyramid projection.
20709 Half equirectangular projection.
20714 Format specific options:
20719 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20721 If diagonal field of view is set it overrides horizontal and vertical field of view.
20726 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20728 If diagonal field of view is set it overrides horizontal and vertical field of view.
20732 Orthographic format.
20734 Format specific options:
20739 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20741 If diagonal field of view is set it overrides horizontal and vertical field of view.
20746 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20748 If diagonal field of view is set it overrides horizontal and vertical field of view.
20752 Octahedron projection.
20756 Set interpolation method.@*
20757 @i{Note: more complex interpolation methods require much more memory to run.}
20767 Bilinear interpolation.
20769 Lagrange9 interpolation.
20772 Bicubic interpolation.
20775 Lanczos interpolation.
20778 Spline16 interpolation.
20781 Gaussian interpolation.
20783 Mitchell interpolation.
20786 Default value is @b{@samp{line}}.
20790 Set the output video resolution.
20792 Default resolution depends on formats.
20796 Set the input/output stereo format.
20807 Default value is @b{@samp{2d}} for input and output format.
20812 Set rotation for the output video. Values in degrees.
20815 Set rotation order for the output video. Choose one item for each position.
20826 Default value is @b{@samp{ypr}}.
20831 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20835 Set if input video is flipped horizontally/vertically. Boolean values.
20838 Set if input video is transposed. Boolean value, by default disabled.
20841 Set if output video needs to be transposed. Boolean value, by default disabled.
20844 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20847 @subsection Examples
20851 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20853 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20856 Extract back view of Equi-Angular Cubemap:
20858 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20861 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20863 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20867 @subsection Commands
20869 This filter supports subset of above options as @ref{commands}.
20871 @section vaguedenoiser
20873 Apply a wavelet based denoiser.
20875 It transforms each frame from the video input into the wavelet domain,
20876 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20877 the obtained coefficients. It does an inverse wavelet transform after.
20878 Due to wavelet properties, it should give a nice smoothed result, and
20879 reduced noise, without blurring picture features.
20881 This filter accepts the following options:
20885 The filtering strength. The higher, the more filtered the video will be.
20886 Hard thresholding can use a higher threshold than soft thresholding
20887 before the video looks overfiltered. Default value is 2.
20890 The filtering method the filter will use.
20892 It accepts the following values:
20895 All values under the threshold will be zeroed.
20898 All values under the threshold will be zeroed. All values above will be
20899 reduced by the threshold.
20902 Scales or nullifies coefficients - intermediary between (more) soft and
20903 (less) hard thresholding.
20906 Default is garrote.
20909 Number of times, the wavelet will decompose the picture. Picture can't
20910 be decomposed beyond a particular point (typically, 8 for a 640x480
20911 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20914 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20917 A list of the planes to process. By default all planes are processed.
20920 The threshold type the filter will use.
20922 It accepts the following values:
20925 Threshold used is same for all decompositions.
20928 Threshold used depends also on each decomposition coefficients.
20931 Default is universal.
20934 @section vectorscope
20936 Display 2 color component values in the two dimensional graph (which is called
20939 This filter accepts the following options:
20943 Set vectorscope mode.
20945 It accepts the following values:
20949 Gray values are displayed on graph, higher brightness means more pixels have
20950 same component color value on location in graph. This is the default mode.
20953 Gray values are displayed on graph. Surrounding pixels values which are not
20954 present in video frame are drawn in gradient of 2 color components which are
20955 set by option @code{x} and @code{y}. The 3rd color component is static.
20958 Actual color components values present in video frame are displayed on graph.
20961 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20962 on graph increases value of another color component, which is luminance by
20963 default values of @code{x} and @code{y}.
20966 Actual colors present in video frame are displayed on graph. If two different
20967 colors map to same position on graph then color with higher value of component
20968 not present in graph is picked.
20971 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20972 component picked from radial gradient.
20976 Set which color component will be represented on X-axis. Default is @code{1}.
20979 Set which color component will be represented on Y-axis. Default is @code{2}.
20982 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20983 of color component which represents frequency of (X, Y) location in graph.
20988 No envelope, this is default.
20991 Instant envelope, even darkest single pixel will be clearly highlighted.
20994 Hold maximum and minimum values presented in graph over time. This way you
20995 can still spot out of range values without constantly looking at vectorscope.
20998 Peak and instant envelope combined together.
21002 Set what kind of graticule to draw.
21011 Set graticule opacity.
21014 Set graticule flags.
21018 Draw graticule for white point.
21021 Draw graticule for black point.
21024 Draw color points short names.
21028 Set background opacity.
21030 @item lthreshold, l
21031 Set low threshold for color component not represented on X or Y axis.
21032 Values lower than this value will be ignored. Default is 0.
21033 Note this value is multiplied with actual max possible value one pixel component
21034 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
21037 @item hthreshold, h
21038 Set high threshold for color component not represented on X or Y axis.
21039 Values higher than this value will be ignored. Default is 1.
21040 Note this value is multiplied with actual max possible value one pixel component
21041 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
21042 is 0.9 * 255 = 230.
21044 @item colorspace, c
21045 Set what kind of colorspace to use when drawing graticule.
21055 Set color tint for gray/tint vectorscope mode. By default both options are zero.
21056 This means no tint, and output will remain gray.
21059 @anchor{vidstabdetect}
21060 @section vidstabdetect
21062 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
21063 @ref{vidstabtransform} for pass 2.
21065 This filter generates a file with relative translation and rotation
21066 transform information about subsequent frames, which is then used by
21067 the @ref{vidstabtransform} filter.
21069 To enable compilation of this filter you need to configure FFmpeg with
21070 @code{--enable-libvidstab}.
21072 This filter accepts the following options:
21076 Set the path to the file used to write the transforms information.
21077 Default value is @file{transforms.trf}.
21080 Set how shaky the video is and how quick the camera is. It accepts an
21081 integer in the range 1-10, a value of 1 means little shakiness, a
21082 value of 10 means strong shakiness. Default value is 5.
21085 Set the accuracy of the detection process. It must be a value in the
21086 range 1-15. A value of 1 means low accuracy, a value of 15 means high
21087 accuracy. Default value is 15.
21090 Set stepsize of the search process. The region around minimum is
21091 scanned with 1 pixel resolution. Default value is 6.
21094 Set minimum contrast. Below this value a local measurement field is
21095 discarded. Must be a floating point value in the range 0-1. Default
21099 Set reference frame number for tripod mode.
21101 If enabled, the motion of the frames is compared to a reference frame
21102 in the filtered stream, identified by the specified number. The idea
21103 is to compensate all movements in a more-or-less static scene and keep
21104 the camera view absolutely still.
21106 If set to 0, it is disabled. The frames are counted starting from 1.
21109 Show fields and transforms in the resulting frames. It accepts an
21110 integer in the range 0-2. Default value is 0, which disables any
21114 @subsection Examples
21118 Use default values:
21124 Analyze strongly shaky movie and put the results in file
21125 @file{mytransforms.trf}:
21127 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
21131 Visualize the result of internal transformations in the resulting
21134 vidstabdetect=show=1
21138 Analyze a video with medium shakiness using @command{ffmpeg}:
21140 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
21144 @anchor{vidstabtransform}
21145 @section vidstabtransform
21147 Video stabilization/deshaking: pass 2 of 2,
21148 see @ref{vidstabdetect} for pass 1.
21150 Read a file with transform information for each frame and
21151 apply/compensate them. Together with the @ref{vidstabdetect}
21152 filter this can be used to deshake videos. See also
21153 @url{http://public.hronopik.de/vid.stab}. It is important to also use
21154 the @ref{unsharp} filter, see below.
21156 To enable compilation of this filter you need to configure FFmpeg with
21157 @code{--enable-libvidstab}.
21159 @subsection Options
21163 Set path to the file used to read the transforms. Default value is
21164 @file{transforms.trf}.
21167 Set the number of frames (value*2 + 1) used for lowpass filtering the
21168 camera movements. Default value is 10.
21170 For example a number of 10 means that 21 frames are used (10 in the
21171 past and 10 in the future) to smoothen the motion in the video. A
21172 larger value leads to a smoother video, but limits the acceleration of
21173 the camera (pan/tilt movements). 0 is a special case where a static
21174 camera is simulated.
21177 Set the camera path optimization algorithm.
21179 Accepted values are:
21182 gaussian kernel low-pass filter on camera motion (default)
21184 averaging on transformations
21188 Set maximal number of pixels to translate frames. Default value is -1,
21192 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
21193 value is -1, meaning no limit.
21196 Specify how to deal with borders that may be visible due to movement
21199 Available values are:
21202 keep image information from previous frame (default)
21204 fill the border black
21208 Invert transforms if set to 1. Default value is 0.
21211 Consider transforms as relative to previous frame if set to 1,
21212 absolute if set to 0. Default value is 0.
21215 Set percentage to zoom. A positive value will result in a zoom-in
21216 effect, a negative value in a zoom-out effect. Default value is 0 (no
21220 Set optimal zooming to avoid borders.
21222 Accepted values are:
21227 optimal static zoom value is determined (only very strong movements
21228 will lead to visible borders) (default)
21230 optimal adaptive zoom value is determined (no borders will be
21231 visible), see @option{zoomspeed}
21234 Note that the value given at zoom is added to the one calculated here.
21237 Set percent to zoom maximally each frame (enabled when
21238 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
21242 Specify type of interpolation.
21244 Available values are:
21249 linear only horizontal
21251 linear in both directions (default)
21253 cubic in both directions (slow)
21257 Enable virtual tripod mode if set to 1, which is equivalent to
21258 @code{relative=0:smoothing=0}. Default value is 0.
21260 Use also @code{tripod} option of @ref{vidstabdetect}.
21263 Increase log verbosity if set to 1. Also the detected global motions
21264 are written to the temporary file @file{global_motions.trf}. Default
21268 @subsection Examples
21272 Use @command{ffmpeg} for a typical stabilization with default values:
21274 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
21277 Note the use of the @ref{unsharp} filter which is always recommended.
21280 Zoom in a bit more and load transform data from a given file:
21282 vidstabtransform=zoom=5:input="mytransforms.trf"
21286 Smoothen the video even more:
21288 vidstabtransform=smoothing=30
21294 Flip the input video vertically.
21296 For example, to vertically flip a video with @command{ffmpeg}:
21298 ffmpeg -i in.avi -vf "vflip" out.avi
21303 Detect variable frame rate video.
21305 This filter tries to detect if the input is variable or constant frame rate.
21307 At end it will output number of frames detected as having variable delta pts,
21308 and ones with constant delta pts.
21309 If there was frames with variable delta, than it will also show min, max and
21310 average delta encountered.
21314 Boost or alter saturation.
21316 The filter accepts the following options:
21319 Set strength of boost if positive value or strength of alter if negative value.
21320 Default is 0. Allowed range is from -2 to 2.
21323 Set the red balance. Default is 1. Allowed range is from -10 to 10.
21326 Set the green balance. Default is 1. Allowed range is from -10 to 10.
21329 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
21332 Set the red luma coefficient.
21335 Set the green luma coefficient.
21338 Set the blue luma coefficient.
21341 If @code{intensity} is negative and this is set to 1, colors will change,
21342 otherwise colors will be less saturated, more towards gray.
21345 @subsection Commands
21347 This filter supports the all above options as @ref{commands}.
21351 Obtain the average VIF (Visual Information Fidelity) between two input videos.
21353 This filter takes two input videos.
21355 Both input videos must have the same resolution and pixel format for
21356 this filter to work correctly. Also it assumes that both inputs
21357 have the same number of frames, which are compared one by one.
21359 The obtained average VIF score is printed through the logging system.
21361 The filter stores the calculated VIF score of each frame.
21363 In the below example the input file @file{main.mpg} being processed is compared
21364 with the reference file @file{ref.mpg}.
21367 ffmpeg -i main.mpg -i ref.mpg -lavfi vif -f null -
21373 Make or reverse a natural vignetting effect.
21375 The filter accepts the following options:
21379 Set lens angle expression as a number of radians.
21381 The value is clipped in the @code{[0,PI/2]} range.
21383 Default value: @code{"PI/5"}
21387 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
21391 Set forward/backward mode.
21393 Available modes are:
21396 The larger the distance from the central point, the darker the image becomes.
21399 The larger the distance from the central point, the brighter the image becomes.
21400 This can be used to reverse a vignette effect, though there is no automatic
21401 detection to extract the lens @option{angle} and other settings (yet). It can
21402 also be used to create a burning effect.
21405 Default value is @samp{forward}.
21408 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
21410 It accepts the following values:
21413 Evaluate expressions only once during the filter initialization.
21416 Evaluate expressions for each incoming frame. This is way slower than the
21417 @samp{init} mode since it requires all the scalers to be re-computed, but it
21418 allows advanced dynamic expressions.
21421 Default value is @samp{init}.
21424 Set dithering to reduce the circular banding effects. Default is @code{1}
21428 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
21429 Setting this value to the SAR of the input will make a rectangular vignetting
21430 following the dimensions of the video.
21432 Default is @code{1/1}.
21435 @subsection Expressions
21437 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
21438 following parameters.
21443 input width and height
21446 the number of input frame, starting from 0
21449 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
21450 @var{TB} units, NAN if undefined
21453 frame rate of the input video, NAN if the input frame rate is unknown
21456 the PTS (Presentation TimeStamp) of the filtered video frame,
21457 expressed in seconds, NAN if undefined
21460 time base of the input video
21464 @subsection Examples
21468 Apply simple strong vignetting effect:
21474 Make a flickering vignetting:
21476 vignette='PI/4+random(1)*PI/50':eval=frame
21481 @section vmafmotion
21483 Obtain the average VMAF motion score of a video.
21484 It is one of the component metrics of VMAF.
21486 The obtained average motion score is printed through the logging system.
21488 The filter accepts the following options:
21492 If specified, the filter will use the named file to save the motion score of
21493 each frame with respect to the previous frame.
21494 When filename equals "-" the data is sent to standard output.
21499 ffmpeg -i ref.mpg -vf vmafmotion -f null -
21503 Stack input videos vertically.
21505 All streams must be of same pixel format and of same width.
21507 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
21508 to create same output.
21510 The filter accepts the following options:
21514 Set number of input streams. Default is 2.
21517 If set to 1, force the output to terminate when the shortest input
21518 terminates. Default value is 0.
21523 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
21524 Deinterlacing Filter").
21526 Based on the process described by Martin Weston for BBC R&D, and
21527 implemented based on the de-interlace algorithm written by Jim
21528 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
21529 uses filter coefficients calculated by BBC R&D.
21531 This filter uses field-dominance information in frame to decide which
21532 of each pair of fields to place first in the output.
21533 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
21535 There are two sets of filter coefficients, so called "simple"
21536 and "complex". Which set of filter coefficients is used can
21537 be set by passing an optional parameter:
21541 Set the interlacing filter coefficients. Accepts one of the following values:
21545 Simple filter coefficient set.
21547 More-complex filter coefficient set.
21549 Default value is @samp{complex}.
21552 The interlacing mode to adopt. It accepts one of the following values:
21556 Output one frame for each frame.
21558 Output one frame for each field.
21561 The default value is @code{field}.
21564 The picture field parity assumed for the input interlaced video. It accepts one
21565 of the following values:
21569 Assume the top field is first.
21571 Assume the bottom field is first.
21573 Enable automatic detection of field parity.
21576 The default value is @code{auto}.
21577 If the interlacing is unknown or the decoder does not export this information,
21578 top field first will be assumed.
21581 Specify which frames to deinterlace. Accepts one of the following values:
21585 Deinterlace all frames,
21587 Only deinterlace frames marked as interlaced.
21590 Default value is @samp{all}.
21593 @subsection Commands
21594 This filter supports same @ref{commands} as options.
21597 Video waveform monitor.
21599 The waveform monitor plots color component intensity. By default luminance
21600 only. Each column of the waveform corresponds to a column of pixels in the
21603 It accepts the following options:
21607 Can be either @code{row}, or @code{column}. Default is @code{column}.
21608 In row mode, the graph on the left side represents color component value 0 and
21609 the right side represents value = 255. In column mode, the top side represents
21610 color component value = 0 and bottom side represents value = 255.
21613 Set intensity. Smaller values are useful to find out how many values of the same
21614 luminance are distributed across input rows/columns.
21615 Default value is @code{0.04}. Allowed range is [0, 1].
21618 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
21619 In mirrored mode, higher values will be represented on the left
21620 side for @code{row} mode and at the top for @code{column} mode. Default is
21621 @code{1} (mirrored).
21625 It accepts the following values:
21628 Presents information identical to that in the @code{parade}, except
21629 that the graphs representing color components are superimposed directly
21632 This display mode makes it easier to spot relative differences or similarities
21633 in overlapping areas of the color components that are supposed to be identical,
21634 such as neutral whites, grays, or blacks.
21637 Display separate graph for the color components side by side in
21638 @code{row} mode or one below the other in @code{column} mode.
21641 Display separate graph for the color components side by side in
21642 @code{column} mode or one below the other in @code{row} mode.
21644 Using this display mode makes it easy to spot color casts in the highlights
21645 and shadows of an image, by comparing the contours of the top and the bottom
21646 graphs of each waveform. Since whites, grays, and blacks are characterized
21647 by exactly equal amounts of red, green, and blue, neutral areas of the picture
21648 should display three waveforms of roughly equal width/height. If not, the
21649 correction is easy to perform by making level adjustments the three waveforms.
21651 Default is @code{stack}.
21653 @item components, c
21654 Set which color components to display. Default is 1, which means only luminance
21655 or red color component if input is in RGB colorspace. If is set for example to
21656 7 it will display all 3 (if) available color components.
21661 No envelope, this is default.
21664 Instant envelope, minimum and maximum values presented in graph will be easily
21665 visible even with small @code{step} value.
21668 Hold minimum and maximum values presented in graph across time. This way you
21669 can still spot out of range values without constantly looking at waveforms.
21672 Peak and instant envelope combined together.
21678 No filtering, this is default.
21681 Luma and chroma combined together.
21684 Similar as above, but shows difference between blue and red chroma.
21687 Similar as above, but use different colors.
21690 Similar as above, but again with different colors.
21693 Displays only chroma.
21696 Displays actual color value on waveform.
21699 Similar as above, but with luma showing frequency of chroma values.
21703 Set which graticule to display.
21707 Do not display graticule.
21710 Display green graticule showing legal broadcast ranges.
21713 Display orange graticule showing legal broadcast ranges.
21716 Display invert graticule showing legal broadcast ranges.
21720 Set graticule opacity.
21723 Set graticule flags.
21727 Draw numbers above lines. By default enabled.
21730 Draw dots instead of lines.
21734 Set scale used for displaying graticule.
21741 Default is digital.
21744 Set background opacity.
21748 Set tint for output.
21749 Only used with lowpass filter and when display is not overlay and input
21750 pixel formats are not RGB.
21753 @section weave, doubleweave
21755 The @code{weave} takes a field-based video input and join
21756 each two sequential fields into single frame, producing a new double
21757 height clip with half the frame rate and half the frame count.
21759 The @code{doubleweave} works same as @code{weave} but without
21760 halving frame rate and frame count.
21762 It accepts the following option:
21766 Set first field. Available values are:
21770 Set the frame as top-field-first.
21773 Set the frame as bottom-field-first.
21777 @subsection Examples
21781 Interlace video using @ref{select} and @ref{separatefields} filter:
21783 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21788 Apply the xBR high-quality magnification filter which is designed for pixel
21789 art. It follows a set of edge-detection rules, see
21790 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21792 It accepts the following option:
21796 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21797 @code{3xBR} and @code{4} for @code{4xBR}.
21798 Default is @code{3}.
21803 Apply cross fade from one input video stream to another input video stream.
21804 The cross fade is applied for specified duration.
21806 Both inputs must be constant frame-rate and have the same resolution, pixel format,
21807 frame rate and timebase.
21809 The filter accepts the following options:
21813 Set one of available transition effects:
21861 Default transition effect is fade.
21864 Set cross fade duration in seconds.
21865 Default duration is 1 second.
21868 Set cross fade start relative to first input stream in seconds.
21869 Default offset is 0.
21872 Set expression for custom transition effect.
21874 The expressions can use the following variables and functions:
21879 The coordinates of the current sample.
21883 The width and height of the image.
21886 Progress of transition effect.
21889 Currently processed plane.
21892 Return value of first input at current location and plane.
21895 Return value of second input at current location and plane.
21901 Return the value of the pixel at location (@var{x},@var{y}) of the
21902 first/second/third/fourth component of first input.
21908 Return the value of the pixel at location (@var{x},@var{y}) of the
21909 first/second/third/fourth component of second input.
21913 @subsection Examples
21917 Cross fade from one input video to another input video, with fade transition and duration of transition
21918 of 2 seconds starting at offset of 5 seconds:
21920 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21925 Pick median pixels from several input videos.
21927 The filter accepts the following options:
21931 Set number of inputs.
21932 Default is 3. Allowed range is from 3 to 255.
21933 If number of inputs is even number, than result will be mean value between two median values.
21936 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21939 Set median percentile. Default value is @code{0.5}.
21940 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21941 minimum values, and @code{1} maximum values.
21944 @subsection Commands
21946 This filter supports all above options as @ref{commands}, excluding option @code{inputs}.
21949 Stack video inputs into custom layout.
21951 All streams must be of same pixel format.
21953 The filter accepts the following options:
21957 Set number of input streams. Default is 2.
21960 Specify layout of inputs.
21961 This option requires the desired layout configuration to be explicitly set by the user.
21962 This sets position of each video input in output. Each input
21963 is separated by '|'.
21964 The first number represents the column, and the second number represents the row.
21965 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21966 where X is video input from which to take width or height.
21967 Multiple values can be used when separated by '+'. In such
21968 case values are summed together.
21970 Note that if inputs are of different sizes gaps may appear, as not all of
21971 the output video frame will be filled. Similarly, videos can overlap each
21972 other if their position doesn't leave enough space for the full frame of
21975 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21976 a layout must be set by the user.
21979 If set to 1, force the output to terminate when the shortest input
21980 terminates. Default value is 0.
21983 If set to valid color, all unused pixels will be filled with that color.
21984 By default fill is set to none, so it is disabled.
21987 @subsection Examples
21991 Display 4 inputs into 2x2 grid.
21995 input1(0, 0) | input3(w0, 0)
21996 input2(0, h0) | input4(w0, h0)
22000 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
22003 Note that if inputs are of different sizes, gaps or overlaps may occur.
22006 Display 4 inputs into 1x4 grid.
22013 input4(0, h0+h1+h2)
22017 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
22020 Note that if inputs are of different widths, unused space will appear.
22023 Display 9 inputs into 3x3 grid.
22027 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
22028 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
22029 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
22033 xstack=inputs=9:layout=0_0|0_h0|0_h0+h1|w0_0|w0_h0|w0_h0+h1|w0+w3_0|w0+w3_h0|w0+w3_h0+h1
22036 Note that if inputs are of different sizes, gaps or overlaps may occur.
22039 Display 16 inputs into 4x4 grid.
22043 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
22044 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
22045 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
22046 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
22050 xstack=inputs=16:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2|w0_0|w0_h0|w0_h0+h1|w0_h0+h1+h2|w0+w4_0|
22051 w0+w4_h0|w0+w4_h0+h1|w0+w4_h0+h1+h2|w0+w4+w8_0|w0+w4+w8_h0|w0+w4+w8_h0+h1|w0+w4+w8_h0+h1+h2
22054 Note that if inputs are of different sizes, gaps or overlaps may occur.
22061 Deinterlace the input video ("yadif" means "yet another deinterlacing
22064 It accepts the following parameters:
22070 The interlacing mode to adopt. It accepts one of the following values:
22073 @item 0, send_frame
22074 Output one frame for each frame.
22075 @item 1, send_field
22076 Output one frame for each field.
22077 @item 2, send_frame_nospatial
22078 Like @code{send_frame}, but it skips the spatial interlacing check.
22079 @item 3, send_field_nospatial
22080 Like @code{send_field}, but it skips the spatial interlacing check.
22083 The default value is @code{send_frame}.
22086 The picture field parity assumed for the input interlaced video. It accepts one
22087 of the following values:
22091 Assume the top field is first.
22093 Assume the bottom field is first.
22095 Enable automatic detection of field parity.
22098 The default value is @code{auto}.
22099 If the interlacing is unknown or the decoder does not export this information,
22100 top field first will be assumed.
22103 Specify which frames to deinterlace. Accepts one of the following
22108 Deinterlace all frames.
22109 @item 1, interlaced
22110 Only deinterlace frames marked as interlaced.
22113 The default value is @code{all}.
22116 @section yadif_cuda
22118 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
22119 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
22122 It accepts the following parameters:
22128 The interlacing mode to adopt. It accepts one of the following values:
22131 @item 0, send_frame
22132 Output one frame for each frame.
22133 @item 1, send_field
22134 Output one frame for each field.
22135 @item 2, send_frame_nospatial
22136 Like @code{send_frame}, but it skips the spatial interlacing check.
22137 @item 3, send_field_nospatial
22138 Like @code{send_field}, but it skips the spatial interlacing check.
22141 The default value is @code{send_frame}.
22144 The picture field parity assumed for the input interlaced video. It accepts one
22145 of the following values:
22149 Assume the top field is first.
22151 Assume the bottom field is first.
22153 Enable automatic detection of field parity.
22156 The default value is @code{auto}.
22157 If the interlacing is unknown or the decoder does not export this information,
22158 top field first will be assumed.
22161 Specify which frames to deinterlace. Accepts one of the following
22166 Deinterlace all frames.
22167 @item 1, interlaced
22168 Only deinterlace frames marked as interlaced.
22171 The default value is @code{all}.
22176 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
22177 The algorithm is described in
22178 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
22180 It accepts the following parameters:
22184 Set the window radius. Default value is 3.
22187 Set which planes to filter. Default is only the first plane.
22190 Set blur strength. Default value is 128.
22193 @subsection Commands
22194 This filter supports same @ref{commands} as options.
22198 Apply Zoom & Pan effect.
22200 This filter accepts the following options:
22204 Set the zoom expression. Range is 1-10. Default is 1.
22208 Set the x and y expression. Default is 0.
22211 Set the duration expression in number of frames.
22212 This sets for how many number of frames effect will last for
22213 single input image. Default is 90.
22216 Set the output image size, default is 'hd720'.
22219 Set the output frame rate, default is '25'.
22222 Each expression can contain the following constants:
22241 Output frame count.
22244 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
22246 @item out_time, time, ot
22247 The output timestamp expressed in seconds.
22251 Last calculated 'x' and 'y' position from 'x' and 'y' expression
22252 for current input frame.
22256 'x' and 'y' of last output frame of previous input frame or 0 when there was
22257 not yet such frame (first input frame).
22260 Last calculated zoom from 'z' expression for current input frame.
22263 Last calculated zoom of last output frame of previous input frame.
22266 Number of output frames for current input frame. Calculated from 'd' expression
22267 for each input frame.
22270 number of output frames created for previous input frame
22273 Rational number: input width / input height
22276 sample aspect ratio
22279 display aspect ratio
22283 @subsection Examples
22287 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
22289 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='if(gte(zoom,1.5),x,x+1/a)':y='if(gte(zoom,1.5),y,y+1)':s=640x360
22293 Zoom in up to 1.5x and pan always at center of picture:
22295 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
22299 Same as above but without pausing:
22301 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
22305 Zoom in 2x into center of picture only for the first second of the input video:
22307 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
22314 Scale (resize) the input video, using the z.lib library:
22315 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
22316 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
22318 The zscale filter forces the output display aspect ratio to be the same
22319 as the input, by changing the output sample aspect ratio.
22321 If the input image format is different from the format requested by
22322 the next filter, the zscale filter will convert the input to the
22325 @subsection Options
22326 The filter accepts the following options.
22331 Set the output video dimension expression. Default value is the input
22334 If the @var{width} or @var{w} value is 0, the input width is used for
22335 the output. If the @var{height} or @var{h} value is 0, the input height
22336 is used for the output.
22338 If one and only one of the values is -n with n >= 1, the zscale filter
22339 will use a value that maintains the aspect ratio of the input image,
22340 calculated from the other specified dimension. After that it will,
22341 however, make sure that the calculated dimension is divisible by n and
22342 adjust the value if necessary.
22344 If both values are -n with n >= 1, the behavior will be identical to
22345 both values being set to 0 as previously detailed.
22347 See below for the list of accepted constants for use in the dimension
22351 Set the video size. For the syntax of this option, check the
22352 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22355 Set the dither type.
22357 Possible values are:
22362 @item error_diffusion
22368 Set the resize filter type.
22370 Possible values are:
22380 Default is bilinear.
22383 Set the color range.
22385 Possible values are:
22392 Default is same as input.
22395 Set the color primaries.
22397 Possible values are:
22407 Default is same as input.
22410 Set the transfer characteristics.
22412 Possible values are:
22426 Default is same as input.
22429 Set the colorspace matrix.
22431 Possible value are:
22442 Default is same as input.
22445 Set the input color range.
22447 Possible values are:
22454 Default is same as input.
22456 @item primariesin, pin
22457 Set the input color primaries.
22459 Possible values are:
22469 Default is same as input.
22471 @item transferin, tin
22472 Set the input transfer characteristics.
22474 Possible values are:
22485 Default is same as input.
22487 @item matrixin, min
22488 Set the input colorspace matrix.
22490 Possible value are:
22502 Set the output chroma location.
22504 Possible values are:
22515 @item chromalin, cin
22516 Set the input chroma location.
22518 Possible values are:
22530 Set the nominal peak luminance.
22533 Parameter A for scaling filters. Parameter "b" for bicubic, and the number of
22534 filter taps for lanczos.
22537 Parameter B for scaling filters. Parameter "c" for bicubic.
22540 The values of the @option{w} and @option{h} options are expressions
22541 containing the following constants:
22546 The input width and height
22550 These are the same as @var{in_w} and @var{in_h}.
22554 The output (scaled) width and height
22558 These are the same as @var{out_w} and @var{out_h}
22561 The same as @var{iw} / @var{ih}
22564 input sample aspect ratio
22567 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
22571 horizontal and vertical input chroma subsample values. For example for the
22572 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22576 horizontal and vertical output chroma subsample values. For example for the
22577 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22580 @subsection Commands
22582 This filter supports the following commands:
22586 Set the output video dimension expression.
22587 The command accepts the same syntax of the corresponding option.
22589 If the specified expression is not valid, it is kept at its current
22593 @c man end VIDEO FILTERS
22595 @chapter OpenCL Video Filters
22596 @c man begin OPENCL VIDEO FILTERS
22598 Below is a description of the currently available OpenCL video filters.
22600 To enable compilation of these filters you need to configure FFmpeg with
22601 @code{--enable-opencl}.
22603 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
22606 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
22607 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
22608 given device parameters.
22610 @item -filter_hw_device @var{name}
22611 Pass the hardware device called @var{name} to all filters in any filter graph.
22615 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
22619 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
22621 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
22625 Since OpenCL filters are not able to access frame data in normal memory, all frame data needs to be uploaded(@ref{hwupload}) to hardware surfaces connected to the appropriate device before being used and then downloaded(@ref{hwdownload}) back to normal memory. Note that @ref{hwupload} will upload to a surface with the same layout as the software frame, so it may be necessary to add a @ref{format} filter immediately before to get the input into the right format and @ref{hwdownload} does not support all formats on the output - it may be necessary to insert an additional @ref{format} filter immediately following in the graph to get the output in a supported format.
22627 @section avgblur_opencl
22629 Apply average blur filter.
22631 The filter accepts the following options:
22635 Set horizontal radius size.
22636 Range is @code{[1, 1024]} and default value is @code{1}.
22639 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22642 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
22645 @subsection Example
22649 Apply average blur filter with horizontal and vertical size of 3, setting each pixel of the output to the average value of the 7x7 region centered on it in the input. For pixels on the edges of the image, the region does not extend beyond the image boundaries, and so out-of-range coordinates are not used in the calculations.
22651 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
22655 @section boxblur_opencl
22657 Apply a boxblur algorithm to the input video.
22659 It accepts the following parameters:
22663 @item luma_radius, lr
22664 @item luma_power, lp
22665 @item chroma_radius, cr
22666 @item chroma_power, cp
22667 @item alpha_radius, ar
22668 @item alpha_power, ap
22672 A description of the accepted options follows.
22675 @item luma_radius, lr
22676 @item chroma_radius, cr
22677 @item alpha_radius, ar
22678 Set an expression for the box radius in pixels used for blurring the
22679 corresponding input plane.
22681 The radius value must be a non-negative number, and must not be
22682 greater than the value of the expression @code{min(w,h)/2} for the
22683 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
22686 Default value for @option{luma_radius} is "2". If not specified,
22687 @option{chroma_radius} and @option{alpha_radius} default to the
22688 corresponding value set for @option{luma_radius}.
22690 The expressions can contain the following constants:
22694 The input width and height in pixels.
22698 The input chroma image width and height in pixels.
22702 The horizontal and vertical chroma subsample values. For example, for the
22703 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22706 @item luma_power, lp
22707 @item chroma_power, cp
22708 @item alpha_power, ap
22709 Specify how many times the boxblur filter is applied to the
22710 corresponding plane.
22712 Default value for @option{luma_power} is 2. If not specified,
22713 @option{chroma_power} and @option{alpha_power} default to the
22714 corresponding value set for @option{luma_power}.
22716 A value of 0 will disable the effect.
22719 @subsection Examples
22721 Apply boxblur filter, setting each pixel of the output to the average value of box-radiuses @var{luma_radius}, @var{chroma_radius}, @var{alpha_radius} for each plane respectively. The filter will apply @var{luma_power}, @var{chroma_power}, @var{alpha_power} times onto the corresponding plane. For pixels on the edges of the image, the radius does not extend beyond the image boundaries, and so out-of-range coordinates are not used in the calculations.
22725 Apply a boxblur filter with the luma, chroma, and alpha radius
22726 set to 2 and luma, chroma, and alpha power set to 3. The filter will run 3 times with box-radius set to 2 for every plane of the image.
22728 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22729 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22733 Apply a boxblur filter with luma radius set to 2, luma_power to 1, chroma_radius to 4, chroma_power to 5, alpha_radius to 3 and alpha_power to 7.
22735 For the luma plane, a 2x2 box radius will be run once.
22737 For the chroma plane, a 4x4 box radius will be run 5 times.
22739 For the alpha plane, a 3x3 box radius will be run 7 times.
22741 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22745 @section colorkey_opencl
22746 RGB colorspace color keying.
22748 The filter accepts the following options:
22752 The color which will be replaced with transparency.
22755 Similarity percentage with the key color.
22757 0.01 matches only the exact key color, while 1.0 matches everything.
22762 0.0 makes pixels either fully transparent, or not transparent at all.
22764 Higher values result in semi-transparent pixels, with a higher transparency
22765 the more similar the pixels color is to the key color.
22768 @subsection Examples
22772 Make every semi-green pixel in the input transparent with some slight blending:
22774 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22778 @section convolution_opencl
22780 Apply convolution of 3x3, 5x5, 7x7 matrix.
22782 The filter accepts the following options:
22789 Set matrix for each plane.
22790 Matrix is sequence of 9, 25 or 49 signed numbers.
22791 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22797 Set multiplier for calculated value for each plane.
22798 If unset or 0, it will be sum of all matrix elements.
22799 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22805 Set bias for each plane. This value is added to the result of the multiplication.
22806 Useful for making the overall image brighter or darker.
22807 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22811 @subsection Examples
22817 -i INPUT -vf "hwupload, convolution_opencl=0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0, hwdownload" OUTPUT
22823 -i INPUT -vf "hwupload, convolution_opencl=1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1/9:1/9:1/9:1/9, hwdownload" OUTPUT
22827 Apply edge enhance:
22829 -i INPUT -vf "hwupload, convolution_opencl=0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:5:1:1:1:0:128:128:128, hwdownload" OUTPUT
22835 -i INPUT -vf "hwupload, convolution_opencl=0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:5:5:5:1:0:128:128:128, hwdownload" OUTPUT
22839 Apply laplacian edge detector which includes diagonals:
22841 -i INPUT -vf "hwupload, convolution_opencl=1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:5:5:5:1:0:128:128:0, hwdownload" OUTPUT
22847 -i INPUT -vf "hwupload, convolution_opencl=-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2, hwdownload" OUTPUT
22851 @section erosion_opencl
22853 Apply erosion effect to the video.
22855 This filter replaces the pixel by the local(3x3) minimum.
22857 It accepts the following options:
22864 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22865 If @code{0}, plane will remain unchanged.
22868 Flag which specifies the pixel to refer to.
22869 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22871 Flags to local 3x3 coordinates region centered on @code{x}:
22880 @subsection Example
22884 Apply erosion filter with threshold0 set to 30, threshold1 set 40, threshold2 set to 50 and coordinates set to 231, setting each pixel of the output to the local minimum between pixels: 1, 2, 3, 6, 7, 8 of the 3x3 region centered on it in the input. If the difference between input pixel and local minimum is more then threshold of the corresponding plane, output pixel will be set to input pixel - threshold of corresponding plane.
22886 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22890 @section deshake_opencl
22891 Feature-point based video stabilization filter.
22893 The filter accepts the following options:
22897 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22900 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22902 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22904 Viewing point matches in the output video is only supported for RGB input.
22906 Defaults to @code{0}.
22908 @item adaptive_crop
22909 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22911 Defaults to @code{1}.
22913 @item refine_features
22914 Whether or not feature points should be refined at a sub-pixel level.
22916 This can be turned off for a slight performance gain at the cost of precision.
22918 Defaults to @code{1}.
22920 @item smooth_strength
22921 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22923 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22925 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22927 Defaults to @code{0.0}.
22929 @item smooth_window_multiplier
22930 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22932 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22934 Acceptable values range from @code{0.1} to @code{10.0}.
22936 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22937 potentially improving smoothness, but also increase latency and memory usage.
22939 Defaults to @code{2.0}.
22943 @subsection Examples
22947 Stabilize a video with a fixed, medium smoothing strength:
22949 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22953 Stabilize a video with debugging (both in console and in rendered video):
22955 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22959 @section dilation_opencl
22961 Apply dilation effect to the video.
22963 This filter replaces the pixel by the local(3x3) maximum.
22965 It accepts the following options:
22972 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22973 If @code{0}, plane will remain unchanged.
22976 Flag which specifies the pixel to refer to.
22977 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22979 Flags to local 3x3 coordinates region centered on @code{x}:
22988 @subsection Example
22992 Apply dilation filter with threshold0 set to 30, threshold1 set 40, threshold2 set to 50 and coordinates set to 231, setting each pixel of the output to the local maximum between pixels: 1, 2, 3, 6, 7, 8 of the 3x3 region centered on it in the input. If the difference between input pixel and local maximum is more then threshold of the corresponding plane, output pixel will be set to input pixel + threshold of corresponding plane.
22994 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22998 @section nlmeans_opencl
23000 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
23002 @section overlay_opencl
23004 Overlay one video on top of another.
23006 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
23007 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
23009 The filter accepts the following options:
23014 Set the x coordinate of the overlaid video on the main video.
23015 Default value is @code{0}.
23018 Set the y coordinate of the overlaid video on the main video.
23019 Default value is @code{0}.
23023 @subsection Examples
23027 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
23029 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
23032 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
23034 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
23039 @section pad_opencl
23041 Add paddings to the input image, and place the original input at the
23042 provided @var{x}, @var{y} coordinates.
23044 It accepts the following options:
23049 Specify an expression for the size of the output image with the
23050 paddings added. If the value for @var{width} or @var{height} is 0, the
23051 corresponding input size is used for the output.
23053 The @var{width} expression can reference the value set by the
23054 @var{height} expression, and vice versa.
23056 The default value of @var{width} and @var{height} is 0.
23060 Specify the offsets to place the input image at within the padded area,
23061 with respect to the top/left border of the output image.
23063 The @var{x} expression can reference the value set by the @var{y}
23064 expression, and vice versa.
23066 The default value of @var{x} and @var{y} is 0.
23068 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
23069 so the input image is centered on the padded area.
23072 Specify the color of the padded area. For the syntax of this option,
23073 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
23074 manual,ffmpeg-utils}.
23077 Pad to an aspect instead to a resolution.
23080 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
23081 options are expressions containing the following constants:
23086 The input video width and height.
23090 These are the same as @var{in_w} and @var{in_h}.
23094 The output width and height (the size of the padded area), as
23095 specified by the @var{width} and @var{height} expressions.
23099 These are the same as @var{out_w} and @var{out_h}.
23103 The x and y offsets as specified by the @var{x} and @var{y}
23104 expressions, or NAN if not yet specified.
23107 same as @var{iw} / @var{ih}
23110 input sample aspect ratio
23113 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
23116 @section prewitt_opencl
23118 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
23120 The filter accepts the following option:
23124 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
23127 Set value which will be multiplied with filtered result.
23128 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
23131 Set value which will be added to filtered result.
23132 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
23135 @subsection Example
23139 Apply the Prewitt operator with scale set to 2 and delta set to 10.
23141 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
23145 @anchor{program_opencl}
23146 @section program_opencl
23148 Filter video using an OpenCL program.
23153 OpenCL program source file.
23156 Kernel name in program.
23159 Number of inputs to the filter. Defaults to 1.
23162 Size of output frames. Defaults to the same as the first input.
23166 The @code{program_opencl} filter also supports the @ref{framesync} options.
23168 The program source file must contain a kernel function with the given name,
23169 which will be run once for each plane of the output. Each run on a plane
23170 gets enqueued as a separate 2D global NDRange with one work-item for each
23171 pixel to be generated. The global ID offset for each work-item is therefore
23172 the coordinates of a pixel in the destination image.
23174 The kernel function needs to take the following arguments:
23177 Destination image, @var{__write_only image2d_t}.
23179 This image will become the output; the kernel should write all of it.
23181 Frame index, @var{unsigned int}.
23183 This is a counter starting from zero and increasing by one for each frame.
23185 Source images, @var{__read_only image2d_t}.
23187 These are the most recent images on each input. The kernel may read from
23188 them to generate the output, but they can't be written to.
23195 Copy the input to the output (output must be the same size as the input).
23197 __kernel void copy(__write_only image2d_t destination,
23198 unsigned int index,
23199 __read_only image2d_t source)
23201 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
23203 int2 location = (int2)(get_global_id(0), get_global_id(1));
23205 float4 value = read_imagef(source, sampler, location);
23207 write_imagef(destination, location, value);
23212 Apply a simple transformation, rotating the input by an amount increasing
23213 with the index counter. Pixel values are linearly interpolated by the
23214 sampler, and the output need not have the same dimensions as the input.
23216 __kernel void rotate_image(__write_only image2d_t dst,
23217 unsigned int index,
23218 __read_only image2d_t src)
23220 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23221 CLK_FILTER_LINEAR);
23223 float angle = (float)index / 100.0f;
23225 float2 dst_dim = convert_float2(get_image_dim(dst));
23226 float2 src_dim = convert_float2(get_image_dim(src));
23228 float2 dst_cen = dst_dim / 2.0f;
23229 float2 src_cen = src_dim / 2.0f;
23231 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
23233 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
23235 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
23236 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
23238 src_pos = src_pos * src_dim / dst_dim;
23240 float2 src_loc = src_pos + src_cen;
23242 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
23243 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
23244 write_imagef(dst, dst_loc, 0.5f);
23246 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
23251 Blend two inputs together, with the amount of each input used varying
23252 with the index counter.
23254 __kernel void blend_images(__write_only image2d_t dst,
23255 unsigned int index,
23256 __read_only image2d_t src1,
23257 __read_only image2d_t src2)
23259 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23260 CLK_FILTER_LINEAR);
23262 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
23264 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
23265 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
23266 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
23268 float4 val1 = read_imagef(src1, sampler, src1_loc);
23269 float4 val2 = read_imagef(src2, sampler, src2_loc);
23271 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
23277 @section roberts_opencl
23278 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
23280 The filter accepts the following option:
23284 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
23287 Set value which will be multiplied with filtered result.
23288 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
23291 Set value which will be added to filtered result.
23292 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
23295 @subsection Example
23299 Apply the Roberts cross operator with scale set to 2 and delta set to 10
23301 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
23305 @section sobel_opencl
23307 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
23309 The filter accepts the following option:
23313 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
23316 Set value which will be multiplied with filtered result.
23317 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
23320 Set value which will be added to filtered result.
23321 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
23324 @subsection Example
23328 Apply sobel operator with scale set to 2 and delta set to 10
23330 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
23334 @section tonemap_opencl
23336 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
23338 It accepts the following parameters:
23342 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
23345 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
23348 Apply desaturation for highlights that exceed this level of brightness. The
23349 higher the parameter, the more color information will be preserved. This
23350 setting helps prevent unnaturally blown-out colors for super-highlights, by
23351 (smoothly) turning into white instead. This makes images feel more natural,
23352 at the cost of reducing information about out-of-range colors.
23354 The default value is 0.5, and the algorithm here is a little different from
23355 the cpu version tonemap currently. A setting of 0.0 disables this option.
23358 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
23359 is used to detect whether the scene has changed or not. If the distance between
23360 the current frame average brightness and the current running average exceeds
23361 a threshold value, we would re-calculate scene average and peak brightness.
23362 The default value is 0.2.
23365 Specify the output pixel format.
23367 Currently supported formats are:
23374 Set the output color range.
23376 Possible values are:
23382 Default is same as input.
23385 Set the output color primaries.
23387 Possible values are:
23393 Default is same as input.
23396 Set the output transfer characteristics.
23398 Possible values are:
23407 Set the output colorspace matrix.
23409 Possible value are:
23415 Default is same as input.
23419 @subsection Example
23423 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
23425 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
23429 @section unsharp_opencl
23431 Sharpen or blur the input video.
23433 It accepts the following parameters:
23436 @item luma_msize_x, lx
23437 Set the luma matrix horizontal size.
23438 Range is @code{[1, 23]} and default value is @code{5}.
23440 @item luma_msize_y, ly
23441 Set the luma matrix vertical size.
23442 Range is @code{[1, 23]} and default value is @code{5}.
23444 @item luma_amount, la
23445 Set the luma effect strength.
23446 Range is @code{[-10, 10]} and default value is @code{1.0}.
23448 Negative values will blur the input video, while positive values will
23449 sharpen it, a value of zero will disable the effect.
23451 @item chroma_msize_x, cx
23452 Set the chroma matrix horizontal size.
23453 Range is @code{[1, 23]} and default value is @code{5}.
23455 @item chroma_msize_y, cy
23456 Set the chroma matrix vertical size.
23457 Range is @code{[1, 23]} and default value is @code{5}.
23459 @item chroma_amount, ca
23460 Set the chroma effect strength.
23461 Range is @code{[-10, 10]} and default value is @code{0.0}.
23463 Negative values will blur the input video, while positive values will
23464 sharpen it, a value of zero will disable the effect.
23468 All parameters are optional and default to the equivalent of the
23469 string '5:5:1.0:5:5:0.0'.
23471 @subsection Examples
23475 Apply strong luma sharpen effect:
23477 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
23481 Apply a strong blur of both luma and chroma parameters:
23483 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
23487 @section xfade_opencl
23489 Cross fade two videos with custom transition effect by using OpenCL.
23491 It accepts the following options:
23495 Set one of possible transition effects.
23499 Select custom transition effect, the actual transition description
23500 will be picked from source and kernel options.
23512 Default transition is fade.
23516 OpenCL program source file for custom transition.
23519 Set name of kernel to use for custom transition from program source file.
23522 Set duration of video transition.
23525 Set time of start of transition relative to first video.
23528 The program source file must contain a kernel function with the given name,
23529 which will be run once for each plane of the output. Each run on a plane
23530 gets enqueued as a separate 2D global NDRange with one work-item for each
23531 pixel to be generated. The global ID offset for each work-item is therefore
23532 the coordinates of a pixel in the destination image.
23534 The kernel function needs to take the following arguments:
23537 Destination image, @var{__write_only image2d_t}.
23539 This image will become the output; the kernel should write all of it.
23542 First Source image, @var{__read_only image2d_t}.
23543 Second Source image, @var{__read_only image2d_t}.
23545 These are the most recent images on each input. The kernel may read from
23546 them to generate the output, but they can't be written to.
23549 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
23556 Apply dots curtain transition effect:
23558 __kernel void blend_images(__write_only image2d_t dst,
23559 __read_only image2d_t src1,
23560 __read_only image2d_t src2,
23563 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23564 CLK_FILTER_LINEAR);
23565 int2 p = (int2)(get_global_id(0), get_global_id(1));
23566 float2 rp = (float2)(get_global_id(0), get_global_id(1));
23567 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
23570 float2 dots = (float2)(20.0, 20.0);
23571 float2 center = (float2)(0,0);
23574 float4 val1 = read_imagef(src1, sampler, p);
23575 float4 val2 = read_imagef(src2, sampler, p);
23576 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
23578 write_imagef(dst, p, next ? val1 : val2);
23584 @c man end OPENCL VIDEO FILTERS
23586 @chapter VAAPI Video Filters
23587 @c man begin VAAPI VIDEO FILTERS
23589 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
23591 To enable compilation of these filters you need to configure FFmpeg with
23592 @code{--enable-vaapi}.
23594 To use vaapi filters, you need to setup the vaapi device correctly. For more information, please read @url{https://trac.ffmpeg.org/wiki/Hardware/VAAPI}
23596 @section tonemap_vaapi
23598 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
23599 It maps the dynamic range of HDR10 content to the SDR content.
23600 It currently only accepts HDR10 as input.
23602 It accepts the following parameters:
23606 Specify the output pixel format.
23608 Currently supported formats are:
23617 Set the output color primaries.
23619 Default is same as input.
23622 Set the output transfer characteristics.
23627 Set the output colorspace matrix.
23629 Default is same as input.
23633 @subsection Example
23637 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
23639 tonemap_vaapi=format=p010:t=bt2020-10
23643 @c man end VAAPI VIDEO FILTERS
23645 @chapter Video Sources
23646 @c man begin VIDEO SOURCES
23648 Below is a description of the currently available video sources.
23652 Buffer video frames, and make them available to the filter chain.
23654 This source is mainly intended for a programmatic use, in particular
23655 through the interface defined in @file{libavfilter/buffersrc.h}.
23657 It accepts the following parameters:
23662 Specify the size (width and height) of the buffered video frames. For the
23663 syntax of this option, check the
23664 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23667 The input video width.
23670 The input video height.
23673 A string representing the pixel format of the buffered video frames.
23674 It may be a number corresponding to a pixel format, or a pixel format
23678 Specify the timebase assumed by the timestamps of the buffered frames.
23681 Specify the frame rate expected for the video stream.
23683 @item pixel_aspect, sar
23684 The sample (pixel) aspect ratio of the input video.
23687 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
23688 to the filtergraph description to specify swscale flags for automatically
23689 inserted scalers. See @ref{Filtergraph syntax}.
23691 @item hw_frames_ctx
23692 When using a hardware pixel format, this should be a reference to an
23693 AVHWFramesContext describing input frames.
23698 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23701 will instruct the source to accept video frames with size 320x240 and
23702 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23703 square pixels (1:1 sample aspect ratio).
23704 Since the pixel format with name "yuv410p" corresponds to the number 6
23705 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23706 this example corresponds to:
23708 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23711 Alternatively, the options can be specified as a flat string, but this
23712 syntax is deprecated:
23714 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23718 Create a pattern generated by an elementary cellular automaton.
23720 The initial state of the cellular automaton can be defined through the
23721 @option{filename} and @option{pattern} options. If such options are
23722 not specified an initial state is created randomly.
23724 At each new frame a new row in the video is filled with the result of
23725 the cellular automaton next generation. The behavior when the whole
23726 frame is filled is defined by the @option{scroll} option.
23728 This source accepts the following options:
23732 Read the initial cellular automaton state, i.e. the starting row, from
23733 the specified file.
23734 In the file, each non-whitespace character is considered an alive
23735 cell, a newline will terminate the row, and further characters in the
23736 file will be ignored.
23739 Read the initial cellular automaton state, i.e. the starting row, from
23740 the specified string.
23742 Each non-whitespace character in the string is considered an alive
23743 cell, a newline will terminate the row, and further characters in the
23744 string will be ignored.
23747 Set the video rate, that is the number of frames generated per second.
23750 @item random_fill_ratio, ratio
23751 Set the random fill ratio for the initial cellular automaton row. It
23752 is a floating point number value ranging from 0 to 1, defaults to
23755 This option is ignored when a file or a pattern is specified.
23757 @item random_seed, seed
23758 Set the seed for filling randomly the initial row, must be an integer
23759 included between 0 and UINT32_MAX. If not specified, or if explicitly
23760 set to -1, the filter will try to use a good random seed on a best
23764 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23765 Default value is 110.
23768 Set the size of the output video. For the syntax of this option, check the
23769 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23771 If @option{filename} or @option{pattern} is specified, the size is set
23772 by default to the width of the specified initial state row, and the
23773 height is set to @var{width} * PHI.
23775 If @option{size} is set, it must contain the width of the specified
23776 pattern string, and the specified pattern will be centered in the
23779 If a filename or a pattern string is not specified, the size value
23780 defaults to "320x518" (used for a randomly generated initial state).
23783 If set to 1, scroll the output upward when all the rows in the output
23784 have been already filled. If set to 0, the new generated row will be
23785 written over the top row just after the bottom row is filled.
23788 @item start_full, full
23789 If set to 1, completely fill the output with generated rows before
23790 outputting the first frame.
23791 This is the default behavior, for disabling set the value to 0.
23794 If set to 1, stitch the left and right row edges together.
23795 This is the default behavior, for disabling set the value to 0.
23798 @subsection Examples
23802 Read the initial state from @file{pattern}, and specify an output of
23805 cellauto=f=pattern:s=200x400
23809 Generate a random initial row with a width of 200 cells, with a fill
23812 cellauto=ratio=2/3:s=200x200
23816 Create a pattern generated by rule 18 starting by a single alive cell
23817 centered on an initial row with width 100:
23819 cellauto=p=@@:s=100x400:full=0:rule=18
23823 Specify a more elaborated initial pattern:
23825 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23830 @anchor{coreimagesrc}
23831 @section coreimagesrc
23832 Video source generated on GPU using Apple's CoreImage API on OSX.
23834 This video source is a specialized version of the @ref{coreimage} video filter.
23835 Use a core image generator at the beginning of the applied filterchain to
23836 generate the content.
23838 The coreimagesrc video source accepts the following options:
23840 @item list_generators
23841 List all available generators along with all their respective options as well as
23842 possible minimum and maximum values along with the default values.
23844 list_generators=true
23848 Specify the size of the sourced video. For the syntax of this option, check the
23849 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23850 The default value is @code{320x240}.
23853 Specify the frame rate of the sourced video, as the number of frames
23854 generated per second. It has to be a string in the format
23855 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23856 number or a valid video frame rate abbreviation. The default value is
23860 Set the sample aspect ratio of the sourced video.
23863 Set the duration of the sourced video. See
23864 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23865 for the accepted syntax.
23867 If not specified, or the expressed duration is negative, the video is
23868 supposed to be generated forever.
23871 Additionally, all options of the @ref{coreimage} video filter are accepted.
23872 A complete filterchain can be used for further processing of the
23873 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23874 and examples for details.
23876 @subsection Examples
23881 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23882 given as complete and escaped command-line for Apple's standard bash shell:
23884 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23886 This example is equivalent to the QRCode example of @ref{coreimage} without the
23887 need for a nullsrc video source.
23892 Generate several gradients.
23896 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23897 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23900 Set frame rate, expressed as number of frames per second. Default
23903 @item c0, c1, c2, c3, c4, c5, c6, c7
23904 Set 8 colors. Default values for colors is to pick random one.
23906 @item x0, y0, y0, y1
23907 Set gradient line source and destination points. If negative or out of range, random ones
23911 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23914 Set seed for picking gradient line points.
23917 Set the duration of the sourced video. See
23918 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23919 for the accepted syntax.
23921 If not specified, or the expressed duration is negative, the video is
23922 supposed to be generated forever.
23925 Set speed of gradients rotation.
23929 @section mandelbrot
23931 Generate a Mandelbrot set fractal, and progressively zoom towards the
23932 point specified with @var{start_x} and @var{start_y}.
23934 This source accepts the following options:
23939 Set the terminal pts value. Default value is 400.
23942 Set the terminal scale value.
23943 Must be a floating point value. Default value is 0.3.
23946 Set the inner coloring mode, that is the algorithm used to draw the
23947 Mandelbrot fractal internal region.
23949 It shall assume one of the following values:
23954 Show time until convergence.
23956 Set color based on point closest to the origin of the iterations.
23961 Default value is @var{mincol}.
23964 Set the bailout value. Default value is 10.0.
23967 Set the maximum of iterations performed by the rendering
23968 algorithm. Default value is 7189.
23971 Set outer coloring mode.
23972 It shall assume one of following values:
23974 @item iteration_count
23975 Set iteration count mode.
23976 @item normalized_iteration_count
23977 set normalized iteration count mode.
23979 Default value is @var{normalized_iteration_count}.
23982 Set frame rate, expressed as number of frames per second. Default
23986 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23987 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23990 Set the initial scale value. Default value is 3.0.
23993 Set the initial x position. Must be a floating point value between
23994 -100 and 100. Default value is -0.743643887037158704752191506114774.
23997 Set the initial y position. Must be a floating point value between
23998 -100 and 100. Default value is -0.131825904205311970493132056385139.
24003 Generate various test patterns, as generated by the MPlayer test filter.
24005 The size of the generated video is fixed, and is 256x256.
24006 This source is useful in particular for testing encoding features.
24008 This source accepts the following options:
24013 Specify the frame rate of the sourced video, as the number of frames
24014 generated per second. It has to be a string in the format
24015 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
24016 number or a valid video frame rate abbreviation. The default value is
24020 Set the duration of the sourced video. See
24021 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
24022 for the accepted syntax.
24024 If not specified, or the expressed duration is negative, the video is
24025 supposed to be generated forever.
24029 Set the number or the name of the test to perform. Supported tests are:
24043 @item max_frames, m
24044 Set the maximum number of frames generated for each test, default value is 30.
24048 Default value is "all", which will cycle through the list of all tests.
24053 mptestsrc=t=dc_luma
24056 will generate a "dc_luma" test pattern.
24058 @section frei0r_src
24060 Provide a frei0r source.
24062 To enable compilation of this filter you need to install the frei0r
24063 header and configure FFmpeg with @code{--enable-frei0r}.
24065 This source accepts the following parameters:
24070 The size of the video to generate. For the syntax of this option, check the
24071 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24074 The framerate of the generated video. It may be a string of the form
24075 @var{num}/@var{den} or a frame rate abbreviation.
24078 The name to the frei0r source to load. For more information regarding frei0r and
24079 how to set the parameters, read the @ref{frei0r} section in the video filters
24082 @item filter_params
24083 A '|'-separated list of parameters to pass to the frei0r source.
24087 For example, to generate a frei0r partik0l source with size 200x200
24088 and frame rate 10 which is overlaid on the overlay filter main input:
24090 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
24095 Generate a life pattern.
24097 This source is based on a generalization of John Conway's life game.
24099 The sourced input represents a life grid, each pixel represents a cell
24100 which can be in one of two possible states, alive or dead. Every cell
24101 interacts with its eight neighbours, which are the cells that are
24102 horizontally, vertically, or diagonally adjacent.
24104 At each interaction the grid evolves according to the adopted rule,
24105 which specifies the number of neighbor alive cells which will make a
24106 cell stay alive or born. The @option{rule} option allows one to specify
24109 This source accepts the following options:
24113 Set the file from which to read the initial grid state. In the file,
24114 each non-whitespace character is considered an alive cell, and newline
24115 is used to delimit the end of each row.
24117 If this option is not specified, the initial grid is generated
24121 Set the video rate, that is the number of frames generated per second.
24124 @item random_fill_ratio, ratio
24125 Set the random fill ratio for the initial random grid. It is a
24126 floating point number value ranging from 0 to 1, defaults to 1/PHI.
24127 It is ignored when a file is specified.
24129 @item random_seed, seed
24130 Set the seed for filling the initial random grid, must be an integer
24131 included between 0 and UINT32_MAX. If not specified, or if explicitly
24132 set to -1, the filter will try to use a good random seed on a best
24138 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
24139 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
24140 @var{NS} specifies the number of alive neighbor cells which make a
24141 live cell stay alive, and @var{NB} the number of alive neighbor cells
24142 which make a dead cell to become alive (i.e. to "born").
24143 "s" and "b" can be used in place of "S" and "B", respectively.
24145 Alternatively a rule can be specified by an 18-bits integer. The 9
24146 high order bits are used to encode the next cell state if it is alive
24147 for each number of neighbor alive cells, the low order bits specify
24148 the rule for "borning" new cells. Higher order bits encode for an
24149 higher number of neighbor cells.
24150 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
24151 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
24153 Default value is "S23/B3", which is the original Conway's game of life
24154 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
24155 cells, and will born a new cell if there are three alive cells around
24159 Set the size of the output video. For the syntax of this option, check the
24160 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24162 If @option{filename} is specified, the size is set by default to the
24163 same size of the input file. If @option{size} is set, it must contain
24164 the size specified in the input file, and the initial grid defined in
24165 that file is centered in the larger resulting area.
24167 If a filename is not specified, the size value defaults to "320x240"
24168 (used for a randomly generated initial grid).
24171 If set to 1, stitch the left and right grid edges together, and the
24172 top and bottom edges also. Defaults to 1.
24175 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
24176 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
24177 value from 0 to 255.
24180 Set the color of living (or new born) cells.
24183 Set the color of dead cells. If @option{mold} is set, this is the first color
24184 used to represent a dead cell.
24187 Set mold color, for definitely dead and moldy cells.
24189 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
24190 ffmpeg-utils manual,ffmpeg-utils}.
24193 @subsection Examples
24197 Read a grid from @file{pattern}, and center it on a grid of size
24200 life=f=pattern:s=300x300
24204 Generate a random grid of size 200x200, with a fill ratio of 2/3:
24206 life=ratio=2/3:s=200x200
24210 Specify a custom rule for evolving a randomly generated grid:
24216 Full example with slow death effect (mold) using @command{ffplay}:
24218 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
24225 @anchor{haldclutsrc}
24228 @anchor{pal100bars}
24229 @anchor{rgbtestsrc}
24231 @anchor{smptehdbars}
24234 @anchor{yuvtestsrc}
24235 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
24237 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
24239 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
24241 The @code{color} source provides an uniformly colored input.
24243 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
24244 @ref{haldclut} filter.
24246 The @code{nullsrc} source returns unprocessed video frames. It is
24247 mainly useful to be employed in analysis / debugging tools, or as the
24248 source for filters which ignore the input data.
24250 The @code{pal75bars} source generates a color bars pattern, based on
24251 EBU PAL recommendations with 75% color levels.
24253 The @code{pal100bars} source generates a color bars pattern, based on
24254 EBU PAL recommendations with 100% color levels.
24256 The @code{rgbtestsrc} source generates an RGB test pattern useful for
24257 detecting RGB vs BGR issues. You should see a red, green and blue
24258 stripe from top to bottom.
24260 The @code{smptebars} source generates a color bars pattern, based on
24261 the SMPTE Engineering Guideline EG 1-1990.
24263 The @code{smptehdbars} source generates a color bars pattern, based on
24264 the SMPTE RP 219-2002.
24266 The @code{testsrc} source generates a test video pattern, showing a
24267 color pattern, a scrolling gradient and a timestamp. This is mainly
24268 intended for testing purposes.
24270 The @code{testsrc2} source is similar to testsrc, but supports more
24271 pixel formats instead of just @code{rgb24}. This allows using it as an
24272 input for other tests without requiring a format conversion.
24274 The @code{yuvtestsrc} source generates an YUV test pattern. You should
24275 see a y, cb and cr stripe from top to bottom.
24277 The sources accept the following parameters:
24282 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
24283 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
24284 pixels to be used as identity matrix for 3D lookup tables. Each component is
24285 coded on a @code{1/(N*N)} scale.
24288 Specify the color of the source, only available in the @code{color}
24289 source. For the syntax of this option, check the
24290 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
24293 Specify the size of the sourced video. For the syntax of this option, check the
24294 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24295 The default value is @code{320x240}.
24297 This option is not available with the @code{allrgb}, @code{allyuv}, and
24298 @code{haldclutsrc} filters.
24301 Specify the frame rate of the sourced video, as the number of frames
24302 generated per second. It has to be a string in the format
24303 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
24304 number or a valid video frame rate abbreviation. The default value is
24308 Set the duration of the sourced video. See
24309 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
24310 for the accepted syntax.
24312 If not specified, or the expressed duration is negative, the video is
24313 supposed to be generated forever.
24315 Since the frame rate is used as time base, all frames including the last one
24316 will have their full duration. If the specified duration is not a multiple
24317 of the frame duration, it will be rounded up.
24320 Set the sample aspect ratio of the sourced video.
24323 Specify the alpha (opacity) of the background, only available in the
24324 @code{testsrc2} source. The value must be between 0 (fully transparent) and
24325 255 (fully opaque, the default).
24328 Set the number of decimals to show in the timestamp, only available in the
24329 @code{testsrc} source.
24331 The displayed timestamp value will correspond to the original
24332 timestamp value multiplied by the power of 10 of the specified
24333 value. Default value is 0.
24336 @subsection Examples
24340 Generate a video with a duration of 5.3 seconds, with size
24341 176x144 and a frame rate of 10 frames per second:
24343 testsrc=duration=5.3:size=qcif:rate=10
24347 The following graph description will generate a red source
24348 with an opacity of 0.2, with size "qcif" and a frame rate of 10
24351 color=c=red@@0.2:s=qcif:r=10
24355 If the input content is to be ignored, @code{nullsrc} can be used. The
24356 following command generates noise in the luminance plane by employing
24357 the @code{geq} filter:
24359 nullsrc=s=256x256, geq=random(1)*255:128:128
24363 @subsection Commands
24365 The @code{color} source supports the following commands:
24369 Set the color of the created image. Accepts the same syntax of the
24370 corresponding @option{color} option.
24375 Generate video using an OpenCL program.
24380 OpenCL program source file.
24383 Kernel name in program.
24386 Size of frames to generate. This must be set.
24389 Pixel format to use for the generated frames. This must be set.
24392 Number of frames generated every second. Default value is '25'.
24396 For details of how the program loading works, see the @ref{program_opencl}
24403 Generate a colour ramp by setting pixel values from the position of the pixel
24404 in the output image. (Note that this will work with all pixel formats, but
24405 the generated output will not be the same.)
24407 __kernel void ramp(__write_only image2d_t dst,
24408 unsigned int index)
24410 int2 loc = (int2)(get_global_id(0), get_global_id(1));
24413 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
24415 write_imagef(dst, loc, val);
24420 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
24422 __kernel void sierpinski_carpet(__write_only image2d_t dst,
24423 unsigned int index)
24425 int2 loc = (int2)(get_global_id(0), get_global_id(1));
24427 float4 value = 0.0f;
24428 int x = loc.x + index;
24429 int y = loc.y + index;
24430 while (x > 0 || y > 0) {
24431 if (x % 3 == 1 && y % 3 == 1) {
24439 write_imagef(dst, loc, value);
24445 @section sierpinski
24447 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
24449 This source accepts the following options:
24453 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
24454 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
24457 Set frame rate, expressed as number of frames per second. Default
24461 Set seed which is used for random panning.
24464 Set max jump for single pan destination. Allowed range is from 1 to 10000.
24467 Set fractal type, can be default @code{carpet} or @code{triangle}.
24470 @c man end VIDEO SOURCES
24472 @chapter Video Sinks
24473 @c man begin VIDEO SINKS
24475 Below is a description of the currently available video sinks.
24477 @section buffersink
24479 Buffer video frames, and make them available to the end of the filter
24482 This sink is mainly intended for programmatic use, in particular
24483 through the interface defined in @file{libavfilter/buffersink.h}
24484 or the options system.
24486 It accepts a pointer to an AVBufferSinkContext structure, which
24487 defines the incoming buffers' formats, to be passed as the opaque
24488 parameter to @code{avfilter_init_filter} for initialization.
24492 Null video sink: do absolutely nothing with the input video. It is
24493 mainly useful as a template and for use in analysis / debugging
24496 @c man end VIDEO SINKS
24498 @chapter Multimedia Filters
24499 @c man begin MULTIMEDIA FILTERS
24501 Below is a description of the currently available multimedia filters.
24505 Convert input audio to a video output, displaying the audio bit scope.
24507 The filter accepts the following options:
24511 Set frame rate, expressed as number of frames per second. Default
24515 Specify the video size for the output. For the syntax of this option, check the
24516 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24517 Default value is @code{1024x256}.
24520 Specify list of colors separated by space or by '|' which will be used to
24521 draw channels. Unrecognized or missing colors will be replaced
24525 @section adrawgraph
24526 Draw a graph using input audio metadata.
24528 See @ref{drawgraph}
24530 @section agraphmonitor
24532 See @ref{graphmonitor}.
24534 @section ahistogram
24536 Convert input audio to a video output, displaying the volume histogram.
24538 The filter accepts the following options:
24542 Specify how histogram is calculated.
24544 It accepts the following values:
24547 Use single histogram for all channels.
24549 Use separate histogram for each channel.
24551 Default is @code{single}.
24554 Set frame rate, expressed as number of frames per second. Default
24558 Specify the video size for the output. For the syntax of this option, check the
24559 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24560 Default value is @code{hd720}.
24565 It accepts the following values:
24576 reverse logarithmic
24578 Default is @code{log}.
24581 Set amplitude scale.
24583 It accepts the following values:
24590 Default is @code{log}.
24593 Set how much frames to accumulate in histogram.
24594 Default is 1. Setting this to -1 accumulates all frames.
24597 Set histogram ratio of window height.
24600 Set sonogram sliding.
24602 It accepts the following values:
24605 replace old rows with new ones.
24607 scroll from top to bottom.
24609 Default is @code{replace}.
24612 @section aphasemeter
24614 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
24615 representing mean phase of current audio frame. A video output can also be produced and is
24616 enabled by default. The audio is passed through as first output.
24618 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
24619 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
24620 and @code{1} means channels are in phase.
24622 The filter accepts the following options, all related to its video output:
24626 Set the output frame rate. Default value is @code{25}.
24629 Set the video size for the output. For the syntax of this option, check the
24630 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24631 Default value is @code{800x400}.
24636 Specify the red, green, blue contrast. Default values are @code{2},
24637 @code{7} and @code{1}.
24638 Allowed range is @code{[0, 255]}.
24641 Set color which will be used for drawing median phase. If color is
24642 @code{none} which is default, no median phase value will be drawn.
24645 Enable video output. Default is enabled.
24648 @subsection phasing detection
24650 The filter also detects out of phase and mono sequences in stereo streams.
24651 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
24653 The filter accepts the following options for this detection:
24657 Enable mono and out of phase detection. Default is disabled.
24660 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
24661 Allowed range is @code{[0, 1]}.
24664 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
24665 Allowed range is @code{[90, 180]}.
24668 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
24671 @subsection Examples
24675 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
24677 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
24681 @section avectorscope
24683 Convert input audio to a video output, representing the audio vector
24686 The filter is used to measure the difference between channels of stereo
24687 audio stream. A monaural signal, consisting of identical left and right
24688 signal, results in straight vertical line. Any stereo separation is visible
24689 as a deviation from this line, creating a Lissajous figure.
24690 If the straight (or deviation from it) but horizontal line appears this
24691 indicates that the left and right channels are out of phase.
24693 The filter accepts the following options:
24697 Set the vectorscope mode.
24699 Available values are:
24702 Lissajous rotated by 45 degrees.
24705 Same as above but not rotated.
24708 Shape resembling half of circle.
24711 Default value is @samp{lissajous}.
24714 Set the video size for the output. For the syntax of this option, check the
24715 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24716 Default value is @code{400x400}.
24719 Set the output frame rate. Default value is @code{25}.
24725 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24726 @code{160}, @code{80} and @code{255}.
24727 Allowed range is @code{[0, 255]}.
24733 Specify the red, green, blue and alpha fade. Default values are @code{15},
24734 @code{10}, @code{5} and @code{5}.
24735 Allowed range is @code{[0, 255]}.
24738 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24739 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24742 Set the vectorscope drawing mode.
24744 Available values are:
24747 Draw dot for each sample.
24750 Draw line between previous and current sample.
24753 Default value is @samp{dot}.
24756 Specify amplitude scale of audio samples.
24758 Available values are:
24774 Swap left channel axis with right channel axis.
24784 Mirror only x axis.
24787 Mirror only y axis.
24795 @subsection Examples
24799 Complete example using @command{ffplay}:
24801 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24802 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24806 @section bench, abench
24808 Benchmark part of a filtergraph.
24810 The filter accepts the following options:
24814 Start or stop a timer.
24816 Available values are:
24819 Get the current time, set it as frame metadata (using the key
24820 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24823 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24824 the input frame metadata to get the time difference. Time difference, average,
24825 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24826 @code{min}) are then printed. The timestamps are expressed in seconds.
24830 @subsection Examples
24834 Benchmark @ref{selectivecolor} filter:
24836 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24842 Concatenate audio and video streams, joining them together one after the
24845 The filter works on segments of synchronized video and audio streams. All
24846 segments must have the same number of streams of each type, and that will
24847 also be the number of streams at output.
24849 The filter accepts the following options:
24854 Set the number of segments. Default is 2.
24857 Set the number of output video streams, that is also the number of video
24858 streams in each segment. Default is 1.
24861 Set the number of output audio streams, that is also the number of audio
24862 streams in each segment. Default is 0.
24865 Activate unsafe mode: do not fail if segments have a different format.
24869 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24870 @var{a} audio outputs.
24872 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24873 segment, in the same order as the outputs, then the inputs for the second
24876 Related streams do not always have exactly the same duration, for various
24877 reasons including codec frame size or sloppy authoring. For that reason,
24878 related synchronized streams (e.g. a video and its audio track) should be
24879 concatenated at once. The concat filter will use the duration of the longest
24880 stream in each segment (except the last one), and if necessary pad shorter
24881 audio streams with silence.
24883 For this filter to work correctly, all segments must start at timestamp 0.
24885 All corresponding streams must have the same parameters in all segments; the
24886 filtering system will automatically select a common pixel format for video
24887 streams, and a common sample format, sample rate and channel layout for
24888 audio streams, but other settings, such as resolution, must be converted
24889 explicitly by the user.
24891 Different frame rates are acceptable but will result in variable frame rate
24892 at output; be sure to configure the output file to handle it.
24894 @subsection Examples
24898 Concatenate an opening, an episode and an ending, all in bilingual version
24899 (video in stream 0, audio in streams 1 and 2):
24901 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24902 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24903 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24904 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24908 Concatenate two parts, handling audio and video separately, using the
24909 (a)movie sources, and adjusting the resolution:
24911 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24912 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24913 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24915 Note that a desync will happen at the stitch if the audio and video streams
24916 do not have exactly the same duration in the first file.
24920 @subsection Commands
24922 This filter supports the following commands:
24925 Close the current segment and step to the next one
24931 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24932 level. By default, it logs a message at a frequency of 10Hz with the
24933 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24934 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24936 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24937 sample format is double-precision floating point. The input stream will be converted to
24938 this specification, if needed. Users may need to insert aformat and/or aresample filters
24939 after this filter to obtain the original parameters.
24941 The filter also has a video output (see the @var{video} option) with a real
24942 time graph to observe the loudness evolution. The graphic contains the logged
24943 message mentioned above, so it is not printed anymore when this option is set,
24944 unless the verbose logging is set. The main graphing area contains the
24945 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24946 the momentary loudness (400 milliseconds), but can optionally be configured
24947 to instead display short-term loudness (see @var{gauge}).
24949 The green area marks a +/- 1LU target range around the target loudness
24950 (-23LUFS by default, unless modified through @var{target}).
24952 More information about the Loudness Recommendation EBU R128 on
24953 @url{http://tech.ebu.ch/loudness}.
24955 The filter accepts the following options:
24960 Activate the video output. The audio stream is passed unchanged whether this
24961 option is set or no. The video stream will be the first output stream if
24962 activated. Default is @code{0}.
24965 Set the video size. This option is for video only. For the syntax of this
24967 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24968 Default and minimum resolution is @code{640x480}.
24971 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24972 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24973 other integer value between this range is allowed.
24976 Set metadata injection. If set to @code{1}, the audio input will be segmented
24977 into 100ms output frames, each of them containing various loudness information
24978 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24980 Default is @code{0}.
24983 Force the frame logging level.
24985 Available values are:
24988 information logging level
24990 verbose logging level
24993 By default, the logging level is set to @var{info}. If the @option{video} or
24994 the @option{metadata} options are set, it switches to @var{verbose}.
24999 Available modes can be cumulated (the option is a @code{flag} type). Possible
25003 Disable any peak mode (default).
25005 Enable sample-peak mode.
25007 Simple peak mode looking for the higher sample value. It logs a message
25008 for sample-peak (identified by @code{SPK}).
25010 Enable true-peak mode.
25012 If enabled, the peak lookup is done on an over-sampled version of the input
25013 stream for better peak accuracy. It logs a message for true-peak.
25014 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
25015 This mode requires a build with @code{libswresample}.
25019 Treat mono input files as "dual mono". If a mono file is intended for playback
25020 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
25021 If set to @code{true}, this option will compensate for this effect.
25022 Multi-channel input files are not affected by this option.
25025 Set a specific pan law to be used for the measurement of dual mono files.
25026 This parameter is optional, and has a default value of -3.01dB.
25029 Set a specific target level (in LUFS) used as relative zero in the visualization.
25030 This parameter is optional and has a default value of -23LUFS as specified
25031 by EBU R128. However, material published online may prefer a level of -16LUFS
25032 (e.g. for use with podcasts or video platforms).
25035 Set the value displayed by the gauge. Valid values are @code{momentary} and s
25036 @code{shortterm}. By default the momentary value will be used, but in certain
25037 scenarios it may be more useful to observe the short term value instead (e.g.
25041 Sets the display scale for the loudness. Valid parameters are @code{absolute}
25042 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
25043 video output, not the summary or continuous log output.
25046 @subsection Examples
25050 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
25052 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
25056 Run an analysis with @command{ffmpeg}:
25058 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
25062 @section interleave, ainterleave
25064 Temporally interleave frames from several inputs.
25066 @code{interleave} works with video inputs, @code{ainterleave} with audio.
25068 These filters read frames from several inputs and send the oldest
25069 queued frame to the output.
25071 Input streams must have well defined, monotonically increasing frame
25074 In order to submit one frame to output, these filters need to enqueue
25075 at least one frame for each input, so they cannot work in case one
25076 input is not yet terminated and will not receive incoming frames.
25078 For example consider the case when one input is a @code{select} filter
25079 which always drops input frames. The @code{interleave} filter will keep
25080 reading from that input, but it will never be able to send new frames
25081 to output until the input sends an end-of-stream signal.
25083 Also, depending on inputs synchronization, the filters will drop
25084 frames in case one input receives more frames than the other ones, and
25085 the queue is already filled.
25087 These filters accept the following options:
25091 Set the number of different inputs, it is 2 by default.
25094 How to determine the end-of-stream.
25098 The duration of the longest input. (default)
25101 The duration of the shortest input.
25104 The duration of the first input.
25109 @subsection Examples
25113 Interleave frames belonging to different streams using @command{ffmpeg}:
25115 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
25119 Add flickering blur effect:
25121 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
25125 @section metadata, ametadata
25127 Manipulate frame metadata.
25129 This filter accepts the following options:
25133 Set mode of operation of the filter.
25135 Can be one of the following:
25139 If both @code{value} and @code{key} is set, select frames
25140 which have such metadata. If only @code{key} is set, select
25141 every frame that has such key in metadata.
25144 Add new metadata @code{key} and @code{value}. If key is already available
25148 Modify value of already present key.
25151 If @code{value} is set, delete only keys that have such value.
25152 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
25156 Print key and its value if metadata was found. If @code{key} is not set print all
25157 metadata values available in frame.
25161 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
25164 Set metadata value which will be used. This option is mandatory for
25165 @code{modify} and @code{add} mode.
25168 Which function to use when comparing metadata value and @code{value}.
25170 Can be one of following:
25174 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
25177 Values are interpreted as strings, returns true if metadata value starts with
25178 the @code{value} option string.
25181 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
25184 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
25187 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
25190 Values are interpreted as floats, returns true if expression from option @code{expr}
25194 Values are interpreted as strings, returns true if metadata value ends with
25195 the @code{value} option string.
25199 Set expression which is used when @code{function} is set to @code{expr}.
25200 The expression is evaluated through the eval API and can contain the following
25205 Float representation of @code{value} from metadata key.
25208 Float representation of @code{value} as supplied by user in @code{value} option.
25212 If specified in @code{print} mode, output is written to the named file. Instead of
25213 plain filename any writable url can be specified. Filename ``-'' is a shorthand
25214 for standard output. If @code{file} option is not set, output is written to the log
25215 with AV_LOG_INFO loglevel.
25218 Reduces buffering in print mode when output is written to a URL set using @var{file}.
25222 @subsection Examples
25226 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
25229 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
25232 Print silencedetect output to file @file{metadata.txt}.
25234 silencedetect,ametadata=mode=print:file=metadata.txt
25237 Direct all metadata to a pipe with file descriptor 4.
25239 metadata=mode=print:file='pipe\:4'
25243 @section perms, aperms
25245 Set read/write permissions for the output frames.
25247 These filters are mainly aimed at developers to test direct path in the
25248 following filter in the filtergraph.
25250 The filters accept the following options:
25254 Select the permissions mode.
25256 It accepts the following values:
25259 Do nothing. This is the default.
25261 Set all the output frames read-only.
25263 Set all the output frames directly writable.
25265 Make the frame read-only if writable, and writable if read-only.
25267 Set each output frame read-only or writable randomly.
25271 Set the seed for the @var{random} mode, must be an integer included between
25272 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
25273 @code{-1}, the filter will try to use a good random seed on a best effort
25277 Note: in case of auto-inserted filter between the permission filter and the
25278 following one, the permission might not be received as expected in that
25279 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
25280 perms/aperms filter can avoid this problem.
25282 @section realtime, arealtime
25284 Slow down filtering to match real time approximately.
25286 These filters will pause the filtering for a variable amount of time to
25287 match the output rate with the input timestamps.
25288 They are similar to the @option{re} option to @code{ffmpeg}.
25290 They accept the following options:
25294 Time limit for the pauses. Any pause longer than that will be considered
25295 a timestamp discontinuity and reset the timer. Default is 2 seconds.
25297 Speed factor for processing. The value must be a float larger than zero.
25298 Values larger than 1.0 will result in faster than realtime processing,
25299 smaller will slow processing down. The @var{limit} is automatically adapted
25300 accordingly. Default is 1.0.
25302 A processing speed faster than what is possible without these filters cannot
25307 @section select, aselect
25309 Select frames to pass in output.
25311 This filter accepts the following options:
25316 Set expression, which is evaluated for each input frame.
25318 If the expression is evaluated to zero, the frame is discarded.
25320 If the evaluation result is negative or NaN, the frame is sent to the
25321 first output; otherwise it is sent to the output with index
25322 @code{ceil(val)-1}, assuming that the input index starts from 0.
25324 For example a value of @code{1.2} corresponds to the output with index
25325 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
25328 Set the number of outputs. The output to which to send the selected
25329 frame is based on the result of the evaluation. Default value is 1.
25332 The expression can contain the following constants:
25336 The (sequential) number of the filtered frame, starting from 0.
25339 The (sequential) number of the selected frame, starting from 0.
25341 @item prev_selected_n
25342 The sequential number of the last selected frame. It's NAN if undefined.
25345 The timebase of the input timestamps.
25348 The PTS (Presentation TimeStamp) of the filtered video frame,
25349 expressed in @var{TB} units. It's NAN if undefined.
25352 The PTS of the filtered video frame,
25353 expressed in seconds. It's NAN if undefined.
25356 The PTS of the previously filtered video frame. It's NAN if undefined.
25358 @item prev_selected_pts
25359 The PTS of the last previously filtered video frame. It's NAN if undefined.
25361 @item prev_selected_t
25362 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
25365 The PTS of the first video frame in the video. It's NAN if undefined.
25368 The time of the first video frame in the video. It's NAN if undefined.
25370 @item pict_type @emph{(video only)}
25371 The type of the filtered frame. It can assume one of the following
25383 @item interlace_type @emph{(video only)}
25384 The frame interlace type. It can assume one of the following values:
25387 The frame is progressive (not interlaced).
25389 The frame is top-field-first.
25391 The frame is bottom-field-first.
25394 @item consumed_sample_n @emph{(audio only)}
25395 the number of selected samples before the current frame
25397 @item samples_n @emph{(audio only)}
25398 the number of samples in the current frame
25400 @item sample_rate @emph{(audio only)}
25401 the input sample rate
25404 This is 1 if the filtered frame is a key-frame, 0 otherwise.
25407 the position in the file of the filtered frame, -1 if the information
25408 is not available (e.g. for synthetic video)
25410 @item scene @emph{(video only)}
25411 value between 0 and 1 to indicate a new scene; a low value reflects a low
25412 probability for the current frame to introduce a new scene, while a higher
25413 value means the current frame is more likely to be one (see the example below)
25415 @item concatdec_select
25416 The concat demuxer can select only part of a concat input file by setting an
25417 inpoint and an outpoint, but the output packets may not be entirely contained
25418 in the selected interval. By using this variable, it is possible to skip frames
25419 generated by the concat demuxer which are not exactly contained in the selected
25422 This works by comparing the frame pts against the @var{lavf.concat.start_time}
25423 and the @var{lavf.concat.duration} packet metadata values which are also
25424 present in the decoded frames.
25426 The @var{concatdec_select} variable is -1 if the frame pts is at least
25427 start_time and either the duration metadata is missing or the frame pts is less
25428 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
25431 That basically means that an input frame is selected if its pts is within the
25432 interval set by the concat demuxer.
25436 The default value of the select expression is "1".
25438 @subsection Examples
25442 Select all frames in input:
25447 The example above is the same as:
25459 Select only I-frames:
25461 select='eq(pict_type\,I)'
25465 Select one frame every 100:
25467 select='not(mod(n\,100))'
25471 Select only frames contained in the 10-20 time interval:
25473 select=between(t\,10\,20)
25477 Select only I-frames contained in the 10-20 time interval:
25479 select=between(t\,10\,20)*eq(pict_type\,I)
25483 Select frames with a minimum distance of 10 seconds:
25485 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
25489 Use aselect to select only audio frames with samples number > 100:
25491 aselect='gt(samples_n\,100)'
25495 Create a mosaic of the first scenes:
25497 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
25500 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
25504 Send even and odd frames to separate outputs, and compose them:
25506 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
25510 Select useful frames from an ffconcat file which is using inpoints and
25511 outpoints but where the source files are not intra frame only.
25513 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
25517 @section sendcmd, asendcmd
25519 Send commands to filters in the filtergraph.
25521 These filters read commands to be sent to other filters in the
25524 @code{sendcmd} must be inserted between two video filters,
25525 @code{asendcmd} must be inserted between two audio filters, but apart
25526 from that they act the same way.
25528 The specification of commands can be provided in the filter arguments
25529 with the @var{commands} option, or in a file specified by the
25530 @var{filename} option.
25532 These filters accept the following options:
25535 Set the commands to be read and sent to the other filters.
25537 Set the filename of the commands to be read and sent to the other
25541 @subsection Commands syntax
25543 A commands description consists of a sequence of interval
25544 specifications, comprising a list of commands to be executed when a
25545 particular event related to that interval occurs. The occurring event
25546 is typically the current frame time entering or leaving a given time
25549 An interval is specified by the following syntax:
25551 @var{START}[-@var{END}] @var{COMMANDS};
25554 The time interval is specified by the @var{START} and @var{END} times.
25555 @var{END} is optional and defaults to the maximum time.
25557 The current frame time is considered within the specified interval if
25558 it is included in the interval [@var{START}, @var{END}), that is when
25559 the time is greater or equal to @var{START} and is lesser than
25562 @var{COMMANDS} consists of a sequence of one or more command
25563 specifications, separated by ",", relating to that interval. The
25564 syntax of a command specification is given by:
25566 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
25569 @var{FLAGS} is optional and specifies the type of events relating to
25570 the time interval which enable sending the specified command, and must
25571 be a non-null sequence of identifier flags separated by "+" or "|" and
25572 enclosed between "[" and "]".
25574 The following flags are recognized:
25577 The command is sent when the current frame timestamp enters the
25578 specified interval. In other words, the command is sent when the
25579 previous frame timestamp was not in the given interval, and the
25583 The command is sent when the current frame timestamp leaves the
25584 specified interval. In other words, the command is sent when the
25585 previous frame timestamp was in the given interval, and the
25589 The command @var{ARG} is interpreted as expression and result of
25590 expression is passed as @var{ARG}.
25592 The expression is evaluated through the eval API and can contain the following
25597 Original position in the file of the frame, or undefined if undefined
25598 for the current frame.
25601 The presentation timestamp in input.
25604 The count of the input frame for video or audio, starting from 0.
25607 The time in seconds of the current frame.
25610 The start time in seconds of the current command interval.
25613 The end time in seconds of the current command interval.
25616 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
25621 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
25624 @var{TARGET} specifies the target of the command, usually the name of
25625 the filter class or a specific filter instance name.
25627 @var{COMMAND} specifies the name of the command for the target filter.
25629 @var{ARG} is optional and specifies the optional list of argument for
25630 the given @var{COMMAND}.
25632 Between one interval specification and another, whitespaces, or
25633 sequences of characters starting with @code{#} until the end of line,
25634 are ignored and can be used to annotate comments.
25636 A simplified BNF description of the commands specification syntax
25639 @var{COMMAND_FLAG} ::= "enter" | "leave"
25640 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
25641 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
25642 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
25643 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
25644 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
25647 @subsection Examples
25651 Specify audio tempo change at second 4:
25653 asendcmd=c='4.0 atempo tempo 1.5',atempo
25657 Target a specific filter instance:
25659 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
25663 Specify a list of drawtext and hue commands in a file.
25665 # show text in the interval 5-10
25666 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
25667 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
25669 # desaturate the image in the interval 15-20
25670 15.0-20.0 [enter] hue s 0,
25671 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
25673 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
25675 # apply an exponential saturation fade-out effect, starting from time 25
25676 25 [enter] hue s exp(25-t)
25679 A filtergraph allowing to read and process the above command list
25680 stored in a file @file{test.cmd}, can be specified with:
25682 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
25687 @section setpts, asetpts
25689 Change the PTS (presentation timestamp) of the input frames.
25691 @code{setpts} works on video frames, @code{asetpts} on audio frames.
25693 This filter accepts the following options:
25698 The expression which is evaluated for each frame to construct its timestamp.
25702 The expression is evaluated through the eval API and can contain the following
25706 @item FRAME_RATE, FR
25707 frame rate, only defined for constant frame-rate video
25710 The presentation timestamp in input
25713 The count of the input frame for video or the number of consumed samples,
25714 not including the current frame for audio, starting from 0.
25716 @item NB_CONSUMED_SAMPLES
25717 The number of consumed samples, not including the current frame (only
25720 @item NB_SAMPLES, S
25721 The number of samples in the current frame (only audio)
25723 @item SAMPLE_RATE, SR
25724 The audio sample rate.
25727 The PTS of the first frame.
25730 the time in seconds of the first frame
25733 State whether the current frame is interlaced.
25736 the time in seconds of the current frame
25739 original position in the file of the frame, or undefined if undefined
25740 for the current frame
25743 The previous input PTS.
25746 previous input time in seconds
25749 The previous output PTS.
25752 previous output time in seconds
25755 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25759 The wallclock (RTC) time at the start of the movie in microseconds.
25762 The timebase of the input timestamps.
25766 @subsection Examples
25770 Start counting PTS from zero
25772 setpts=PTS-STARTPTS
25776 Apply fast motion effect:
25782 Apply slow motion effect:
25788 Set fixed rate of 25 frames per second:
25794 Set fixed rate 25 fps with some jitter:
25796 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25800 Apply an offset of 10 seconds to the input PTS:
25806 Generate timestamps from a "live source" and rebase onto the current timebase:
25808 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25812 Generate timestamps by counting samples:
25821 Force color range for the output video frame.
25823 The @code{setrange} filter marks the color range property for the
25824 output frames. It does not change the input frame, but only sets the
25825 corresponding property, which affects how the frame is treated by
25828 The filter accepts the following options:
25833 Available values are:
25837 Keep the same color range property.
25839 @item unspecified, unknown
25840 Set the color range as unspecified.
25842 @item limited, tv, mpeg
25843 Set the color range as limited.
25845 @item full, pc, jpeg
25846 Set the color range as full.
25850 @section settb, asettb
25852 Set the timebase to use for the output frames timestamps.
25853 It is mainly useful for testing timebase configuration.
25855 It accepts the following parameters:
25860 The expression which is evaluated into the output timebase.
25864 The value for @option{tb} is an arithmetic expression representing a
25865 rational. The expression can contain the constants "AVTB" (the default
25866 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25867 audio only). Default value is "intb".
25869 @subsection Examples
25873 Set the timebase to 1/25:
25879 Set the timebase to 1/10:
25885 Set the timebase to 1001/1000:
25891 Set the timebase to 2*intb:
25897 Set the default timebase value:
25904 Convert input audio to a video output representing frequency spectrum
25905 logarithmically using Brown-Puckette constant Q transform algorithm with
25906 direct frequency domain coefficient calculation (but the transform itself
25907 is not really constant Q, instead the Q factor is actually variable/clamped),
25908 with musical tone scale, from E0 to D#10.
25910 The filter accepts the following options:
25914 Specify the video size for the output. It must be even. For the syntax of this option,
25915 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25916 Default value is @code{1920x1080}.
25919 Set the output frame rate. Default value is @code{25}.
25922 Set the bargraph height. It must be even. Default value is @code{-1} which
25923 computes the bargraph height automatically.
25926 Set the axis height. It must be even. Default value is @code{-1} which computes
25927 the axis height automatically.
25930 Set the sonogram height. It must be even. Default value is @code{-1} which
25931 computes the sonogram height automatically.
25934 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25935 instead. Default value is @code{1}.
25937 @item sono_v, volume
25938 Specify the sonogram volume expression. It can contain variables:
25941 the @var{bar_v} evaluated expression
25942 @item frequency, freq, f
25943 the frequency where it is evaluated
25944 @item timeclamp, tc
25945 the value of @var{timeclamp} option
25949 @item a_weighting(f)
25950 A-weighting of equal loudness
25951 @item b_weighting(f)
25952 B-weighting of equal loudness
25953 @item c_weighting(f)
25954 C-weighting of equal loudness.
25956 Default value is @code{16}.
25958 @item bar_v, volume2
25959 Specify the bargraph volume expression. It can contain variables:
25962 the @var{sono_v} evaluated expression
25963 @item frequency, freq, f
25964 the frequency where it is evaluated
25965 @item timeclamp, tc
25966 the value of @var{timeclamp} option
25970 @item a_weighting(f)
25971 A-weighting of equal loudness
25972 @item b_weighting(f)
25973 B-weighting of equal loudness
25974 @item c_weighting(f)
25975 C-weighting of equal loudness.
25977 Default value is @code{sono_v}.
25979 @item sono_g, gamma
25980 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25981 higher gamma makes the spectrum having more range. Default value is @code{3}.
25982 Acceptable range is @code{[1, 7]}.
25984 @item bar_g, gamma2
25985 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25989 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25990 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25992 @item timeclamp, tc
25993 Specify the transform timeclamp. At low frequency, there is trade-off between
25994 accuracy in time domain and frequency domain. If timeclamp is lower,
25995 event in time domain is represented more accurately (such as fast bass drum),
25996 otherwise event in frequency domain is represented more accurately
25997 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
26000 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
26001 limits future samples by applying asymmetric windowing in time domain, useful
26002 when low latency is required. Accepted range is @code{[0, 1]}.
26005 Specify the transform base frequency. Default value is @code{20.01523126408007475},
26006 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
26009 Specify the transform end frequency. Default value is @code{20495.59681441799654},
26010 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
26013 This option is deprecated and ignored.
26016 Specify the transform length in time domain. Use this option to control accuracy
26017 trade-off between time domain and frequency domain at every frequency sample.
26018 It can contain variables:
26020 @item frequency, freq, f
26021 the frequency where it is evaluated
26022 @item timeclamp, tc
26023 the value of @var{timeclamp} option.
26025 Default value is @code{384*tc/(384+tc*f)}.
26028 Specify the transform count for every video frame. Default value is @code{6}.
26029 Acceptable range is @code{[1, 30]}.
26032 Specify the transform count for every single pixel. Default value is @code{0},
26033 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
26036 Specify font file for use with freetype to draw the axis. If not specified,
26037 use embedded font. Note that drawing with font file or embedded font is not
26038 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
26042 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
26043 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
26047 Specify font color expression. This is arithmetic expression that should return
26048 integer value 0xRRGGBB. It can contain variables:
26050 @item frequency, freq, f
26051 the frequency where it is evaluated
26052 @item timeclamp, tc
26053 the value of @var{timeclamp} option
26058 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
26059 @item r(x), g(x), b(x)
26060 red, green, and blue value of intensity x.
26062 Default value is @code{st(0, (midi(f)-59.5)/12);
26063 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
26064 r(1-ld(1)) + b(ld(1))}.
26067 Specify image file to draw the axis. This option override @var{fontfile} and
26068 @var{fontcolor} option.
26071 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
26072 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
26073 Default value is @code{1}.
26076 Set colorspace. The accepted values are:
26079 Unspecified (default)
26088 BT.470BG or BT.601-6 625
26091 SMPTE-170M or BT.601-6 525
26097 BT.2020 with non-constant luminance
26102 Set spectrogram color scheme. This is list of floating point values with format
26103 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
26104 The default is @code{1|0.5|0|0|0.5|1}.
26108 @subsection Examples
26112 Playing audio while showing the spectrum:
26114 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
26118 Same as above, but with frame rate 30 fps:
26120 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
26124 Playing at 1280x720:
26126 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
26130 Disable sonogram display:
26136 A1 and its harmonics: A1, A2, (near)E3, A3:
26138 ffplay -f lavfi 'aevalsrc=0.1*sin(2*PI*55*t)+0.1*sin(4*PI*55*t)+0.1*sin(6*PI*55*t)+0.1*sin(8*PI*55*t),
26139 asplit[a][out1]; [a] showcqt [out0]'
26143 Same as above, but with more accuracy in frequency domain:
26145 ffplay -f lavfi 'aevalsrc=0.1*sin(2*PI*55*t)+0.1*sin(4*PI*55*t)+0.1*sin(6*PI*55*t)+0.1*sin(8*PI*55*t),
26146 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
26152 bar_v=10:sono_v=bar_v*a_weighting(f)
26156 Custom gamma, now spectrum is linear to the amplitude.
26162 Custom tlength equation:
26164 tc=0.33:tlength='st(0,0.17); 384*tc / (384 / ld(0) + tc*f /(1-ld(0))) + 384*tc / (tc*f / ld(0) + 384 /(1-ld(0)))'
26168 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
26170 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
26174 Custom font using fontconfig:
26176 font='Courier New,Monospace,mono|bold'
26180 Custom frequency range with custom axis using image file:
26182 axisfile=myaxis.png:basefreq=40:endfreq=10000
26188 Convert input audio to video output representing the audio power spectrum.
26189 Audio amplitude is on Y-axis while frequency is on X-axis.
26191 The filter accepts the following options:
26195 Specify size of video. For the syntax of this option, check the
26196 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26197 Default is @code{1024x512}.
26201 This set how each frequency bin will be represented.
26203 It accepts the following values:
26209 Default is @code{bar}.
26212 Set amplitude scale.
26214 It accepts the following values:
26228 Default is @code{log}.
26231 Set frequency scale.
26233 It accepts the following values:
26242 Reverse logarithmic scale.
26244 Default is @code{lin}.
26247 Set window size. Allowed range is from 16 to 65536.
26249 Default is @code{2048}
26252 Set windowing function.
26254 It accepts the following values:
26277 Default is @code{hanning}.
26280 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26281 which means optimal overlap for selected window function will be picked.
26284 Set time averaging. Setting this to 0 will display current maximal peaks.
26285 Default is @code{1}, which means time averaging is disabled.
26288 Specify list of colors separated by space or by '|' which will be used to
26289 draw channel frequencies. Unrecognized or missing colors will be replaced
26293 Set channel display mode.
26295 It accepts the following values:
26300 Default is @code{combined}.
26303 Set minimum amplitude used in @code{log} amplitude scaler.
26306 Set data display mode.
26308 It accepts the following values:
26314 Default is @code{magnitude}.
26317 @section showspatial
26319 Convert stereo input audio to a video output, representing the spatial relationship
26320 between two channels.
26322 The filter accepts the following options:
26326 Specify the video size for the output. For the syntax of this option, check the
26327 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26328 Default value is @code{512x512}.
26331 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
26334 Set window function.
26336 It accepts the following values:
26361 Default value is @code{hann}.
26364 Set ratio of overlap window. Default value is @code{0.5}.
26365 When value is @code{1} overlap is set to recommended size for specific
26366 window function currently used.
26369 @anchor{showspectrum}
26370 @section showspectrum
26372 Convert input audio to a video output, representing the audio frequency
26375 The filter accepts the following options:
26379 Specify the video size for the output. For the syntax of this option, check the
26380 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26381 Default value is @code{640x512}.
26384 Specify how the spectrum should slide along the window.
26386 It accepts the following values:
26389 the samples start again on the left when they reach the right
26391 the samples scroll from right to left
26393 frames are only produced when the samples reach the right
26395 the samples scroll from left to right
26398 Default value is @code{replace}.
26401 Specify display mode.
26403 It accepts the following values:
26406 all channels are displayed in the same row
26408 all channels are displayed in separate rows
26411 Default value is @samp{combined}.
26414 Specify display color mode.
26416 It accepts the following values:
26419 each channel is displayed in a separate color
26421 each channel is displayed using the same color scheme
26423 each channel is displayed using the rainbow color scheme
26425 each channel is displayed using the moreland color scheme
26427 each channel is displayed using the nebulae color scheme
26429 each channel is displayed using the fire color scheme
26431 each channel is displayed using the fiery color scheme
26433 each channel is displayed using the fruit color scheme
26435 each channel is displayed using the cool color scheme
26437 each channel is displayed using the magma color scheme
26439 each channel is displayed using the green color scheme
26441 each channel is displayed using the viridis color scheme
26443 each channel is displayed using the plasma color scheme
26445 each channel is displayed using the cividis color scheme
26447 each channel is displayed using the terrain color scheme
26450 Default value is @samp{channel}.
26453 Specify scale used for calculating intensity color values.
26455 It accepts the following values:
26460 square root, default
26471 Default value is @samp{sqrt}.
26474 Specify frequency scale.
26476 It accepts the following values:
26484 Default value is @samp{lin}.
26487 Set saturation modifier for displayed colors. Negative values provide
26488 alternative color scheme. @code{0} is no saturation at all.
26489 Saturation must be in [-10.0, 10.0] range.
26490 Default value is @code{1}.
26493 Set window function.
26495 It accepts the following values:
26520 Default value is @code{hann}.
26523 Set orientation of time vs frequency axis. Can be @code{vertical} or
26524 @code{horizontal}. Default is @code{vertical}.
26527 Set ratio of overlap window. Default value is @code{0}.
26528 When value is @code{1} overlap is set to recommended size for specific
26529 window function currently used.
26532 Set scale gain for calculating intensity color values.
26533 Default value is @code{1}.
26536 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
26539 Set color rotation, must be in [-1.0, 1.0] range.
26540 Default value is @code{0}.
26543 Set start frequency from which to display spectrogram. Default is @code{0}.
26546 Set stop frequency to which to display spectrogram. Default is @code{0}.
26549 Set upper frame rate limit. Default is @code{auto}, unlimited.
26552 Draw time and frequency axes and legends. Default is disabled.
26555 The usage is very similar to the showwaves filter; see the examples in that
26558 @subsection Examples
26562 Large window with logarithmic color scaling:
26564 showspectrum=s=1280x480:scale=log
26568 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
26570 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
26571 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
26575 @section showspectrumpic
26577 Convert input audio to a single video frame, representing the audio frequency
26580 The filter accepts the following options:
26584 Specify the video size for the output. For the syntax of this option, check the
26585 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26586 Default value is @code{4096x2048}.
26589 Specify display mode.
26591 It accepts the following values:
26594 all channels are displayed in the same row
26596 all channels are displayed in separate rows
26598 Default value is @samp{combined}.
26601 Specify display color mode.
26603 It accepts the following values:
26606 each channel is displayed in a separate color
26608 each channel is displayed using the same color scheme
26610 each channel is displayed using the rainbow color scheme
26612 each channel is displayed using the moreland color scheme
26614 each channel is displayed using the nebulae color scheme
26616 each channel is displayed using the fire color scheme
26618 each channel is displayed using the fiery color scheme
26620 each channel is displayed using the fruit color scheme
26622 each channel is displayed using the cool color scheme
26624 each channel is displayed using the magma color scheme
26626 each channel is displayed using the green color scheme
26628 each channel is displayed using the viridis color scheme
26630 each channel is displayed using the plasma color scheme
26632 each channel is displayed using the cividis color scheme
26634 each channel is displayed using the terrain color scheme
26636 Default value is @samp{intensity}.
26639 Specify scale used for calculating intensity color values.
26641 It accepts the following values:
26646 square root, default
26656 Default value is @samp{log}.
26659 Specify frequency scale.
26661 It accepts the following values:
26669 Default value is @samp{lin}.
26672 Set saturation modifier for displayed colors. Negative values provide
26673 alternative color scheme. @code{0} is no saturation at all.
26674 Saturation must be in [-10.0, 10.0] range.
26675 Default value is @code{1}.
26678 Set window function.
26680 It accepts the following values:
26704 Default value is @code{hann}.
26707 Set orientation of time vs frequency axis. Can be @code{vertical} or
26708 @code{horizontal}. Default is @code{vertical}.
26711 Set scale gain for calculating intensity color values.
26712 Default value is @code{1}.
26715 Draw time and frequency axes and legends. Default is enabled.
26718 Set color rotation, must be in [-1.0, 1.0] range.
26719 Default value is @code{0}.
26722 Set start frequency from which to display spectrogram. Default is @code{0}.
26725 Set stop frequency to which to display spectrogram. Default is @code{0}.
26728 @subsection Examples
26732 Extract an audio spectrogram of a whole audio track
26733 in a 1024x1024 picture using @command{ffmpeg}:
26735 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26739 @section showvolume
26741 Convert input audio volume to a video output.
26743 The filter accepts the following options:
26750 Set border width, allowed range is [0, 5]. Default is 1.
26753 Set channel width, allowed range is [80, 8192]. Default is 400.
26756 Set channel height, allowed range is [1, 900]. Default is 20.
26759 Set fade, allowed range is [0, 1]. Default is 0.95.
26762 Set volume color expression.
26764 The expression can use the following variables:
26768 Current max volume of channel in dB.
26774 Current channel number, starting from 0.
26778 If set, displays channel names. Default is enabled.
26781 If set, displays volume values. Default is enabled.
26784 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26785 default is @code{h}.
26788 Set step size, allowed range is [0, 5]. Default is 0, which means
26792 Set background opacity, allowed range is [0, 1]. Default is 0.
26795 Set metering mode, can be peak: @code{p} or rms: @code{r},
26796 default is @code{p}.
26799 Set display scale, can be linear: @code{lin} or log: @code{log},
26800 default is @code{lin}.
26804 If set to > 0., display a line for the max level
26805 in the previous seconds.
26806 default is disabled: @code{0.}
26809 The color of the max line. Use when @code{dm} option is set to > 0.
26810 default is: @code{orange}
26815 Convert input audio to a video output, representing the samples waves.
26817 The filter accepts the following options:
26821 Specify the video size for the output. For the syntax of this option, check the
26822 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26823 Default value is @code{600x240}.
26828 Available values are:
26831 Draw a point for each sample.
26834 Draw a vertical line for each sample.
26837 Draw a point for each sample and a line between them.
26840 Draw a centered vertical line for each sample.
26843 Default value is @code{point}.
26846 Set the number of samples which are printed on the same column. A
26847 larger value will decrease the frame rate. Must be a positive
26848 integer. This option can be set only if the value for @var{rate}
26849 is not explicitly specified.
26852 Set the (approximate) output frame rate. This is done by setting the
26853 option @var{n}. Default value is "25".
26855 @item split_channels
26856 Set if channels should be drawn separately or overlap. Default value is 0.
26859 Set colors separated by '|' which are going to be used for drawing of each channel.
26862 Set amplitude scale.
26864 Available values are:
26882 Set the draw mode. This is mostly useful to set for high @var{n}.
26884 Available values are:
26887 Scale pixel values for each drawn sample.
26890 Draw every sample directly.
26893 Default value is @code{scale}.
26896 @subsection Examples
26900 Output the input file audio and the corresponding video representation
26903 amovie=a.mp3,asplit[out0],showwaves[out1]
26907 Create a synthetic signal and show it with showwaves, forcing a
26908 frame rate of 30 frames per second:
26910 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26914 @section showwavespic
26916 Convert input audio to a single video frame, representing the samples waves.
26918 The filter accepts the following options:
26922 Specify the video size for the output. For the syntax of this option, check the
26923 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26924 Default value is @code{600x240}.
26926 @item split_channels
26927 Set if channels should be drawn separately or overlap. Default value is 0.
26930 Set colors separated by '|' which are going to be used for drawing of each channel.
26933 Set amplitude scale.
26935 Available values are:
26955 Available values are:
26958 Scale pixel values for each drawn sample.
26961 Draw every sample directly.
26964 Default value is @code{scale}.
26967 Set the filter mode.
26969 Available values are:
26972 Use average samples values for each drawn sample.
26975 Use peak samples values for each drawn sample.
26978 Default value is @code{average}.
26981 @subsection Examples
26985 Extract a channel split representation of the wave form of a whole audio track
26986 in a 1024x800 picture using @command{ffmpeg}:
26988 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26992 @section sidedata, asidedata
26994 Delete frame side data, or select frames based on it.
26996 This filter accepts the following options:
27000 Set mode of operation of the filter.
27002 Can be one of the following:
27006 Select every frame with side data of @code{type}.
27009 Delete side data of @code{type}. If @code{type} is not set, delete all side
27015 Set side data type used with all modes. Must be set for @code{select} mode. For
27016 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
27017 in @file{libavutil/frame.h}. For example, to choose
27018 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
27022 @section spectrumsynth
27024 Synthesize audio from 2 input video spectrums, first input stream represents
27025 magnitude across time and second represents phase across time.
27026 The filter will transform from frequency domain as displayed in videos back
27027 to time domain as presented in audio output.
27029 This filter is primarily created for reversing processed @ref{showspectrum}
27030 filter outputs, but can synthesize sound from other spectrograms too.
27031 But in such case results are going to be poor if the phase data is not
27032 available, because in such cases phase data need to be recreated, usually
27033 it's just recreated from random noise.
27034 For best results use gray only output (@code{channel} color mode in
27035 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
27036 @code{lin} scale for phase video. To produce phase, for 2nd video, use
27037 @code{data} option. Inputs videos should generally use @code{fullframe}
27038 slide mode as that saves resources needed for decoding video.
27040 The filter accepts the following options:
27044 Specify sample rate of output audio, the sample rate of audio from which
27045 spectrum was generated may differ.
27048 Set number of channels represented in input video spectrums.
27051 Set scale which was used when generating magnitude input spectrum.
27052 Can be @code{lin} or @code{log}. Default is @code{log}.
27055 Set slide which was used when generating inputs spectrums.
27056 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
27057 Default is @code{fullframe}.
27060 Set window function used for resynthesis.
27063 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
27064 which means optimal overlap for selected window function will be picked.
27067 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
27068 Default is @code{vertical}.
27071 @subsection Examples
27075 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
27076 then resynthesize videos back to audio with spectrumsynth:
27078 ffmpeg -i input.flac -lavfi showspectrum=mode=separate:scale=log:overlap=0.875:color=channel:slide=fullframe:data=magnitude -an -c:v rawvideo magnitude.nut
27079 ffmpeg -i input.flac -lavfi showspectrum=mode=separate:scale=lin:overlap=0.875:color=channel:slide=fullframe:data=phase -an -c:v rawvideo phase.nut
27080 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
27084 @section split, asplit
27086 Split input into several identical outputs.
27088 @code{asplit} works with audio input, @code{split} with video.
27090 The filter accepts a single parameter which specifies the number of outputs. If
27091 unspecified, it defaults to 2.
27093 @subsection Examples
27097 Create two separate outputs from the same input:
27099 [in] split [out0][out1]
27103 To create 3 or more outputs, you need to specify the number of
27106 [in] asplit=3 [out0][out1][out2]
27110 Create two separate outputs from the same input, one cropped and
27113 [in] split [splitout1][splitout2];
27114 [splitout1] crop=100:100:0:0 [cropout];
27115 [splitout2] pad=200:200:100:100 [padout];
27119 Create 5 copies of the input audio with @command{ffmpeg}:
27121 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
27127 Receive commands sent through a libzmq client, and forward them to
27128 filters in the filtergraph.
27130 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
27131 must be inserted between two video filters, @code{azmq} between two
27132 audio filters. Both are capable to send messages to any filter type.
27134 To enable these filters you need to install the libzmq library and
27135 headers and configure FFmpeg with @code{--enable-libzmq}.
27137 For more information about libzmq see:
27138 @url{http://www.zeromq.org/}
27140 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
27141 receives messages sent through a network interface defined by the
27142 @option{bind_address} (or the abbreviation "@option{b}") option.
27143 Default value of this option is @file{tcp://localhost:5555}. You may
27144 want to alter this value to your needs, but do not forget to escape any
27145 ':' signs (see @ref{filtergraph escaping}).
27147 The received message must be in the form:
27149 @var{TARGET} @var{COMMAND} [@var{ARG}]
27152 @var{TARGET} specifies the target of the command, usually the name of
27153 the filter class or a specific filter instance name. The default
27154 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
27155 but you can override this by using the @samp{filter_name@@id} syntax
27156 (see @ref{Filtergraph syntax}).
27158 @var{COMMAND} specifies the name of the command for the target filter.
27160 @var{ARG} is optional and specifies the optional argument list for the
27161 given @var{COMMAND}.
27163 Upon reception, the message is processed and the corresponding command
27164 is injected into the filtergraph. Depending on the result, the filter
27165 will send a reply to the client, adopting the format:
27167 @var{ERROR_CODE} @var{ERROR_REASON}
27171 @var{MESSAGE} is optional.
27173 @subsection Examples
27175 Look at @file{tools/zmqsend} for an example of a zmq client which can
27176 be used to send commands processed by these filters.
27178 Consider the following filtergraph generated by @command{ffplay}.
27179 In this example the last overlay filter has an instance name. All other
27180 filters will have default instance names.
27183 ffplay -dumpgraph 1 -f lavfi "
27184 color=s=100x100:c=red [l];
27185 color=s=100x100:c=blue [r];
27186 nullsrc=s=200x100, zmq [bg];
27187 [bg][l] overlay [bg+l];
27188 [bg+l][r] overlay@@my=x=100 "
27191 To change the color of the left side of the video, the following
27192 command can be used:
27194 echo Parsed_color_0 c yellow | tools/zmqsend
27197 To change the right side:
27199 echo Parsed_color_1 c pink | tools/zmqsend
27202 To change the position of the right side:
27204 echo overlay@@my x 150 | tools/zmqsend
27208 @c man end MULTIMEDIA FILTERS
27210 @chapter Multimedia Sources
27211 @c man begin MULTIMEDIA SOURCES
27213 Below is a description of the currently available multimedia sources.
27217 This is the same as @ref{movie} source, except it selects an audio
27223 Read audio and/or video stream(s) from a movie container.
27225 It accepts the following parameters:
27229 The name of the resource to read (not necessarily a file; it can also be a
27230 device or a stream accessed through some protocol).
27232 @item format_name, f
27233 Specifies the format assumed for the movie to read, and can be either
27234 the name of a container or an input device. If not specified, the
27235 format is guessed from @var{movie_name} or by probing.
27237 @item seek_point, sp
27238 Specifies the seek point in seconds. The frames will be output
27239 starting from this seek point. The parameter is evaluated with
27240 @code{av_strtod}, so the numerical value may be suffixed by an IS
27241 postfix. The default value is "0".
27244 Specifies the streams to read. Several streams can be specified,
27245 separated by "+". The source will then have as many outputs, in the
27246 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
27247 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
27248 respectively the default (best suited) video and audio stream. Default
27249 is "dv", or "da" if the filter is called as "amovie".
27251 @item stream_index, si
27252 Specifies the index of the video stream to read. If the value is -1,
27253 the most suitable video stream will be automatically selected. The default
27254 value is "-1". Deprecated. If the filter is called "amovie", it will select
27255 audio instead of video.
27258 Specifies how many times to read the stream in sequence.
27259 If the value is 0, the stream will be looped infinitely.
27260 Default value is "1".
27262 Note that when the movie is looped the source timestamps are not
27263 changed, so it will generate non monotonically increasing timestamps.
27265 @item discontinuity
27266 Specifies the time difference between frames above which the point is
27267 considered a timestamp discontinuity which is removed by adjusting the later
27271 It allows overlaying a second video on top of the main input of
27272 a filtergraph, as shown in this graph:
27274 input -----------> deltapts0 --> overlay --> output
27277 movie --> scale--> deltapts1 -------+
27279 @subsection Examples
27283 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
27284 on top of the input labelled "in":
27286 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
27287 [in] setpts=PTS-STARTPTS [main];
27288 [main][over] overlay=16:16 [out]
27292 Read from a video4linux2 device, and overlay it on top of the input
27295 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
27296 [in] setpts=PTS-STARTPTS [main];
27297 [main][over] overlay=16:16 [out]
27301 Read the first video stream and the audio stream with id 0x81 from
27302 dvd.vob; the video is connected to the pad named "video" and the audio is
27303 connected to the pad named "audio":
27305 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
27309 @subsection Commands
27311 Both movie and amovie support the following commands:
27314 Perform seek using "av_seek_frame".
27315 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
27318 @var{stream_index}: If stream_index is -1, a default
27319 stream is selected, and @var{timestamp} is automatically converted
27320 from AV_TIME_BASE units to the stream specific time_base.
27322 @var{timestamp}: Timestamp in AVStream.time_base units
27323 or, if no stream is specified, in AV_TIME_BASE units.
27325 @var{flags}: Flags which select direction and seeking mode.
27329 Get movie duration in AV_TIME_BASE units.
27333 @c man end MULTIMEDIA SOURCES