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/install_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 @anchor{dnn_processing}
10131 @section dnn_processing
10133 Do image processing with deep neural networks. It works together with another filter
10134 which converts the pixel format of the Frame to what the dnn network requires.
10136 The filter accepts the following options:
10140 Specify which DNN backend to use for model loading and execution. This option accepts
10141 the following values:
10145 Native implementation of DNN loading and execution.
10148 TensorFlow backend. To enable this backend you
10149 need to install the TensorFlow for C library (see
10150 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
10151 @code{--enable-libtensorflow}
10154 OpenVINO backend. To enable this backend you
10155 need to build and install the OpenVINO for C library (see
10156 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
10157 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
10158 be needed if the header files and libraries are not installed into system path)
10162 Default value is @samp{native}.
10165 Set path to model file specifying network architecture and its parameters.
10166 Note that different backends use different file formats. TensorFlow, OpenVINO and native
10167 backend can load files for only its format.
10169 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
10172 Set the input name of the dnn network.
10175 Set the output name of the dnn network.
10178 use DNN async execution if set (default: set),
10179 roll back to sync execution if the backend does not support async.
10183 @subsection Examples
10187 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
10189 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
10193 Halve the pixel value of the frame with format gray32f:
10195 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
10199 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
10201 ./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
10205 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
10207 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
10214 Draw a colored box on the input image.
10216 It accepts the following parameters:
10221 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10225 The expressions which specify the width and height of the box; if 0 they are interpreted as
10226 the input width and height. It defaults to 0.
10229 Specify the color of the box to write. For the general syntax of this option,
10230 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10231 value @code{invert} is used, the box edge color is the same as the
10232 video with inverted luma.
10235 The expression which sets the thickness of the box edge.
10236 A value of @code{fill} will create a filled box. Default value is @code{3}.
10238 See below for the list of accepted constants.
10241 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10242 will overwrite the video's color and alpha pixels.
10243 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10246 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10247 following constants:
10251 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10255 horizontal and vertical chroma subsample values. For example for the
10256 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10260 The input width and height.
10263 The input sample aspect ratio.
10267 The x and y offset coordinates where the box is drawn.
10271 The width and height of the drawn box.
10274 The thickness of the drawn box.
10276 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10277 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10281 @subsection Examples
10285 Draw a black box around the edge of the input image:
10291 Draw a box with color red and an opacity of 50%:
10293 drawbox=10:20:200:60:red@@0.5
10296 The previous example can be specified as:
10298 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10302 Fill the box with pink color:
10304 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10308 Draw a 2-pixel red 2.40:1 mask:
10310 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
10314 @subsection Commands
10315 This filter supports same commands as options.
10316 The command accepts the same syntax of the corresponding option.
10318 If the specified expression is not valid, it is kept at its current
10323 Draw a graph using input video metadata.
10325 It accepts the following parameters:
10329 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10332 Set 1st foreground color expression.
10335 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10338 Set 2nd foreground color expression.
10341 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10344 Set 3rd foreground color expression.
10347 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10350 Set 4th foreground color expression.
10353 Set minimal value of metadata value.
10356 Set maximal value of metadata value.
10359 Set graph background color. Default is white.
10364 Available values for mode is:
10371 Default is @code{line}.
10376 Available values for slide is:
10379 Draw new frame when right border is reached.
10382 Replace old columns with new ones.
10385 Scroll from right to left.
10388 Scroll from left to right.
10391 Draw single picture.
10394 Default is @code{frame}.
10397 Set size of graph video. For the syntax of this option, check the
10398 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10399 The default value is @code{900x256}.
10402 Set the output frame rate. Default value is @code{25}.
10404 The foreground color expressions can use the following variables:
10407 Minimal value of metadata value.
10410 Maximal value of metadata value.
10413 Current metadata key value.
10416 The color is defined as 0xAABBGGRR.
10419 Example using metadata from @ref{signalstats} filter:
10421 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10424 Example using metadata from @ref{ebur128} filter:
10426 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10431 Draw a grid on the input image.
10433 It accepts the following parameters:
10438 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10442 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10443 input width and height, respectively, minus @code{thickness}, so image gets
10444 framed. Default to 0.
10447 Specify the color of the grid. For the general syntax of this option,
10448 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10449 value @code{invert} is used, the grid color is the same as the
10450 video with inverted luma.
10453 The expression which sets the thickness of the grid line. Default value is @code{1}.
10455 See below for the list of accepted constants.
10458 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10459 will overwrite the video's color and alpha pixels.
10460 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10463 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10464 following constants:
10468 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10472 horizontal and vertical chroma subsample values. For example for the
10473 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10477 The input grid cell width and height.
10480 The input sample aspect ratio.
10484 The x and y coordinates of some point of grid intersection (meant to configure offset).
10488 The width and height of the drawn cell.
10491 The thickness of the drawn cell.
10493 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10494 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10498 @subsection Examples
10502 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10504 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10508 Draw a white 3x3 grid with an opacity of 50%:
10510 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10514 @subsection Commands
10515 This filter supports same commands as options.
10516 The command accepts the same syntax of the corresponding option.
10518 If the specified expression is not valid, it is kept at its current
10524 Draw a text string or text from a specified file on top of a video, using the
10525 libfreetype library.
10527 To enable compilation of this filter, you need to configure FFmpeg with
10528 @code{--enable-libfreetype}.
10529 To enable default font fallback and the @var{font} option you need to
10530 configure FFmpeg with @code{--enable-libfontconfig}.
10531 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10532 @code{--enable-libfribidi}.
10536 It accepts the following parameters:
10541 Used to draw a box around text using the background color.
10542 The value must be either 1 (enable) or 0 (disable).
10543 The default value of @var{box} is 0.
10546 Set the width of the border to be drawn around the box using @var{boxcolor}.
10547 The default value of @var{boxborderw} is 0.
10550 The color to be used for drawing box around text. For the syntax of this
10551 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10553 The default value of @var{boxcolor} is "white".
10556 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10557 The default value of @var{line_spacing} is 0.
10560 Set the width of the border to be drawn around the text using @var{bordercolor}.
10561 The default value of @var{borderw} is 0.
10564 Set the color to be used for drawing border around text. For the syntax of this
10565 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10567 The default value of @var{bordercolor} is "black".
10570 Select how the @var{text} is expanded. Can be either @code{none},
10571 @code{strftime} (deprecated) or
10572 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10576 Set a start time for the count. Value is in microseconds. Only applied
10577 in the deprecated strftime expansion mode. To emulate in normal expansion
10578 mode use the @code{pts} function, supplying the start time (in seconds)
10579 as the second argument.
10582 If true, check and fix text coords to avoid clipping.
10585 The color to be used for drawing fonts. For the syntax of this option, check
10586 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10588 The default value of @var{fontcolor} is "black".
10590 @item fontcolor_expr
10591 String which is expanded the same way as @var{text} to obtain dynamic
10592 @var{fontcolor} value. By default this option has empty value and is not
10593 processed. When this option is set, it overrides @var{fontcolor} option.
10596 The font family to be used for drawing text. By default Sans.
10599 The font file to be used for drawing text. The path must be included.
10600 This parameter is mandatory if the fontconfig support is disabled.
10603 Draw the text applying alpha blending. The value can
10604 be a number between 0.0 and 1.0.
10605 The expression accepts the same variables @var{x, y} as well.
10606 The default value is 1.
10607 Please see @var{fontcolor_expr}.
10610 The font size to be used for drawing text.
10611 The default value of @var{fontsize} is 16.
10614 If set to 1, attempt to shape the text (for example, reverse the order of
10615 right-to-left text and join Arabic characters) before drawing it.
10616 Otherwise, just draw the text exactly as given.
10617 By default 1 (if supported).
10619 @item ft_load_flags
10620 The flags to be used for loading the fonts.
10622 The flags map the corresponding flags supported by libfreetype, and are
10623 a combination of the following values:
10630 @item vertical_layout
10631 @item force_autohint
10634 @item ignore_global_advance_width
10636 @item ignore_transform
10638 @item linear_design
10642 Default value is "default".
10644 For more information consult the documentation for the FT_LOAD_*
10648 The color to be used for drawing a shadow behind the drawn text. For the
10649 syntax of this option, check the @ref{color syntax,,"Color" section in the
10650 ffmpeg-utils manual,ffmpeg-utils}.
10652 The default value of @var{shadowcolor} is "black".
10656 The x and y offsets for the text shadow position with respect to the
10657 position of the text. They can be either positive or negative
10658 values. The default value for both is "0".
10661 The starting frame number for the n/frame_num variable. The default value
10665 The size in number of spaces to use for rendering the tab.
10666 Default value is 4.
10669 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10670 format. It can be used with or without text parameter. @var{timecode_rate}
10671 option must be specified.
10673 @item timecode_rate, rate, r
10674 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10675 integer. Minimum value is "1".
10676 Drop-frame timecode is supported for frame rates 30 & 60.
10679 If set to 1, the output of the timecode option will wrap around at 24 hours.
10680 Default is 0 (disabled).
10683 The text string to be drawn. The text must be a sequence of UTF-8
10684 encoded characters.
10685 This parameter is mandatory if no file is specified with the parameter
10689 A text file containing text to be drawn. The text must be a sequence
10690 of UTF-8 encoded characters.
10692 This parameter is mandatory if no text string is specified with the
10693 parameter @var{text}.
10695 If both @var{text} and @var{textfile} are specified, an error is thrown.
10698 If set to 1, the @var{textfile} will be reloaded before each frame.
10699 Be sure to update it atomically, or it may be read partially, or even fail.
10703 The expressions which specify the offsets where text will be drawn
10704 within the video frame. They are relative to the top/left border of the
10707 The default value of @var{x} and @var{y} is "0".
10709 See below for the list of accepted constants and functions.
10712 The parameters for @var{x} and @var{y} are expressions containing the
10713 following constants and functions:
10717 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10721 horizontal and vertical chroma subsample values. For example for the
10722 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10725 the height of each text line
10733 @item max_glyph_a, ascent
10734 the maximum distance from the baseline to the highest/upper grid
10735 coordinate used to place a glyph outline point, for all the rendered
10737 It is a positive value, due to the grid's orientation with the Y axis
10740 @item max_glyph_d, descent
10741 the maximum distance from the baseline to the lowest grid coordinate
10742 used to place a glyph outline point, for all the rendered glyphs.
10743 This is a negative value, due to the grid's orientation, with the Y axis
10747 maximum glyph height, that is the maximum height for all the glyphs
10748 contained in the rendered text, it is equivalent to @var{ascent} -
10752 maximum glyph width, that is the maximum width for all the glyphs
10753 contained in the rendered text
10756 the number of input frame, starting from 0
10758 @item rand(min, max)
10759 return a random number included between @var{min} and @var{max}
10762 The input sample aspect ratio.
10765 timestamp expressed in seconds, NAN if the input timestamp is unknown
10768 the height of the rendered text
10771 the width of the rendered text
10775 the x and y offset coordinates where the text is drawn.
10777 These parameters allow the @var{x} and @var{y} expressions to refer
10778 to each other, so you can for example specify @code{y=x/dar}.
10781 A one character description of the current frame's picture type.
10784 The current packet's position in the input file or stream
10785 (in bytes, from the start of the input). A value of -1 indicates
10786 this info is not available.
10789 The current packet's duration, in seconds.
10792 The current packet's size (in bytes).
10795 @anchor{drawtext_expansion}
10796 @subsection Text expansion
10798 If @option{expansion} is set to @code{strftime},
10799 the filter recognizes strftime() sequences in the provided text and
10800 expands them accordingly. Check the documentation of strftime(). This
10801 feature is deprecated.
10803 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10805 If @option{expansion} is set to @code{normal} (which is the default),
10806 the following expansion mechanism is used.
10808 The backslash character @samp{\}, followed by any character, always expands to
10809 the second character.
10811 Sequences of the form @code{%@{...@}} are expanded. The text between the
10812 braces is a function name, possibly followed by arguments separated by ':'.
10813 If the arguments contain special characters or delimiters (':' or '@}'),
10814 they should be escaped.
10816 Note that they probably must also be escaped as the value for the
10817 @option{text} option in the filter argument string and as the filter
10818 argument in the filtergraph description, and possibly also for the shell,
10819 that makes up to four levels of escaping; using a text file avoids these
10822 The following functions are available:
10827 The expression evaluation result.
10829 It must take one argument specifying the expression to be evaluated,
10830 which accepts the same constants and functions as the @var{x} and
10831 @var{y} values. Note that not all constants should be used, for
10832 example the text size is not known when evaluating the expression, so
10833 the constants @var{text_w} and @var{text_h} will have an undefined
10836 @item expr_int_format, eif
10837 Evaluate the expression's value and output as formatted integer.
10839 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10840 The second argument specifies the output format. Allowed values are @samp{x},
10841 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10842 @code{printf} function.
10843 The third parameter is optional and sets the number of positions taken by the output.
10844 It can be used to add padding with zeros from the left.
10847 The time at which the filter is running, expressed in UTC.
10848 It can accept an argument: a strftime() format string.
10851 The time at which the filter is running, expressed in the local time zone.
10852 It can accept an argument: a strftime() format string.
10855 Frame metadata. Takes one or two arguments.
10857 The first argument is mandatory and specifies the metadata key.
10859 The second argument is optional and specifies a default value, used when the
10860 metadata key is not found or empty.
10862 Available metadata can be identified by inspecting entries
10863 starting with TAG included within each frame section
10864 printed by running @code{ffprobe -show_frames}.
10866 String metadata generated in filters leading to
10867 the drawtext filter are also available.
10870 The frame number, starting from 0.
10873 A one character description of the current picture type.
10876 The timestamp of the current frame.
10877 It can take up to three arguments.
10879 The first argument is the format of the timestamp; it defaults to @code{flt}
10880 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10881 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10882 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10883 @code{localtime} stands for the timestamp of the frame formatted as
10884 local time zone time.
10886 The second argument is an offset added to the timestamp.
10888 If the format is set to @code{hms}, a third argument @code{24HH} may be
10889 supplied to present the hour part of the formatted timestamp in 24h format
10892 If the format is set to @code{localtime} or @code{gmtime},
10893 a third argument may be supplied: a strftime() format string.
10894 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10897 @subsection Commands
10899 This filter supports altering parameters via commands:
10902 Alter existing filter parameters.
10904 Syntax for the argument is the same as for filter invocation, e.g.
10907 fontsize=56:fontcolor=green:text='Hello World'
10910 Full filter invocation with sendcmd would look like this:
10913 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10917 If the entire argument can't be parsed or applied as valid values then the filter will
10918 continue with its existing parameters.
10920 @subsection Examples
10924 Draw "Test Text" with font FreeSerif, using the default values for the
10925 optional parameters.
10928 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10932 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10933 and y=50 (counting from the top-left corner of the screen), text is
10934 yellow with a red box around it. Both the text and the box have an
10938 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10939 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10942 Note that the double quotes are not necessary if spaces are not used
10943 within the parameter list.
10946 Show the text at the center of the video frame:
10948 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10952 Show the text at a random position, switching to a new position every 30 seconds:
10954 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)"
10958 Show a text line sliding from right to left in the last row of the video
10959 frame. The file @file{LONG_LINE} is assumed to contain a single line
10962 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10966 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10968 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10972 Draw a single green letter "g", at the center of the input video.
10973 The glyph baseline is placed at half screen height.
10975 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10979 Show text for 1 second every 3 seconds:
10981 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10985 Use fontconfig to set the font. Note that the colons need to be escaped.
10987 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10991 Draw "Test Text" with font size dependent on height of the video.
10993 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10997 Print the date of a real-time encoding (see strftime(3)):
10999 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
11003 Show text fading in and out (appearing/disappearing):
11006 DS=1.0 # display start
11007 DE=10.0 # display end
11008 FID=1.5 # fade in duration
11009 FOD=5 # fade out duration
11010 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 @}"
11014 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
11015 and the @option{fontsize} value are included in the @option{y} offset.
11017 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
11018 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
11022 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
11023 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
11024 must have option @option{-export_path_metadata 1} for the special metadata fields
11025 to be available for filters.
11027 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
11032 For more information about libfreetype, check:
11033 @url{http://www.freetype.org/}.
11035 For more information about fontconfig, check:
11036 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
11038 For more information about libfribidi, check:
11039 @url{http://fribidi.org/}.
11041 @section edgedetect
11043 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
11045 The filter accepts the following options:
11050 Set low and high threshold values used by the Canny thresholding
11053 The high threshold selects the "strong" edge pixels, which are then
11054 connected through 8-connectivity with the "weak" edge pixels selected
11055 by the low threshold.
11057 @var{low} and @var{high} threshold values must be chosen in the range
11058 [0,1], and @var{low} should be lesser or equal to @var{high}.
11060 Default value for @var{low} is @code{20/255}, and default value for @var{high}
11064 Define the drawing mode.
11068 Draw white/gray wires on black background.
11071 Mix the colors to create a paint/cartoon effect.
11074 Apply Canny edge detector on all selected planes.
11076 Default value is @var{wires}.
11079 Select planes for filtering. By default all available planes are filtered.
11082 @subsection Examples
11086 Standard edge detection with custom values for the hysteresis thresholding:
11088 edgedetect=low=0.1:high=0.4
11092 Painting effect without thresholding:
11094 edgedetect=mode=colormix:high=0
11100 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
11102 For each input image, the filter will compute the optimal mapping from
11103 the input to the output given the codebook length, that is the number
11104 of distinct output colors.
11106 This filter accepts the following options.
11109 @item codebook_length, l
11110 Set codebook length. The value must be a positive integer, and
11111 represents the number of distinct output colors. Default value is 256.
11114 Set the maximum number of iterations to apply for computing the optimal
11115 mapping. The higher the value the better the result and the higher the
11116 computation time. Default value is 1.
11119 Set a random seed, must be an integer included between 0 and
11120 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
11121 will try to use a good random seed on a best effort basis.
11124 Set pal8 output pixel format. This option does not work with codebook
11125 length greater than 256.
11130 Measure graylevel entropy in histogram of color channels of video frames.
11132 It accepts the following parameters:
11136 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
11138 @var{diff} mode measures entropy of histogram delta values, absolute differences
11139 between neighbour histogram values.
11143 Apply the EPX magnification filter which is designed for pixel art.
11145 It accepts the following option:
11149 Set the scaling dimension: @code{2} for @code{2xEPX}, @code{3} for
11151 Default is @code{3}.
11155 Set brightness, contrast, saturation and approximate gamma adjustment.
11157 The filter accepts the following options:
11161 Set the contrast expression. The value must be a float value in range
11162 @code{-1000.0} to @code{1000.0}. The default value is "1".
11165 Set the brightness expression. The value must be a float value in
11166 range @code{-1.0} to @code{1.0}. The default value is "0".
11169 Set the saturation expression. The value must be a float in
11170 range @code{0.0} to @code{3.0}. The default value is "1".
11173 Set the gamma expression. The value must be a float in range
11174 @code{0.1} to @code{10.0}. The default value is "1".
11177 Set the gamma expression for red. The value must be a float in
11178 range @code{0.1} to @code{10.0}. The default value is "1".
11181 Set the gamma expression for green. The value must be a float in range
11182 @code{0.1} to @code{10.0}. The default value is "1".
11185 Set the gamma expression for blue. The value must be a float in range
11186 @code{0.1} to @code{10.0}. The default value is "1".
11189 Set the gamma weight expression. It can be used to reduce the effect
11190 of a high gamma value on bright image areas, e.g. keep them from
11191 getting overamplified and just plain white. The value must be a float
11192 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
11193 gamma correction all the way down while @code{1.0} leaves it at its
11194 full strength. Default is "1".
11197 Set when the expressions for brightness, contrast, saturation and
11198 gamma expressions are evaluated.
11200 It accepts the following values:
11203 only evaluate expressions once during the filter initialization or
11204 when a command is processed
11207 evaluate expressions for each incoming frame
11210 Default value is @samp{init}.
11213 The expressions accept the following parameters:
11216 frame count of the input frame starting from 0
11219 byte position of the corresponding packet in the input file, NAN if
11223 frame rate of the input video, NAN if the input frame rate is unknown
11226 timestamp expressed in seconds, NAN if the input timestamp is unknown
11229 @subsection Commands
11230 The filter supports the following commands:
11234 Set the contrast expression.
11237 Set the brightness expression.
11240 Set the saturation expression.
11243 Set the gamma expression.
11246 Set the gamma_r expression.
11249 Set gamma_g expression.
11252 Set gamma_b expression.
11255 Set gamma_weight expression.
11257 The command accepts the same syntax of the corresponding option.
11259 If the specified expression is not valid, it is kept at its current
11266 Apply erosion effect to the video.
11268 This filter replaces the pixel by the local(3x3) minimum.
11270 It accepts the following options:
11277 Limit the maximum change for each plane, default is 65535.
11278 If 0, plane will remain unchanged.
11281 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11284 Flags to local 3x3 coordinates maps like this:
11291 @subsection Commands
11293 This filter supports the all above options as @ref{commands}.
11297 Deinterlace the input video ("estdif" stands for "Edge Slope
11298 Tracing Deinterlacing Filter").
11300 Spatial only filter that uses edge slope tracing algorithm
11301 to interpolate missing lines.
11302 It accepts the following parameters:
11306 The interlacing mode to adopt. It accepts one of the following values:
11310 Output one frame for each frame.
11312 Output one frame for each field.
11315 The default value is @code{field}.
11318 The picture field parity assumed for the input interlaced video. It accepts one
11319 of the following values:
11323 Assume the top field is first.
11325 Assume the bottom field is first.
11327 Enable automatic detection of field parity.
11330 The default value is @code{auto}.
11331 If the interlacing is unknown or the decoder does not export this information,
11332 top field first will be assumed.
11335 Specify which frames to deinterlace. Accepts one of the following
11340 Deinterlace all frames.
11342 Only deinterlace frames marked as interlaced.
11345 The default value is @code{all}.
11348 Specify the search radius for edge slope tracing. Default value is 1.
11349 Allowed range is from 1 to 15.
11352 Specify the search radius for best edge matching. Default value is 2.
11353 Allowed range is from 0 to 15.
11356 Specify the interpolation used. Default is 4-point interpolation. It accepts one
11357 of the following values:
11361 Two-point interpolation.
11363 Four-point interpolation.
11365 Six-point interpolation.
11369 @subsection Commands
11370 This filter supports same @ref{commands} as options.
11373 Adjust exposure of the video stream.
11375 The filter accepts the following options:
11379 Set the exposure correction in EV. Allowed range is from -3.0 to 3.0 EV
11380 Default value is 0 EV.
11383 Set the black level correction. Allowed range is from -1.0 to 1.0.
11384 Default value is 0.
11387 @subsection Commands
11389 This filter supports same @ref{commands} as options.
11391 @section extractplanes
11393 Extract color channel components from input video stream into
11394 separate grayscale video streams.
11396 The filter accepts the following option:
11400 Set plane(s) to extract.
11402 Available values for planes are:
11413 Choosing planes not available in the input will result in an error.
11414 That means you cannot select @code{r}, @code{g}, @code{b} planes
11415 with @code{y}, @code{u}, @code{v} planes at same time.
11418 @subsection Examples
11422 Extract luma, u and v color channel component from input video frame
11423 into 3 grayscale outputs:
11425 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
11431 Apply a fade-in/out effect to the input video.
11433 It accepts the following parameters:
11437 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11439 Default is @code{in}.
11441 @item start_frame, s
11442 Specify the number of the frame to start applying the fade
11443 effect at. Default is 0.
11446 The number of frames that the fade effect lasts. At the end of the
11447 fade-in effect, the output video will have the same intensity as the input video.
11448 At the end of the fade-out transition, the output video will be filled with the
11449 selected @option{color}.
11453 If set to 1, fade only alpha channel, if one exists on the input.
11454 Default value is 0.
11456 @item start_time, st
11457 Specify the timestamp (in seconds) of the frame to start to apply the fade
11458 effect. If both start_frame and start_time are specified, the fade will start at
11459 whichever comes last. Default is 0.
11462 The number of seconds for which the fade effect has to last. At the end of the
11463 fade-in effect the output video will have the same intensity as the input video,
11464 at the end of the fade-out transition the output video will be filled with the
11465 selected @option{color}.
11466 If both duration and nb_frames are specified, duration is used. Default is 0
11467 (nb_frames is used by default).
11470 Specify the color of the fade. Default is "black".
11473 @subsection Examples
11477 Fade in the first 30 frames of video:
11482 The command above is equivalent to:
11488 Fade out the last 45 frames of a 200-frame video:
11491 fade=type=out:start_frame=155:nb_frames=45
11495 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11497 fade=in:0:25, fade=out:975:25
11501 Make the first 5 frames yellow, then fade in from frame 5-24:
11503 fade=in:5:20:color=yellow
11507 Fade in alpha over first 25 frames of video:
11509 fade=in:0:25:alpha=1
11513 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11515 fade=t=in:st=5.5:d=0.5
11521 Denoise frames using 3D FFT (frequency domain filtering).
11523 The filter accepts the following options:
11527 Set the noise sigma constant. This sets denoising strength.
11528 Default value is 1. Allowed range is from 0 to 30.
11529 Using very high sigma with low overlap may give blocking artifacts.
11532 Set amount of denoising. By default all detected noise is reduced.
11533 Default value is 1. Allowed range is from 0 to 1.
11536 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11537 Actual size of block in pixels is 2 to power of @var{block}, so by default
11538 block size in pixels is 2^4 which is 16.
11541 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11544 Set number of previous frames to use for denoising. By default is set to 0.
11547 Set number of next frames to to use for denoising. By default is set to 0.
11550 Set planes which will be filtered, by default are all available filtered
11555 Apply arbitrary expressions to samples in frequency domain
11559 Adjust the dc value (gain) of the luma plane of the image. The filter
11560 accepts an integer value in range @code{0} to @code{1000}. The default
11561 value is set to @code{0}.
11564 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11565 filter accepts an integer value in range @code{0} to @code{1000}. The
11566 default value is set to @code{0}.
11569 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11570 filter accepts an integer value in range @code{0} to @code{1000}. The
11571 default value is set to @code{0}.
11574 Set the frequency domain weight expression for the luma plane.
11577 Set the frequency domain weight expression for the 1st chroma plane.
11580 Set the frequency domain weight expression for the 2nd chroma plane.
11583 Set when the expressions are evaluated.
11585 It accepts the following values:
11588 Only evaluate expressions once during the filter initialization.
11591 Evaluate expressions for each incoming frame.
11594 Default value is @samp{init}.
11596 The filter accepts the following variables:
11599 The coordinates of the current sample.
11603 The width and height of the image.
11606 The number of input frame, starting from 0.
11609 @subsection Examples
11615 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11621 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11627 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11633 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11640 Extract a single field from an interlaced image using stride
11641 arithmetic to avoid wasting CPU time. The output frames are marked as
11644 The filter accepts the following options:
11648 Specify whether to extract the top (if the value is @code{0} or
11649 @code{top}) or the bottom field (if the value is @code{1} or
11655 Create new frames by copying the top and bottom fields from surrounding frames
11656 supplied as numbers by the hint file.
11660 Set file containing hints: absolute/relative frame numbers.
11662 There must be one line for each frame in a clip. Each line must contain two
11663 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11664 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11665 is current frame number for @code{absolute} mode or out of [-1, 1] range
11666 for @code{relative} mode. First number tells from which frame to pick up top
11667 field and second number tells from which frame to pick up bottom field.
11669 If optionally followed by @code{+} output frame will be marked as interlaced,
11670 else if followed by @code{-} output frame will be marked as progressive, else
11671 it will be marked same as input frame.
11672 If optionally followed by @code{t} output frame will use only top field, or in
11673 case of @code{b} it will use only bottom field.
11674 If line starts with @code{#} or @code{;} that line is skipped.
11677 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11680 Example of first several lines of @code{hint} file for @code{relative} mode:
11682 0,0 - # first frame
11683 1,0 - # second frame, use third's frame top field and second's frame bottom field
11684 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11699 @section fieldmatch
11701 Field matching filter for inverse telecine. It is meant to reconstruct the
11702 progressive frames from a telecined stream. The filter does not drop duplicated
11703 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11704 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11706 The separation of the field matching and the decimation is notably motivated by
11707 the possibility of inserting a de-interlacing filter fallback between the two.
11708 If the source has mixed telecined and real interlaced content,
11709 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11710 But these remaining combed frames will be marked as interlaced, and thus can be
11711 de-interlaced by a later filter such as @ref{yadif} before decimation.
11713 In addition to the various configuration options, @code{fieldmatch} can take an
11714 optional second stream, activated through the @option{ppsrc} option. If
11715 enabled, the frames reconstruction will be based on the fields and frames from
11716 this second stream. This allows the first input to be pre-processed in order to
11717 help the various algorithms of the filter, while keeping the output lossless
11718 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11719 or brightness/contrast adjustments can help.
11721 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11722 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11723 which @code{fieldmatch} is based on. While the semantic and usage are very
11724 close, some behaviour and options names can differ.
11726 The @ref{decimate} filter currently only works for constant frame rate input.
11727 If your input has mixed telecined (30fps) and progressive content with a lower
11728 framerate like 24fps use the following filterchain to produce the necessary cfr
11729 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11731 The filter accepts the following options:
11735 Specify the assumed field order of the input stream. Available values are:
11739 Auto detect parity (use FFmpeg's internal parity value).
11741 Assume bottom field first.
11743 Assume top field first.
11746 Note that it is sometimes recommended not to trust the parity announced by the
11749 Default value is @var{auto}.
11752 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11753 sense that it won't risk creating jerkiness due to duplicate frames when
11754 possible, but if there are bad edits or blended fields it will end up
11755 outputting combed frames when a good match might actually exist. On the other
11756 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11757 but will almost always find a good frame if there is one. The other values are
11758 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11759 jerkiness and creating duplicate frames versus finding good matches in sections
11760 with bad edits, orphaned fields, blended fields, etc.
11762 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11764 Available values are:
11768 2-way matching (p/c)
11770 2-way matching, and trying 3rd match if still combed (p/c + n)
11772 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11774 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11775 still combed (p/c + n + u/b)
11777 3-way matching (p/c/n)
11779 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11780 detected as combed (p/c/n + u/b)
11783 The parenthesis at the end indicate the matches that would be used for that
11784 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11787 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11790 Default value is @var{pc_n}.
11793 Mark the main input stream as a pre-processed input, and enable the secondary
11794 input stream as the clean source to pick the fields from. See the filter
11795 introduction for more details. It is similar to the @option{clip2} feature from
11798 Default value is @code{0} (disabled).
11801 Set the field to match from. It is recommended to set this to the same value as
11802 @option{order} unless you experience matching failures with that setting. In
11803 certain circumstances changing the field that is used to match from can have a
11804 large impact on matching performance. Available values are:
11808 Automatic (same value as @option{order}).
11810 Match from the bottom field.
11812 Match from the top field.
11815 Default value is @var{auto}.
11818 Set whether or not chroma is included during the match comparisons. In most
11819 cases it is recommended to leave this enabled. You should set this to @code{0}
11820 only if your clip has bad chroma problems such as heavy rainbowing or other
11821 artifacts. Setting this to @code{0} could also be used to speed things up at
11822 the cost of some accuracy.
11824 Default value is @code{1}.
11828 These define an exclusion band which excludes the lines between @option{y0} and
11829 @option{y1} from being included in the field matching decision. An exclusion
11830 band can be used to ignore subtitles, a logo, or other things that may
11831 interfere with the matching. @option{y0} sets the starting scan line and
11832 @option{y1} sets the ending line; all lines in between @option{y0} and
11833 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11834 @option{y0} and @option{y1} to the same value will disable the feature.
11835 @option{y0} and @option{y1} defaults to @code{0}.
11838 Set the scene change detection threshold as a percentage of maximum change on
11839 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11840 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11841 @option{scthresh} is @code{[0.0, 100.0]}.
11843 Default value is @code{12.0}.
11846 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11847 account the combed scores of matches when deciding what match to use as the
11848 final match. Available values are:
11852 No final matching based on combed scores.
11854 Combed scores are only used when a scene change is detected.
11856 Use combed scores all the time.
11859 Default is @var{sc}.
11862 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11863 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11864 Available values are:
11868 No forced calculation.
11870 Force p/c/n calculations.
11872 Force p/c/n/u/b calculations.
11875 Default value is @var{none}.
11878 This is the area combing threshold used for combed frame detection. This
11879 essentially controls how "strong" or "visible" combing must be to be detected.
11880 Larger values mean combing must be more visible and smaller values mean combing
11881 can be less visible or strong and still be detected. Valid settings are from
11882 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11883 be detected as combed). This is basically a pixel difference value. A good
11884 range is @code{[8, 12]}.
11886 Default value is @code{9}.
11889 Sets whether or not chroma is considered in the combed frame decision. Only
11890 disable this if your source has chroma problems (rainbowing, etc.) that are
11891 causing problems for the combed frame detection with chroma enabled. Actually,
11892 using @option{chroma}=@var{0} is usually more reliable, except for the case
11893 where there is chroma only combing in the source.
11895 Default value is @code{0}.
11899 Respectively set the x-axis and y-axis size of the window used during combed
11900 frame detection. This has to do with the size of the area in which
11901 @option{combpel} pixels are required to be detected as combed for a frame to be
11902 declared combed. See the @option{combpel} parameter description for more info.
11903 Possible values are any number that is a power of 2 starting at 4 and going up
11906 Default value is @code{16}.
11909 The number of combed pixels inside any of the @option{blocky} by
11910 @option{blockx} size blocks on the frame for the frame to be detected as
11911 combed. While @option{cthresh} controls how "visible" the combing must be, this
11912 setting controls "how much" combing there must be in any localized area (a
11913 window defined by the @option{blockx} and @option{blocky} settings) on the
11914 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11915 which point no frames will ever be detected as combed). This setting is known
11916 as @option{MI} in TFM/VFM vocabulary.
11918 Default value is @code{80}.
11921 @anchor{p/c/n/u/b meaning}
11922 @subsection p/c/n/u/b meaning
11924 @subsubsection p/c/n
11926 We assume the following telecined stream:
11929 Top fields: 1 2 2 3 4
11930 Bottom fields: 1 2 3 4 4
11933 The numbers correspond to the progressive frame the fields relate to. Here, the
11934 first two frames are progressive, the 3rd and 4th are combed, and so on.
11936 When @code{fieldmatch} is configured to run a matching from bottom
11937 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11942 B 1 2 3 4 4 <-- matching reference
11951 As a result of the field matching, we can see that some frames get duplicated.
11952 To perform a complete inverse telecine, you need to rely on a decimation filter
11953 after this operation. See for instance the @ref{decimate} filter.
11955 The same operation now matching from top fields (@option{field}=@var{top})
11960 T 1 2 2 3 4 <-- matching reference
11970 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11971 basically, they refer to the frame and field of the opposite parity:
11974 @item @var{p} matches the field of the opposite parity in the previous frame
11975 @item @var{c} matches the field of the opposite parity in the current frame
11976 @item @var{n} matches the field of the opposite parity in the next frame
11981 The @var{u} and @var{b} matching are a bit special in the sense that they match
11982 from the opposite parity flag. In the following examples, we assume that we are
11983 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11984 'x' is placed above and below each matched fields.
11986 With bottom matching (@option{field}=@var{bottom}):
11991 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11992 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
12000 With top matching (@option{field}=@var{top}):
12005 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
12006 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
12014 @subsection Examples
12016 Simple IVTC of a top field first telecined stream:
12018 fieldmatch=order=tff:combmatch=none, decimate
12021 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
12023 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
12026 @section fieldorder
12028 Transform the field order of the input video.
12030 It accepts the following parameters:
12035 The output field order. Valid values are @var{tff} for top field first or @var{bff}
12036 for bottom field first.
12039 The default value is @samp{tff}.
12041 The transformation is done by shifting the picture content up or down
12042 by one line, and filling the remaining line with appropriate picture content.
12043 This method is consistent with most broadcast field order converters.
12045 If the input video is not flagged as being interlaced, or it is already
12046 flagged as being of the required output field order, then this filter does
12047 not alter the incoming video.
12049 It is very useful when converting to or from PAL DV material,
12050 which is bottom field first.
12054 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
12057 @section fifo, afifo
12059 Buffer input images and send them when they are requested.
12061 It is mainly useful when auto-inserted by the libavfilter
12064 It does not take parameters.
12066 @section fillborders
12068 Fill borders of the input video, without changing video stream dimensions.
12069 Sometimes video can have garbage at the four edges and you may not want to
12070 crop video input to keep size multiple of some number.
12072 This filter accepts the following options:
12076 Number of pixels to fill from left border.
12079 Number of pixels to fill from right border.
12082 Number of pixels to fill from top border.
12085 Number of pixels to fill from bottom border.
12090 It accepts the following values:
12093 fill pixels using outermost pixels
12096 fill pixels using mirroring (half sample symmetric)
12099 fill pixels with constant value
12102 fill pixels using reflecting (whole sample symmetric)
12105 fill pixels using wrapping
12108 fade pixels to constant value
12111 Default is @var{smear}.
12114 Set color for pixels in fixed or fade mode. Default is @var{black}.
12117 @subsection Commands
12118 This filter supports same @ref{commands} as options.
12119 The command accepts the same syntax of the corresponding option.
12121 If the specified expression is not valid, it is kept at its current
12126 Find a rectangular object
12128 It accepts the following options:
12132 Filepath of the object image, needs to be in gray8.
12135 Detection threshold, default is 0.5.
12138 Number of mipmaps, default is 3.
12140 @item xmin, ymin, xmax, ymax
12141 Specifies the rectangle in which to search.
12144 @subsection Examples
12148 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
12150 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
12156 Flood area with values of same pixel components with another values.
12158 It accepts the following options:
12161 Set pixel x coordinate.
12164 Set pixel y coordinate.
12167 Set source #0 component value.
12170 Set source #1 component value.
12173 Set source #2 component value.
12176 Set source #3 component value.
12179 Set destination #0 component value.
12182 Set destination #1 component value.
12185 Set destination #2 component value.
12188 Set destination #3 component value.
12194 Convert the input video to one of the specified pixel formats.
12195 Libavfilter will try to pick one that is suitable as input to
12198 It accepts the following parameters:
12202 A '|'-separated list of pixel format names, such as
12203 "pix_fmts=yuv420p|monow|rgb24".
12207 @subsection Examples
12211 Convert the input video to the @var{yuv420p} format
12213 format=pix_fmts=yuv420p
12216 Convert the input video to any of the formats in the list
12218 format=pix_fmts=yuv420p|yuv444p|yuv410p
12225 Convert the video to specified constant frame rate by duplicating or dropping
12226 frames as necessary.
12228 It accepts the following parameters:
12232 The desired output frame rate. The default is @code{25}.
12235 Assume the first PTS should be the given value, in seconds. This allows for
12236 padding/trimming at the start of stream. By default, no assumption is made
12237 about the first frame's expected PTS, so no padding or trimming is done.
12238 For example, this could be set to 0 to pad the beginning with duplicates of
12239 the first frame if a video stream starts after the audio stream or to trim any
12240 frames with a negative PTS.
12243 Timestamp (PTS) rounding method.
12245 Possible values are:
12252 round towards -infinity
12254 round towards +infinity
12258 The default is @code{near}.
12261 Action performed when reading the last frame.
12263 Possible values are:
12266 Use same timestamp rounding method as used for other frames.
12268 Pass through last frame if input duration has not been reached yet.
12270 The default is @code{round}.
12274 Alternatively, the options can be specified as a flat string:
12275 @var{fps}[:@var{start_time}[:@var{round}]].
12277 See also the @ref{setpts} filter.
12279 @subsection Examples
12283 A typical usage in order to set the fps to 25:
12289 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
12291 fps=fps=film:round=near
12297 Pack two different video streams into a stereoscopic video, setting proper
12298 metadata on supported codecs. The two views should have the same size and
12299 framerate and processing will stop when the shorter video ends. Please note
12300 that you may conveniently adjust view properties with the @ref{scale} and
12303 It accepts the following parameters:
12307 The desired packing format. Supported values are:
12312 The views are next to each other (default).
12315 The views are on top of each other.
12318 The views are packed by line.
12321 The views are packed by column.
12324 The views are temporally interleaved.
12333 # Convert left and right views into a frame-sequential video
12334 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12336 # Convert views into a side-by-side video with the same output resolution as the input
12337 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
12342 Change the frame rate by interpolating new video output frames from the source
12345 This filter is not designed to function correctly with interlaced media. If
12346 you wish to change the frame rate of interlaced media then you are required
12347 to deinterlace before this filter and re-interlace after this filter.
12349 A description of the accepted options follows.
12353 Specify the output frames per second. This option can also be specified
12354 as a value alone. The default is @code{50}.
12357 Specify the start of a range where the output frame will be created as a
12358 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12359 the default is @code{15}.
12362 Specify the end of a range where the output frame will be created as a
12363 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12364 the default is @code{240}.
12367 Specify the level at which a scene change is detected as a value between
12368 0 and 100 to indicate a new scene; a low value reflects a low
12369 probability for the current frame to introduce a new scene, while a higher
12370 value means the current frame is more likely to be one.
12371 The default is @code{8.2}.
12374 Specify flags influencing the filter process.
12376 Available value for @var{flags} is:
12379 @item scene_change_detect, scd
12380 Enable scene change detection using the value of the option @var{scene}.
12381 This flag is enabled by default.
12387 Select one frame every N-th frame.
12389 This filter accepts the following option:
12392 Select frame after every @code{step} frames.
12393 Allowed values are positive integers higher than 0. Default value is @code{1}.
12396 @section freezedetect
12398 Detect frozen video.
12400 This filter logs a message and sets frame metadata when it detects that the
12401 input video has no significant change in content during a specified duration.
12402 Video freeze detection calculates the mean average absolute difference of all
12403 the components of video frames and compares it to a noise floor.
12405 The printed times and duration are expressed in seconds. The
12406 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12407 whose timestamp equals or exceeds the detection duration and it contains the
12408 timestamp of the first frame of the freeze. The
12409 @code{lavfi.freezedetect.freeze_duration} and
12410 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12413 The filter accepts the following options:
12417 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12418 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12422 Set freeze duration until notification (default is 2 seconds).
12425 @section freezeframes
12427 Freeze video frames.
12429 This filter freezes video frames using frame from 2nd input.
12431 The filter accepts the following options:
12435 Set number of first frame from which to start freeze.
12438 Set number of last frame from which to end freeze.
12441 Set number of frame from 2nd input which will be used instead of replaced frames.
12447 Apply a frei0r effect to the input video.
12449 To enable the compilation of this filter, you need to install the frei0r
12450 header and configure FFmpeg with @code{--enable-frei0r}.
12452 It accepts the following parameters:
12457 The name of the frei0r effect to load. If the environment variable
12458 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12459 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12460 Otherwise, the standard frei0r paths are searched, in this order:
12461 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12462 @file{/usr/lib/frei0r-1/}.
12464 @item filter_params
12465 A '|'-separated list of parameters to pass to the frei0r effect.
12469 A frei0r effect parameter can be a boolean (its value is either
12470 "y" or "n"), a double, a color (specified as
12471 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12472 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12473 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12474 a position (specified as @var{X}/@var{Y}, where
12475 @var{X} and @var{Y} are floating point numbers) and/or a string.
12477 The number and types of parameters depend on the loaded effect. If an
12478 effect parameter is not specified, the default value is set.
12480 @subsection Examples
12484 Apply the distort0r effect, setting the first two double parameters:
12486 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12490 Apply the colordistance effect, taking a color as the first parameter:
12492 frei0r=colordistance:0.2/0.3/0.4
12493 frei0r=colordistance:violet
12494 frei0r=colordistance:0x112233
12498 Apply the perspective effect, specifying the top left and top right image
12501 frei0r=perspective:0.2/0.2|0.8/0.2
12505 For more information, see
12506 @url{http://frei0r.dyne.org}
12508 @subsection Commands
12510 This filter supports the @option{filter_params} option as @ref{commands}.
12514 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12516 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12517 processing filter, one of them is performed once per block, not per pixel.
12518 This allows for much higher speed.
12520 The filter accepts the following options:
12524 Set quality. This option defines the number of levels for averaging. It accepts
12525 an integer in the range 4-5. Default value is @code{4}.
12528 Force a constant quantization parameter. It accepts an integer in range 0-63.
12529 If not set, the filter will use the QP from the video stream (if available).
12532 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12533 more details but also more artifacts, while higher values make the image smoother
12534 but also blurrier. Default value is @code{0} − PSNR optimal.
12536 @item use_bframe_qp
12537 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12538 option may cause flicker since the B-Frames have often larger QP. Default is
12539 @code{0} (not enabled).
12545 Apply Gaussian blur filter.
12547 The filter accepts the following options:
12551 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12554 Set number of steps for Gaussian approximation. Default is @code{1}.
12557 Set which planes to filter. By default all planes are filtered.
12560 Set vertical sigma, if negative it will be same as @code{sigma}.
12561 Default is @code{-1}.
12564 @subsection Commands
12565 This filter supports same commands as options.
12566 The command accepts the same syntax of the corresponding option.
12568 If the specified expression is not valid, it is kept at its current
12573 Apply generic equation to each pixel.
12575 The filter accepts the following options:
12578 @item lum_expr, lum
12579 Set the luminance expression.
12581 Set the chrominance blue expression.
12583 Set the chrominance red expression.
12584 @item alpha_expr, a
12585 Set the alpha expression.
12587 Set the red expression.
12588 @item green_expr, g
12589 Set the green expression.
12591 Set the blue expression.
12594 The colorspace is selected according to the specified options. If one
12595 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12596 options is specified, the filter will automatically select a YCbCr
12597 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12598 @option{blue_expr} options is specified, it will select an RGB
12601 If one of the chrominance expression is not defined, it falls back on the other
12602 one. If no alpha expression is specified it will evaluate to opaque value.
12603 If none of chrominance expressions are specified, they will evaluate
12604 to the luminance expression.
12606 The expressions can use the following variables and functions:
12610 The sequential number of the filtered frame, starting from @code{0}.
12614 The coordinates of the current sample.
12618 The width and height of the image.
12622 Width and height scale depending on the currently filtered plane. It is the
12623 ratio between the corresponding luma plane number of pixels and the current
12624 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12625 @code{0.5,0.5} for chroma planes.
12628 Time of the current frame, expressed in seconds.
12631 Return the value of the pixel at location (@var{x},@var{y}) of the current
12635 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12639 Return the value of the pixel at location (@var{x},@var{y}) of the
12640 blue-difference chroma plane. Return 0 if there is no such plane.
12643 Return the value of the pixel at location (@var{x},@var{y}) of the
12644 red-difference chroma plane. Return 0 if there is no such plane.
12649 Return the value of the pixel at location (@var{x},@var{y}) of the
12650 red/green/blue component. Return 0 if there is no such component.
12653 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12654 plane. Return 0 if there is no such plane.
12656 @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)
12657 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12658 sums of samples within a rectangle. See the functions without the sum postfix.
12660 @item interpolation
12661 Set one of interpolation methods:
12666 Default is bilinear.
12669 For functions, if @var{x} and @var{y} are outside the area, the value will be
12670 automatically clipped to the closer edge.
12672 Please note that this filter can use multiple threads in which case each slice
12673 will have its own expression state. If you want to use only a single expression
12674 state because your expressions depend on previous state then you should limit
12675 the number of filter threads to 1.
12677 @subsection Examples
12681 Flip the image horizontally:
12687 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12688 wavelength of 100 pixels:
12690 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12694 Generate a fancy enigmatic moving light:
12696 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
12700 Generate a quick emboss effect:
12702 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12706 Modify RGB components depending on pixel position:
12708 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12712 Create a radial gradient that is the same size as the input (also see
12713 the @ref{vignette} filter):
12715 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12721 Fix the banding artifacts that are sometimes introduced into nearly flat
12722 regions by truncation to 8-bit color depth.
12723 Interpolate the gradients that should go where the bands are, and
12726 It is designed for playback only. Do not use it prior to
12727 lossy compression, because compression tends to lose the dither and
12728 bring back the bands.
12730 It accepts the following parameters:
12735 The maximum amount by which the filter will change any one pixel. This is also
12736 the threshold for detecting nearly flat regions. Acceptable values range from
12737 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12741 The neighborhood to fit the gradient to. A larger radius makes for smoother
12742 gradients, but also prevents the filter from modifying the pixels near detailed
12743 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12744 values will be clipped to the valid range.
12748 Alternatively, the options can be specified as a flat string:
12749 @var{strength}[:@var{radius}]
12751 @subsection Examples
12755 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12761 Specify radius, omitting the strength (which will fall-back to the default
12769 @anchor{graphmonitor}
12770 @section graphmonitor
12771 Show various filtergraph stats.
12773 With this filter one can debug complete filtergraph.
12774 Especially issues with links filling with queued frames.
12776 The filter accepts the following options:
12780 Set video output size. Default is @var{hd720}.
12783 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12786 Set output mode, can be @var{fulll} or @var{compact}.
12787 In @var{compact} mode only filters with some queued frames have displayed stats.
12790 Set flags which enable which stats are shown in video.
12792 Available values for flags are:
12795 Display number of queued frames in each link.
12797 @item frame_count_in
12798 Display number of frames taken from filter.
12800 @item frame_count_out
12801 Display number of frames given out from filter.
12804 Display current filtered frame pts.
12807 Display current filtered frame time.
12810 Display time base for filter link.
12813 Display used format for filter link.
12816 Display video size or number of audio channels in case of audio used by filter link.
12819 Display video frame rate or sample rate in case of audio used by filter link.
12822 Display link output status.
12826 Set upper limit for video rate of output stream, Default value is @var{25}.
12827 This guarantee that output video frame rate will not be higher than this value.
12831 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12832 and corrects the scene colors accordingly.
12834 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12836 The filter accepts the following options:
12840 The order of differentiation to be applied on the scene. Must be chosen in the range
12841 [0,2] and default value is 1.
12844 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12845 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12846 max value instead of calculating Minkowski distance.
12849 The standard deviation of Gaussian blur to be applied on the scene. Must be
12850 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12851 can't be equal to 0 if @var{difford} is greater than 0.
12854 @subsection Examples
12860 greyedge=difford=1:minknorm=5:sigma=2
12866 greyedge=difford=1:minknorm=0:sigma=2
12874 Apply a Hald CLUT to a video stream.
12876 First input is the video stream to process, and second one is the Hald CLUT.
12877 The Hald CLUT input can be a simple picture or a complete video stream.
12879 The filter accepts the following options:
12883 Force termination when the shortest input terminates. Default is @code{0}.
12885 Continue applying the last CLUT after the end of the stream. A value of
12886 @code{0} disable the filter after the last frame of the CLUT is reached.
12887 Default is @code{1}.
12890 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12891 filters share the same internals).
12893 This filter also supports the @ref{framesync} options.
12895 More information about the Hald CLUT can be found on Eskil Steenberg's website
12896 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12898 @subsection Commands
12900 This filter supports the @code{interp} option as @ref{commands}.
12902 @subsection Workflow examples
12904 @subsubsection Hald CLUT video stream
12906 Generate an identity Hald CLUT stream altered with various effects:
12908 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
12911 Note: make sure you use a lossless codec.
12913 Then use it with @code{haldclut} to apply it on some random stream:
12915 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12918 The Hald CLUT will be applied to the 10 first seconds (duration of
12919 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12920 to the remaining frames of the @code{mandelbrot} stream.
12922 @subsubsection Hald CLUT with preview
12924 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12925 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12926 biggest possible square starting at the top left of the picture. The remaining
12927 padding pixels (bottom or right) will be ignored. This area can be used to add
12928 a preview of the Hald CLUT.
12930 Typically, the following generated Hald CLUT will be supported by the
12931 @code{haldclut} filter:
12934 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12935 pad=iw+320 [padded_clut];
12936 smptebars=s=320x256, split [a][b];
12937 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12938 [main][b] overlay=W-320" -frames:v 1 clut.png
12941 It contains the original and a preview of the effect of the CLUT: SMPTE color
12942 bars are displayed on the right-top, and below the same color bars processed by
12945 Then, the effect of this Hald CLUT can be visualized with:
12947 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12952 Flip the input video horizontally.
12954 For example, to horizontally flip the input video with @command{ffmpeg}:
12956 ffmpeg -i in.avi -vf "hflip" out.avi
12960 This filter applies a global color histogram equalization on a
12963 It can be used to correct video that has a compressed range of pixel
12964 intensities. The filter redistributes the pixel intensities to
12965 equalize their distribution across the intensity range. It may be
12966 viewed as an "automatically adjusting contrast filter". This filter is
12967 useful only for correcting degraded or poorly captured source
12970 The filter accepts the following options:
12974 Determine the amount of equalization to be applied. As the strength
12975 is reduced, the distribution of pixel intensities more-and-more
12976 approaches that of the input frame. The value must be a float number
12977 in the range [0,1] and defaults to 0.200.
12980 Set the maximum intensity that can generated and scale the output
12981 values appropriately. The strength should be set as desired and then
12982 the intensity can be limited if needed to avoid washing-out. The value
12983 must be a float number in the range [0,1] and defaults to 0.210.
12986 Set the antibanding level. If enabled the filter will randomly vary
12987 the luminance of output pixels by a small amount to avoid banding of
12988 the histogram. Possible values are @code{none}, @code{weak} or
12989 @code{strong}. It defaults to @code{none}.
12995 Compute and draw a color distribution histogram for the input video.
12997 The computed histogram is a representation of the color component
12998 distribution in an image.
13000 Standard histogram displays the color components distribution in an image.
13001 Displays color graph for each color component. Shows distribution of
13002 the Y, U, V, A or R, G, B components, depending on input format, in the
13003 current frame. Below each graph a color component scale meter is shown.
13005 The filter accepts the following options:
13009 Set height of level. Default value is @code{200}.
13010 Allowed range is [50, 2048].
13013 Set height of color scale. Default value is @code{12}.
13014 Allowed range is [0, 40].
13018 It accepts the following values:
13021 Per color component graphs are placed below each other.
13024 Per color component graphs are placed side by side.
13027 Presents information identical to that in the @code{parade}, except
13028 that the graphs representing color components are superimposed directly
13031 Default is @code{stack}.
13034 Set mode. Can be either @code{linear}, or @code{logarithmic}.
13035 Default is @code{linear}.
13038 Set what color components to display.
13039 Default is @code{7}.
13042 Set foreground opacity. Default is @code{0.7}.
13045 Set background opacity. Default is @code{0.5}.
13048 @subsection Examples
13053 Calculate and draw histogram:
13055 ffplay -i input -vf histogram
13063 This is a high precision/quality 3d denoise filter. It aims to reduce
13064 image noise, producing smooth images and making still images really
13065 still. It should enhance compressibility.
13067 It accepts the following optional parameters:
13071 A non-negative floating point number which specifies spatial luma strength.
13072 It defaults to 4.0.
13074 @item chroma_spatial
13075 A non-negative floating point number which specifies spatial chroma strength.
13076 It defaults to 3.0*@var{luma_spatial}/4.0.
13079 A floating point number which specifies luma temporal strength. It defaults to
13080 6.0*@var{luma_spatial}/4.0.
13083 A floating point number which specifies chroma temporal strength. It defaults to
13084 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
13087 @subsection Commands
13088 This filter supports same @ref{commands} as options.
13089 The command accepts the same syntax of the corresponding option.
13091 If the specified expression is not valid, it is kept at its current
13094 @anchor{hwdownload}
13095 @section hwdownload
13097 Download hardware frames to system memory.
13099 The input must be in hardware frames, and the output a non-hardware format.
13100 Not all formats will be supported on the output - it may be necessary to insert
13101 an additional @option{format} filter immediately following in the graph to get
13102 the output in a supported format.
13106 Map hardware frames to system memory or to another device.
13108 This filter has several different modes of operation; which one is used depends
13109 on the input and output formats:
13112 Hardware frame input, normal frame output
13114 Map the input frames to system memory and pass them to the output. If the
13115 original hardware frame is later required (for example, after overlaying
13116 something else on part of it), the @option{hwmap} filter can be used again
13117 in the next mode to retrieve it.
13119 Normal frame input, hardware frame output
13121 If the input is actually a software-mapped hardware frame, then unmap it -
13122 that is, return the original hardware frame.
13124 Otherwise, a device must be provided. Create new hardware surfaces on that
13125 device for the output, then map them back to the software format at the input
13126 and give those frames to the preceding filter. This will then act like the
13127 @option{hwupload} filter, but may be able to avoid an additional copy when
13128 the input is already in a compatible format.
13130 Hardware frame input and output
13132 A device must be supplied for the output, either directly or with the
13133 @option{derive_device} option. The input and output devices must be of
13134 different types and compatible - the exact meaning of this is
13135 system-dependent, but typically it means that they must refer to the same
13136 underlying hardware context (for example, refer to the same graphics card).
13138 If the input frames were originally created on the output device, then unmap
13139 to retrieve the original frames.
13141 Otherwise, map the frames to the output device - create new hardware frames
13142 on the output corresponding to the frames on the input.
13145 The following additional parameters are accepted:
13149 Set the frame mapping mode. Some combination of:
13152 The mapped frame should be readable.
13154 The mapped frame should be writeable.
13156 The mapping will always overwrite the entire frame.
13158 This may improve performance in some cases, as the original contents of the
13159 frame need not be loaded.
13161 The mapping must not involve any copying.
13163 Indirect mappings to copies of frames are created in some cases where either
13164 direct mapping is not possible or it would have unexpected properties.
13165 Setting this flag ensures that the mapping is direct and will fail if that is
13168 Defaults to @var{read+write} if not specified.
13170 @item derive_device @var{type}
13171 Rather than using the device supplied at initialisation, instead derive a new
13172 device of type @var{type} from the device the input frames exist on.
13175 In a hardware to hardware mapping, map in reverse - create frames in the sink
13176 and map them back to the source. This may be necessary in some cases where
13177 a mapping in one direction is required but only the opposite direction is
13178 supported by the devices being used.
13180 This option is dangerous - it may break the preceding filter in undefined
13181 ways if there are any additional constraints on that filter's output.
13182 Do not use it without fully understanding the implications of its use.
13188 Upload system memory frames to hardware surfaces.
13190 The device to upload to must be supplied when the filter is initialised. If
13191 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
13192 option or with the @option{derive_device} option. The input and output devices
13193 must be of different types and compatible - the exact meaning of this is
13194 system-dependent, but typically it means that they must refer to the same
13195 underlying hardware context (for example, refer to the same graphics card).
13197 The following additional parameters are accepted:
13200 @item derive_device @var{type}
13201 Rather than using the device supplied at initialisation, instead derive a new
13202 device of type @var{type} from the device the input frames exist on.
13205 @anchor{hwupload_cuda}
13206 @section hwupload_cuda
13208 Upload system memory frames to a CUDA device.
13210 It accepts the following optional parameters:
13214 The number of the CUDA device to use
13219 Apply a high-quality magnification filter designed for pixel art. This filter
13220 was originally created by Maxim Stepin.
13222 It accepts the following option:
13226 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
13227 @code{hq3x} and @code{4} for @code{hq4x}.
13228 Default is @code{3}.
13232 Stack input videos horizontally.
13234 All streams must be of same pixel format and of same height.
13236 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
13237 to create same output.
13239 The filter accepts the following option:
13243 Set number of input streams. Default is 2.
13246 If set to 1, force the output to terminate when the shortest input
13247 terminates. Default value is 0.
13252 Modify the hue and/or the saturation of the input.
13254 It accepts the following parameters:
13258 Specify the hue angle as a number of degrees. It accepts an expression,
13259 and defaults to "0".
13262 Specify the saturation in the [-10,10] range. It accepts an expression and
13266 Specify the hue angle as a number of radians. It accepts an
13267 expression, and defaults to "0".
13270 Specify the brightness in the [-10,10] range. It accepts an expression and
13274 @option{h} and @option{H} are mutually exclusive, and can't be
13275 specified at the same time.
13277 The @option{b}, @option{h}, @option{H} and @option{s} option values are
13278 expressions containing the following constants:
13282 frame count of the input frame starting from 0
13285 presentation timestamp of the input frame expressed in time base units
13288 frame rate of the input video, NAN if the input frame rate is unknown
13291 timestamp expressed in seconds, NAN if the input timestamp is unknown
13294 time base of the input video
13297 @subsection Examples
13301 Set the hue to 90 degrees and the saturation to 1.0:
13307 Same command but expressing the hue in radians:
13313 Rotate hue and make the saturation swing between 0
13314 and 2 over a period of 1 second:
13316 hue="H=2*PI*t: s=sin(2*PI*t)+1"
13320 Apply a 3 seconds saturation fade-in effect starting at 0:
13322 hue="s=min(t/3\,1)"
13325 The general fade-in expression can be written as:
13327 hue="s=min(0\, max((t-START)/DURATION\, 1))"
13331 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13333 hue="s=max(0\, min(1\, (8-t)/3))"
13336 The general fade-out expression can be written as:
13338 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13343 @subsection Commands
13345 This filter supports the following commands:
13351 Modify the hue and/or the saturation and/or brightness of the input video.
13352 The command accepts the same syntax of the corresponding option.
13354 If the specified expression is not valid, it is kept at its current
13358 @section hysteresis
13360 Grow first stream into second stream by connecting components.
13361 This makes it possible to build more robust edge masks.
13363 This filter accepts the following options:
13367 Set which planes will be processed as bitmap, unprocessed planes will be
13368 copied from first stream.
13369 By default value 0xf, all planes will be processed.
13372 Set threshold which is used in filtering. If pixel component value is higher than
13373 this value filter algorithm for connecting components is activated.
13374 By default value is 0.
13377 The @code{hysteresis} filter also supports the @ref{framesync} options.
13381 Obtain the identity score between two input videos.
13383 This filter takes two input videos.
13385 Both input videos must have the same resolution and pixel format for
13386 this filter to work correctly. Also it assumes that both inputs
13387 have the same number of frames, which are compared one by one.
13389 The obtained per component, average, min and max identity score is printed through
13390 the logging system.
13392 The filter stores the calculated identity scores of each frame in frame metadata.
13394 In the below example the input file @file{main.mpg} being processed is compared
13395 with the reference file @file{ref.mpg}.
13398 ffmpeg -i main.mpg -i ref.mpg -lavfi identity -f null -
13403 Detect video interlacing type.
13405 This filter tries to detect if the input frames are interlaced, progressive,
13406 top or bottom field first. It will also try to detect fields that are
13407 repeated between adjacent frames (a sign of telecine).
13409 Single frame detection considers only immediately adjacent frames when classifying each frame.
13410 Multiple frame detection incorporates the classification history of previous frames.
13412 The filter will log these metadata values:
13415 @item single.current_frame
13416 Detected type of current frame using single-frame detection. One of:
13417 ``tff'' (top field first), ``bff'' (bottom field first),
13418 ``progressive'', or ``undetermined''
13421 Cumulative number of frames detected as top field first using single-frame detection.
13424 Cumulative number of frames detected as top field first using multiple-frame detection.
13427 Cumulative number of frames detected as bottom field first using single-frame detection.
13429 @item multiple.current_frame
13430 Detected type of current frame using multiple-frame detection. One of:
13431 ``tff'' (top field first), ``bff'' (bottom field first),
13432 ``progressive'', or ``undetermined''
13435 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13437 @item single.progressive
13438 Cumulative number of frames detected as progressive using single-frame detection.
13440 @item multiple.progressive
13441 Cumulative number of frames detected as progressive using multiple-frame detection.
13443 @item single.undetermined
13444 Cumulative number of frames that could not be classified using single-frame detection.
13446 @item multiple.undetermined
13447 Cumulative number of frames that could not be classified using multiple-frame detection.
13449 @item repeated.current_frame
13450 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13452 @item repeated.neither
13453 Cumulative number of frames with no repeated field.
13456 Cumulative number of frames with the top field repeated from the previous frame's top field.
13458 @item repeated.bottom
13459 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13462 The filter accepts the following options:
13466 Set interlacing threshold.
13468 Set progressive threshold.
13470 Threshold for repeated field detection.
13472 Number of frames after which a given frame's contribution to the
13473 statistics is halved (i.e., it contributes only 0.5 to its
13474 classification). The default of 0 means that all frames seen are given
13475 full weight of 1.0 forever.
13476 @item analyze_interlaced_flag
13477 When this is not 0 then idet will use the specified number of frames to determine
13478 if the interlaced flag is accurate, it will not count undetermined frames.
13479 If the flag is found to be accurate it will be used without any further
13480 computations, if it is found to be inaccurate it will be cleared without any
13481 further computations. This allows inserting the idet filter as a low computational
13482 method to clean up the interlaced flag
13487 Deinterleave or interleave fields.
13489 This filter allows one to process interlaced images fields without
13490 deinterlacing them. Deinterleaving splits the input frame into 2
13491 fields (so called half pictures). Odd lines are moved to the top
13492 half of the output image, even lines to the bottom half.
13493 You can process (filter) them independently and then re-interleave them.
13495 The filter accepts the following options:
13499 @item chroma_mode, c
13500 @item alpha_mode, a
13501 Available values for @var{luma_mode}, @var{chroma_mode} and
13502 @var{alpha_mode} are:
13508 @item deinterleave, d
13509 Deinterleave fields, placing one above the other.
13511 @item interleave, i
13512 Interleave fields. Reverse the effect of deinterleaving.
13514 Default value is @code{none}.
13516 @item luma_swap, ls
13517 @item chroma_swap, cs
13518 @item alpha_swap, as
13519 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13522 @subsection Commands
13524 This filter supports the all above options as @ref{commands}.
13528 Apply inflate effect to the video.
13530 This filter replaces the pixel by the local(3x3) average by taking into account
13531 only values higher than the pixel.
13533 It accepts the following options:
13540 Limit the maximum change for each plane, default is 65535.
13541 If 0, plane will remain unchanged.
13544 @subsection Commands
13546 This filter supports the all above options as @ref{commands}.
13550 Simple interlacing filter from progressive contents. This interleaves upper (or
13551 lower) lines from odd frames with lower (or upper) lines from even frames,
13552 halving the frame rate and preserving image height.
13555 Original Original New Frame
13556 Frame 'j' Frame 'j+1' (tff)
13557 ========== =========== ==================
13558 Line 0 --------------------> Frame 'j' Line 0
13559 Line 1 Line 1 ----> Frame 'j+1' Line 1
13560 Line 2 ---------------------> Frame 'j' Line 2
13561 Line 3 Line 3 ----> Frame 'j+1' Line 3
13563 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13566 It accepts the following optional parameters:
13570 This determines whether the interlaced frame is taken from the even
13571 (tff - default) or odd (bff) lines of the progressive frame.
13574 Vertical lowpass filter to avoid twitter interlacing and
13575 reduce moire patterns.
13579 Disable vertical lowpass filter
13582 Enable linear filter (default)
13585 Enable complex filter. This will slightly less reduce twitter and moire
13586 but better retain detail and subjective sharpness impression.
13593 Deinterlace input video by applying Donald Graft's adaptive kernel
13594 deinterling. Work on interlaced parts of a video to produce
13595 progressive frames.
13597 The description of the accepted parameters follows.
13601 Set the threshold which affects the filter's tolerance when
13602 determining if a pixel line must be processed. It must be an integer
13603 in the range [0,255] and defaults to 10. A value of 0 will result in
13604 applying the process on every pixels.
13607 Paint pixels exceeding the threshold value to white if set to 1.
13611 Set the fields order. Swap fields if set to 1, leave fields alone if
13615 Enable additional sharpening if set to 1. Default is 0.
13618 Enable twoway sharpening if set to 1. Default is 0.
13621 @subsection Examples
13625 Apply default values:
13627 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13631 Enable additional sharpening:
13637 Paint processed pixels in white:
13644 Apply kirsch operator to input video stream.
13646 The filter accepts the following option:
13650 Set which planes will be processed, unprocessed planes will be copied.
13651 By default value 0xf, all planes will be processed.
13654 Set value which will be multiplied with filtered result.
13657 Set value which will be added to filtered result.
13660 @subsection Commands
13662 This filter supports the all above options as @ref{commands}.
13666 Slowly update darker pixels.
13668 This filter makes short flashes of light appear longer.
13669 This filter accepts the following options:
13673 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13676 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13679 @subsection Commands
13681 This filter supports the all above options as @ref{commands}.
13683 @section lenscorrection
13685 Correct radial lens distortion
13687 This filter can be used to correct for radial distortion as can result from the use
13688 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13689 one can use tools available for example as part of opencv or simply trial-and-error.
13690 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13691 and extract the k1 and k2 coefficients from the resulting matrix.
13693 Note that effectively the same filter is available in the open-source tools Krita and
13694 Digikam from the KDE project.
13696 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13697 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13698 brightness distribution, so you may want to use both filters together in certain
13699 cases, though you will have to take care of ordering, i.e. whether vignetting should
13700 be applied before or after lens correction.
13702 @subsection Options
13704 The filter accepts the following options:
13708 Relative x-coordinate of the focal point of the image, and thereby the center of the
13709 distortion. This value has a range [0,1] and is expressed as fractions of the image
13710 width. Default is 0.5.
13712 Relative y-coordinate of the focal point of the image, and thereby the center of the
13713 distortion. This value has a range [0,1] and is expressed as fractions of the image
13714 height. Default is 0.5.
13716 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13717 no correction. Default is 0.
13719 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13720 0 means no correction. Default is 0.
13722 Set interpolation type. Can be @code{nearest} or @code{bilinear}.
13723 Default is @code{nearest}.
13725 Specify the color of the unmapped pixels. For the syntax of this option,
13726 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
13727 manual,ffmpeg-utils}. Default color is @code{black@@0}.
13730 The formula that generates the correction is:
13732 @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)
13734 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13735 distances from the focal point in the source and target images, respectively.
13737 @subsection Commands
13739 This filter supports the all above options as @ref{commands}.
13743 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13745 The @code{lensfun} filter requires the camera make, camera model, and lens model
13746 to apply the lens correction. The filter will load the lensfun database and
13747 query it to find the corresponding camera and lens entries in the database. As
13748 long as these entries can be found with the given options, the filter can
13749 perform corrections on frames. Note that incomplete strings will result in the
13750 filter choosing the best match with the given options, and the filter will
13751 output the chosen camera and lens models (logged with level "info"). You must
13752 provide the make, camera model, and lens model as they are required.
13754 The filter accepts the following options:
13758 The make of the camera (for example, "Canon"). This option is required.
13761 The model of the camera (for example, "Canon EOS 100D"). This option is
13765 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13766 option is required.
13769 The type of correction to apply. The following values are valid options:
13773 Enables fixing lens vignetting.
13776 Enables fixing lens geometry. This is the default.
13779 Enables fixing chromatic aberrations.
13782 Enables fixing lens vignetting and lens geometry.
13785 Enables fixing lens vignetting and chromatic aberrations.
13788 Enables fixing both lens geometry and chromatic aberrations.
13791 Enables all possible corrections.
13795 The focal length of the image/video (zoom; expected constant for video). For
13796 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13797 range should be chosen when using that lens. Default 18.
13800 The aperture of the image/video (expected constant for video). Note that
13801 aperture is only used for vignetting correction. Default 3.5.
13803 @item focus_distance
13804 The focus distance of the image/video (expected constant for video). Note that
13805 focus distance is only used for vignetting and only slightly affects the
13806 vignetting correction process. If unknown, leave it at the default value (which
13810 The scale factor which is applied after transformation. After correction the
13811 video is no longer necessarily rectangular. This parameter controls how much of
13812 the resulting image is visible. The value 0 means that a value will be chosen
13813 automatically such that there is little or no unmapped area in the output
13814 image. 1.0 means that no additional scaling is done. Lower values may result
13815 in more of the corrected image being visible, while higher values may avoid
13816 unmapped areas in the output.
13818 @item target_geometry
13819 The target geometry of the output image/video. The following values are valid
13823 @item rectilinear (default)
13826 @item equirectangular
13827 @item fisheye_orthographic
13828 @item fisheye_stereographic
13829 @item fisheye_equisolid
13830 @item fisheye_thoby
13833 Apply the reverse of image correction (instead of correcting distortion, apply
13836 @item interpolation
13837 The type of interpolation used when correcting distortion. The following values
13842 @item linear (default)
13847 @subsection Examples
13851 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13852 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13856 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
13860 Apply the same as before, but only for the first 5 seconds of video.
13863 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
13870 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13871 score between two input videos.
13873 The obtained VMAF score is printed through the logging system.
13875 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13876 After installing the library it can be enabled using:
13877 @code{./configure --enable-libvmaf}.
13878 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13880 The filter has following options:
13884 Set the model path which is to be used for SVM.
13885 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13888 Set the file path to be used to store logs.
13891 Set the format of the log file (csv, json or xml).
13893 @item enable_transform
13894 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13895 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13896 Default value: @code{false}
13899 Invokes the phone model which will generate VMAF scores higher than in the
13900 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13901 Default value: @code{false}
13904 Enables computing psnr along with vmaf.
13905 Default value: @code{false}
13908 Enables computing ssim along with vmaf.
13909 Default value: @code{false}
13912 Enables computing ms_ssim along with vmaf.
13913 Default value: @code{false}
13916 Set the pool method to be used for computing vmaf.
13917 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13920 Set number of threads to be used when computing vmaf.
13921 Default value: @code{0}, which makes use of all available logical processors.
13924 Set interval for frame subsampling used when computing vmaf.
13925 Default value: @code{1}
13927 @item enable_conf_interval
13928 Enables confidence interval.
13929 Default value: @code{false}
13932 This filter also supports the @ref{framesync} options.
13934 @subsection Examples
13937 On the below examples the input file @file{main.mpg} being processed is
13938 compared with the reference file @file{ref.mpg}.
13941 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13945 Example with options:
13947 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13951 Example with options and different containers:
13953 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 -
13959 Limits the pixel components values to the specified range [min, max].
13961 The filter accepts the following options:
13965 Lower bound. Defaults to the lowest allowed value for the input.
13968 Upper bound. Defaults to the highest allowed value for the input.
13971 Specify which planes will be processed. Defaults to all available.
13974 @subsection Commands
13976 This filter supports the all above options as @ref{commands}.
13982 The filter accepts the following options:
13986 Set the number of loops. Setting this value to -1 will result in infinite loops.
13990 Set maximal size in number of frames. Default is 0.
13993 Set first frame of loop. Default is 0.
13996 @subsection Examples
14000 Loop single first frame infinitely:
14002 loop=loop=-1:size=1:start=0
14006 Loop single first frame 10 times:
14008 loop=loop=10:size=1:start=0
14012 Loop 10 first frames 5 times:
14014 loop=loop=5:size=10:start=0
14020 Apply a 1D LUT to an input video.
14022 The filter accepts the following options:
14026 Set the 1D LUT file name.
14028 Currently supported formats:
14037 Select interpolation mode.
14039 Available values are:
14043 Use values from the nearest defined point.
14045 Interpolate values using the linear interpolation.
14047 Interpolate values using the cosine interpolation.
14049 Interpolate values using the cubic interpolation.
14051 Interpolate values using the spline interpolation.
14055 @subsection Commands
14057 This filter supports the all above options as @ref{commands}.
14062 Apply a 3D LUT to an input video.
14064 The filter accepts the following options:
14068 Set the 3D LUT file name.
14070 Currently supported formats:
14084 Select interpolation mode.
14086 Available values are:
14090 Use values from the nearest defined point.
14092 Interpolate values using the 8 points defining a cube.
14094 Interpolate values using a tetrahedron.
14096 Interpolate values using a pyramid.
14098 Interpolate values using a prism.
14102 @subsection Commands
14104 This filter supports the @code{interp} option as @ref{commands}.
14108 Turn certain luma values into transparency.
14110 The filter accepts the following options:
14114 Set the luma which will be used as base for transparency.
14115 Default value is @code{0}.
14118 Set the range of luma values to be keyed out.
14119 Default value is @code{0.01}.
14122 Set the range of softness. Default value is @code{0}.
14123 Use this to control gradual transition from zero to full transparency.
14126 @subsection Commands
14127 This filter supports same @ref{commands} as options.
14128 The command accepts the same syntax of the corresponding option.
14130 If the specified expression is not valid, it is kept at its current
14133 @section lut, lutrgb, lutyuv
14135 Compute a look-up table for binding each pixel component input value
14136 to an output value, and apply it to the input video.
14138 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
14139 to an RGB input video.
14141 These filters accept the following parameters:
14144 set first pixel component expression
14146 set second pixel component expression
14148 set third pixel component expression
14150 set fourth pixel component expression, corresponds to the alpha component
14153 set red component expression
14155 set green component expression
14157 set blue component expression
14159 alpha component expression
14162 set Y/luminance component expression
14164 set U/Cb component expression
14166 set V/Cr component expression
14169 Each of them specifies the expression to use for computing the lookup table for
14170 the corresponding pixel component values.
14172 The exact component associated to each of the @var{c*} options depends on the
14175 The @var{lut} filter requires either YUV or RGB pixel formats in input,
14176 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
14178 The expressions can contain the following constants and functions:
14183 The input width and height.
14186 The input value for the pixel component.
14189 The input value, clipped to the @var{minval}-@var{maxval} range.
14192 The maximum value for the pixel component.
14195 The minimum value for the pixel component.
14198 The negated value for the pixel component value, clipped to the
14199 @var{minval}-@var{maxval} range; it corresponds to the expression
14200 "maxval-clipval+minval".
14203 The computed value in @var{val}, clipped to the
14204 @var{minval}-@var{maxval} range.
14206 @item gammaval(gamma)
14207 The computed gamma correction value of the pixel component value,
14208 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
14210 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
14214 All expressions default to "val".
14216 @subsection Commands
14218 This filter supports same @ref{commands} as options.
14220 @subsection Examples
14224 Negate input video:
14226 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
14227 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
14230 The above is the same as:
14232 lutrgb="r=negval:g=negval:b=negval"
14233 lutyuv="y=negval:u=negval:v=negval"
14243 Remove chroma components, turning the video into a graytone image:
14245 lutyuv="u=128:v=128"
14249 Apply a luma burning effect:
14255 Remove green and blue components:
14261 Set a constant alpha channel value on input:
14263 format=rgba,lutrgb=a="maxval-minval/2"
14267 Correct luminance gamma by a factor of 0.5:
14269 lutyuv=y=gammaval(0.5)
14273 Discard least significant bits of luma:
14275 lutyuv=y='bitand(val, 128+64+32)'
14279 Technicolor like effect:
14281 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
14285 @section lut2, tlut2
14287 The @code{lut2} filter takes two input streams and outputs one
14290 The @code{tlut2} (time lut2) filter takes two consecutive frames
14291 from one single stream.
14293 This filter accepts the following parameters:
14296 set first pixel component expression
14298 set second pixel component expression
14300 set third pixel component expression
14302 set fourth pixel component expression, corresponds to the alpha component
14305 set output bit depth, only available for @code{lut2} filter. By default is 0,
14306 which means bit depth is automatically picked from first input format.
14309 The @code{lut2} filter also supports the @ref{framesync} options.
14311 Each of them specifies the expression to use for computing the lookup table for
14312 the corresponding pixel component values.
14314 The exact component associated to each of the @var{c*} options depends on the
14317 The expressions can contain the following constants:
14322 The input width and height.
14325 The first input value for the pixel component.
14328 The second input value for the pixel component.
14331 The first input video bit depth.
14334 The second input video bit depth.
14337 All expressions default to "x".
14339 @subsection Commands
14341 This filter supports the all above options as @ref{commands} except option @code{d}.
14343 @subsection Examples
14347 Highlight differences between two RGB video streams:
14349 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)'
14353 Highlight differences between two YUV video streams:
14355 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)'
14359 Show max difference between two video streams:
14361 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)))'
14365 @section maskedclamp
14367 Clamp the first input stream with the second input and third input stream.
14369 Returns the value of first stream to be between second input
14370 stream - @code{undershoot} and third input stream + @code{overshoot}.
14372 This filter accepts the following options:
14375 Default value is @code{0}.
14378 Default value is @code{0}.
14381 Set which planes will be processed as bitmap, unprocessed planes will be
14382 copied from first stream.
14383 By default value 0xf, all planes will be processed.
14386 @subsection Commands
14388 This filter supports the all above options as @ref{commands}.
14392 Merge the second and third input stream into output stream using absolute differences
14393 between second input stream and first input stream and absolute difference between
14394 third input stream and first input stream. The picked value will be from second input
14395 stream if second absolute difference is greater than first one or from third input stream
14398 This filter accepts the following options:
14401 Set which planes will be processed as bitmap, unprocessed planes will be
14402 copied from first stream.
14403 By default value 0xf, all planes will be processed.
14406 @subsection Commands
14408 This filter supports the all above options as @ref{commands}.
14410 @section maskedmerge
14412 Merge the first input stream with the second input stream using per pixel
14413 weights in the third input stream.
14415 A value of 0 in the third stream pixel component means that pixel component
14416 from first stream is returned unchanged, while maximum value (eg. 255 for
14417 8-bit videos) means that pixel component from second stream is returned
14418 unchanged. Intermediate values define the amount of merging between both
14419 input stream's pixel components.
14421 This filter accepts the following options:
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 less 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 maskedthreshold
14454 Pick pixels comparing absolute difference of two video streams with fixed
14457 If absolute difference between pixel component of first and second video
14458 stream is equal or lower than user supplied threshold than pixel component
14459 from first video stream is picked, otherwise pixel component from second
14460 video stream is picked.
14462 This filter accepts the following options:
14465 Set threshold used when picking pixels from absolute difference from two input
14469 Set which planes will be processed as bitmap, unprocessed planes will be
14470 copied from second stream.
14471 By default value 0xf, all planes will be processed.
14474 @subsection Commands
14476 This filter supports the all above options as @ref{commands}.
14479 Create mask from input video.
14481 For example it is useful to create motion masks after @code{tblend} filter.
14483 This filter accepts the following options:
14487 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14490 Set high threshold. Any pixel component higher than this value will be set to max value
14491 allowed for current pixel format.
14494 Set planes to filter, by default all available planes are filtered.
14497 Fill all frame pixels with this value.
14500 Set max average pixel value for frame. If sum of all pixel components is higher that this
14501 average, output frame will be completely filled with value set by @var{fill} option.
14502 Typically useful for scene changes when used in combination with @code{tblend} filter.
14505 @subsection Commands
14507 This filter supports the all above options as @ref{commands}.
14511 Apply motion-compensation deinterlacing.
14513 It needs one field per frame as input and must thus be used together
14514 with yadif=1/3 or equivalent.
14516 This filter accepts the following options:
14519 Set the deinterlacing mode.
14521 It accepts one of the following values:
14526 use iterative motion estimation
14528 like @samp{slow}, but use multiple reference frames.
14530 Default value is @samp{fast}.
14533 Set the picture field parity assumed for the input video. It must be
14534 one of the following values:
14538 assume top field first
14540 assume bottom field first
14543 Default value is @samp{bff}.
14546 Set per-block quantization parameter (QP) used by the internal
14549 Higher values should result in a smoother motion vector field but less
14550 optimal individual vectors. Default value is 1.
14555 Pick median pixel from certain rectangle defined by radius.
14557 This filter accepts the following options:
14561 Set horizontal radius size. Default value is @code{1}.
14562 Allowed range is integer from 1 to 127.
14565 Set which planes to process. Default is @code{15}, which is all available planes.
14568 Set vertical radius size. Default value is @code{0}.
14569 Allowed range is integer from 0 to 127.
14570 If it is 0, value will be picked from horizontal @code{radius} option.
14573 Set median percentile. Default value is @code{0.5}.
14574 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14575 minimum values, and @code{1} maximum values.
14578 @subsection Commands
14579 This filter supports same @ref{commands} as options.
14580 The command accepts the same syntax of the corresponding option.
14582 If the specified expression is not valid, it is kept at its current
14585 @section mergeplanes
14587 Merge color channel components from several video streams.
14589 The filter accepts up to 4 input streams, and merge selected input
14590 planes to the output video.
14592 This filter accepts the following options:
14595 Set input to output plane mapping. Default is @code{0}.
14597 The mappings is specified as a bitmap. It should be specified as a
14598 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14599 mapping for the first plane of the output stream. 'A' sets the number of
14600 the input stream to use (from 0 to 3), and 'a' the plane number of the
14601 corresponding input to use (from 0 to 3). The rest of the mappings is
14602 similar, 'Bb' describes the mapping for the output stream second
14603 plane, 'Cc' describes the mapping for the output stream third plane and
14604 'Dd' describes the mapping for the output stream fourth plane.
14607 Set output pixel format. Default is @code{yuva444p}.
14610 @subsection Examples
14614 Merge three gray video streams of same width and height into single video stream:
14616 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14620 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14622 [a0][a1]mergeplanes=0x00010210:yuva444p
14626 Swap Y and A plane in yuva444p stream:
14628 format=yuva444p,mergeplanes=0x03010200:yuva444p
14632 Swap U and V plane in yuv420p stream:
14634 format=yuv420p,mergeplanes=0x000201:yuv420p
14638 Cast a rgb24 clip to yuv444p:
14640 format=rgb24,mergeplanes=0x000102:yuv444p
14646 Estimate and export motion vectors using block matching algorithms.
14647 Motion vectors are stored in frame side data to be used by other filters.
14649 This filter accepts the following options:
14652 Specify the motion estimation method. Accepts one of the following values:
14656 Exhaustive search algorithm.
14658 Three step search algorithm.
14660 Two dimensional logarithmic search algorithm.
14662 New three step search algorithm.
14664 Four step search algorithm.
14666 Diamond search algorithm.
14668 Hexagon-based search algorithm.
14670 Enhanced predictive zonal search algorithm.
14672 Uneven multi-hexagon search algorithm.
14674 Default value is @samp{esa}.
14677 Macroblock size. Default @code{16}.
14680 Search parameter. Default @code{7}.
14683 @section midequalizer
14685 Apply Midway Image Equalization effect using two video streams.
14687 Midway Image Equalization adjusts a pair of images to have the same
14688 histogram, while maintaining their dynamics as much as possible. It's
14689 useful for e.g. matching exposures from a pair of stereo cameras.
14691 This filter has two inputs and one output, which must be of same pixel format, but
14692 may be of different sizes. The output of filter is first input adjusted with
14693 midway histogram of both inputs.
14695 This filter accepts the following option:
14699 Set which planes to process. Default is @code{15}, which is all available planes.
14702 @section minterpolate
14704 Convert the video to specified frame rate using motion interpolation.
14706 This filter accepts the following options:
14709 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}.
14712 Motion interpolation mode. Following values are accepted:
14715 Duplicate previous or next frame for interpolating new ones.
14717 Blend source frames. Interpolated frame is mean of previous and next frames.
14719 Motion compensated interpolation. Following options are effective when this mode is selected:
14723 Motion compensation mode. Following values are accepted:
14726 Overlapped block motion compensation.
14728 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14730 Default mode is @samp{obmc}.
14733 Motion estimation mode. Following values are accepted:
14736 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14738 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14740 Default mode is @samp{bilat}.
14743 The algorithm to be used for motion estimation. Following values are accepted:
14746 Exhaustive search algorithm.
14748 Three step search algorithm.
14750 Two dimensional logarithmic search algorithm.
14752 New three step search algorithm.
14754 Four step search algorithm.
14756 Diamond search algorithm.
14758 Hexagon-based search algorithm.
14760 Enhanced predictive zonal search algorithm.
14762 Uneven multi-hexagon search algorithm.
14764 Default algorithm is @samp{epzs}.
14767 Macroblock size. Default @code{16}.
14770 Motion estimation search parameter. Default @code{32}.
14773 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).
14778 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:
14781 Disable scene change detection.
14783 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14785 Default method is @samp{fdiff}.
14787 @item scd_threshold
14788 Scene change detection threshold. Default is @code{10.}.
14793 Mix several video input streams into one video stream.
14795 A description of the accepted options follows.
14799 The number of inputs. If unspecified, it defaults to 2.
14802 Specify weight of each input video stream as sequence.
14803 Each weight is separated by space. If number of weights
14804 is smaller than number of @var{frames} last specified
14805 weight will be used for all remaining unset weights.
14808 Specify scale, if it is set it will be multiplied with sum
14809 of each weight multiplied with pixel values to give final destination
14810 pixel value. By default @var{scale} is auto scaled to sum of weights.
14813 Specify how end of stream is determined.
14816 The duration of the longest input. (default)
14819 The duration of the shortest input.
14822 The duration of the first input.
14826 @subsection Commands
14828 This filter supports the following commands:
14832 Syntax is same as option with same name.
14835 @section monochrome
14836 Convert video to gray using custom color filter.
14838 A description of the accepted options follows.
14842 Set the chroma blue spot. Allowed range is from -1 to 1.
14843 Default value is 0.
14846 Set the chroma red spot. Allowed range is from -1 to 1.
14847 Default value is 0.
14850 Set the color filter size. Allowed range is from .1 to 10.
14851 Default value is 1.
14854 Set the highlights strength. Allowed range is from 0 to 1.
14855 Default value is 0.
14858 @subsection Commands
14860 This filter supports the all above options as @ref{commands}.
14862 @section mpdecimate
14864 Drop frames that do not differ greatly from the previous frame in
14865 order to reduce frame rate.
14867 The main use of this filter is for very-low-bitrate encoding
14868 (e.g. streaming over dialup modem), but it could in theory be used for
14869 fixing movies that were inverse-telecined incorrectly.
14871 A description of the accepted options follows.
14875 Set the maximum number of consecutive frames which can be dropped (if
14876 positive), or the minimum interval between dropped frames (if
14877 negative). If the value is 0, the frame is dropped disregarding the
14878 number of previous sequentially dropped frames.
14880 Default value is 0.
14885 Set the dropping threshold values.
14887 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14888 represent actual pixel value differences, so a threshold of 64
14889 corresponds to 1 unit of difference for each pixel, or the same spread
14890 out differently over the block.
14892 A frame is a candidate for dropping if no 8x8 blocks differ by more
14893 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14894 meaning the whole image) differ by more than a threshold of @option{lo}.
14896 Default value for @option{hi} is 64*12, default value for @option{lo} is
14897 64*5, and default value for @option{frac} is 0.33.
14902 Obtain the MSAD (Mean Sum of Absolute Differences) between two input videos.
14904 This filter takes two input videos.
14906 Both input videos must have the same resolution and pixel format for
14907 this filter to work correctly. Also it assumes that both inputs
14908 have the same number of frames, which are compared one by one.
14910 The obtained per component, average, min and max MSAD is printed through
14911 the logging system.
14913 The filter stores the calculated MSAD of each frame in frame metadata.
14915 In the below example the input file @file{main.mpg} being processed is compared
14916 with the reference file @file{ref.mpg}.
14919 ffmpeg -i main.mpg -i ref.mpg -lavfi msad -f null -
14924 Negate (invert) the input video.
14926 It accepts the following option:
14931 With value 1, it negates the alpha component, if present. Default value is 0.
14934 @subsection Commands
14936 This filter supports same @ref{commands} as options.
14941 Denoise frames using Non-Local Means algorithm.
14943 Each pixel is adjusted by looking for other pixels with similar contexts. This
14944 context similarity is defined by comparing their surrounding patches of size
14945 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14948 Note that the research area defines centers for patches, which means some
14949 patches will be made of pixels outside that research area.
14951 The filter accepts the following options.
14955 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14958 Set patch size. Default is 7. Must be odd number in range [0, 99].
14961 Same as @option{p} but for chroma planes.
14963 The default value is @var{0} and means automatic.
14966 Set research size. Default is 15. Must be odd number in range [0, 99].
14969 Same as @option{r} but for chroma planes.
14971 The default value is @var{0} and means automatic.
14976 Deinterlace video using neural network edge directed interpolation.
14978 This filter accepts the following options:
14982 Mandatory option, without binary file filter can not work.
14983 Currently file can be found here:
14984 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14987 Set which frames to deinterlace, by default it is @code{all}.
14988 Can be @code{all} or @code{interlaced}.
14991 Set mode of operation.
14993 Can be one of the following:
14997 Use frame flags, both fields.
14999 Use frame flags, single field.
15001 Use top field only.
15003 Use bottom field only.
15005 Use both fields, top first.
15007 Use both fields, bottom first.
15011 Set which planes to process, by default filter process all frames.
15014 Set size of local neighborhood around each pixel, used by the predictor neural
15017 Can be one of the following:
15030 Set the number of neurons in predictor neural network.
15031 Can be one of the following:
15042 Controls the number of different neural network predictions that are blended
15043 together to compute the final output value. Can be @code{fast}, default or
15047 Set which set of weights to use in the predictor.
15048 Can be one of the following:
15052 weights trained to minimize absolute error
15054 weights trained to minimize squared error
15058 Controls whether or not the prescreener neural network is used to decide
15059 which pixels should be processed by the predictor neural network and which
15060 can be handled by simple cubic interpolation.
15061 The prescreener is trained to know whether cubic interpolation will be
15062 sufficient for a pixel or whether it should be predicted by the predictor nn.
15063 The computational complexity of the prescreener nn is much less than that of
15064 the predictor nn. Since most pixels can be handled by cubic interpolation,
15065 using the prescreener generally results in much faster processing.
15066 The prescreener is pretty accurate, so the difference between using it and not
15067 using it is almost always unnoticeable.
15069 Can be one of the following:
15079 Default is @code{new}.
15082 @subsection Commands
15083 This filter supports same @ref{commands} as options, excluding @var{weights} option.
15087 Force libavfilter not to use any of the specified pixel formats for the
15088 input to the next filter.
15090 It accepts the following parameters:
15094 A '|'-separated list of pixel format names, such as
15095 pix_fmts=yuv420p|monow|rgb24".
15099 @subsection Examples
15103 Force libavfilter to use a format different from @var{yuv420p} for the
15104 input to the vflip filter:
15106 noformat=pix_fmts=yuv420p,vflip
15110 Convert the input video to any of the formats not contained in the list:
15112 noformat=yuv420p|yuv444p|yuv410p
15118 Add noise on video input frame.
15120 The filter accepts the following options:
15128 Set noise seed for specific pixel component or all pixel components in case
15129 of @var{all_seed}. Default value is @code{123457}.
15131 @item all_strength, alls
15132 @item c0_strength, c0s
15133 @item c1_strength, c1s
15134 @item c2_strength, c2s
15135 @item c3_strength, c3s
15136 Set noise strength for specific pixel component or all pixel components in case
15137 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
15139 @item all_flags, allf
15140 @item c0_flags, c0f
15141 @item c1_flags, c1f
15142 @item c2_flags, c2f
15143 @item c3_flags, c3f
15144 Set pixel component flags or set flags for all components if @var{all_flags}.
15145 Available values for component flags are:
15148 averaged temporal noise (smoother)
15150 mix random noise with a (semi)regular pattern
15152 temporal noise (noise pattern changes between frames)
15154 uniform noise (gaussian otherwise)
15158 @subsection Examples
15160 Add temporal and uniform noise to input video:
15162 noise=alls=20:allf=t+u
15167 Normalize RGB video (aka histogram stretching, contrast stretching).
15168 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
15170 For each channel of each frame, the filter computes the input range and maps
15171 it linearly to the user-specified output range. The output range defaults
15172 to the full dynamic range from pure black to pure white.
15174 Temporal smoothing can be used on the input range to reduce flickering (rapid
15175 changes in brightness) caused when small dark or bright objects enter or leave
15176 the scene. This is similar to the auto-exposure (automatic gain control) on a
15177 video camera, and, like a video camera, it may cause a period of over- or
15178 under-exposure of the video.
15180 The R,G,B channels can be normalized independently, which may cause some
15181 color shifting, or linked together as a single channel, which prevents
15182 color shifting. Linked normalization preserves hue. Independent normalization
15183 does not, so it can be used to remove some color casts. Independent and linked
15184 normalization can be combined in any ratio.
15186 The normalize filter accepts the following options:
15191 Colors which define the output range. The minimum input value is mapped to
15192 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
15193 The defaults are black and white respectively. Specifying white for
15194 @var{blackpt} and black for @var{whitept} will give color-inverted,
15195 normalized video. Shades of grey can be used to reduce the dynamic range
15196 (contrast). Specifying saturated colors here can create some interesting
15200 The number of previous frames to use for temporal smoothing. The input range
15201 of each channel is smoothed using a rolling average over the current frame
15202 and the @var{smoothing} previous frames. The default is 0 (no temporal
15206 Controls the ratio of independent (color shifting) channel normalization to
15207 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
15208 independent. Defaults to 1.0 (fully independent).
15211 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
15212 expensive no-op. Defaults to 1.0 (full strength).
15216 @subsection Commands
15217 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
15218 The command accepts the same syntax of the corresponding option.
15220 If the specified expression is not valid, it is kept at its current
15223 @subsection Examples
15225 Stretch video contrast to use the full dynamic range, with no temporal
15226 smoothing; may flicker depending on the source content:
15228 normalize=blackpt=black:whitept=white:smoothing=0
15231 As above, but with 50 frames of temporal smoothing; flicker should be
15232 reduced, depending on the source content:
15234 normalize=blackpt=black:whitept=white:smoothing=50
15237 As above, but with hue-preserving linked channel normalization:
15239 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
15242 As above, but with half strength:
15244 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
15247 Map the darkest input color to red, the brightest input color to cyan:
15249 normalize=blackpt=red:whitept=cyan
15254 Pass the video source unchanged to the output.
15257 Optical Character Recognition
15259 This filter uses Tesseract for optical character recognition. To enable
15260 compilation of this filter, you need to configure FFmpeg with
15261 @code{--enable-libtesseract}.
15263 It accepts the following options:
15267 Set datapath to tesseract data. Default is to use whatever was
15268 set at installation.
15271 Set language, default is "eng".
15274 Set character whitelist.
15277 Set character blacklist.
15280 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
15281 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
15285 Apply a video transform using libopencv.
15287 To enable this filter, install the libopencv library and headers and
15288 configure FFmpeg with @code{--enable-libopencv}.
15290 It accepts the following parameters:
15295 The name of the libopencv filter to apply.
15297 @item filter_params
15298 The parameters to pass to the libopencv filter. If not specified, the default
15299 values are assumed.
15303 Refer to the official libopencv documentation for more precise
15305 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
15307 Several libopencv filters are supported; see the following subsections.
15312 Dilate an image by using a specific structuring element.
15313 It corresponds to the libopencv function @code{cvDilate}.
15315 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
15317 @var{struct_el} represents a structuring element, and has the syntax:
15318 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
15320 @var{cols} and @var{rows} represent the number of columns and rows of
15321 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
15322 point, and @var{shape} the shape for the structuring element. @var{shape}
15323 must be "rect", "cross", "ellipse", or "custom".
15325 If the value for @var{shape} is "custom", it must be followed by a
15326 string of the form "=@var{filename}". The file with name
15327 @var{filename} is assumed to represent a binary image, with each
15328 printable character corresponding to a bright pixel. When a custom
15329 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
15330 or columns and rows of the read file are assumed instead.
15332 The default value for @var{struct_el} is "3x3+0x0/rect".
15334 @var{nb_iterations} specifies the number of times the transform is
15335 applied to the image, and defaults to 1.
15339 # Use the default values
15342 # Dilate using a structuring element with a 5x5 cross, iterating two times
15343 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
15345 # Read the shape from the file diamond.shape, iterating two times.
15346 # The file diamond.shape may contain a pattern of characters like this
15352 # The specified columns and rows are ignored
15353 # but the anchor point coordinates are not
15354 ocv=dilate:0x0+2x2/custom=diamond.shape|2
15359 Erode an image by using a specific structuring element.
15360 It corresponds to the libopencv function @code{cvErode}.
15362 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
15363 with the same syntax and semantics as the @ref{dilate} filter.
15367 Smooth the input video.
15369 The filter takes the following parameters:
15370 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
15372 @var{type} is the type of smooth filter to apply, and must be one of
15373 the following values: "blur", "blur_no_scale", "median", "gaussian",
15374 or "bilateral". The default value is "gaussian".
15376 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
15377 depends on the smooth type. @var{param1} and
15378 @var{param2} accept integer positive values or 0. @var{param3} and
15379 @var{param4} accept floating point values.
15381 The default value for @var{param1} is 3. The default value for the
15382 other parameters is 0.
15384 These parameters correspond to the parameters assigned to the
15385 libopencv function @code{cvSmooth}.
15387 @section oscilloscope
15389 2D Video Oscilloscope.
15391 Useful to measure spatial impulse, step responses, chroma delays, etc.
15393 It accepts the following parameters:
15397 Set scope center x position.
15400 Set scope center y position.
15403 Set scope size, relative to frame diagonal.
15406 Set scope tilt/rotation.
15412 Set trace center x position.
15415 Set trace center y position.
15418 Set trace width, relative to width of frame.
15421 Set trace height, relative to height of frame.
15424 Set which components to trace. By default it traces first three components.
15427 Draw trace grid. By default is enabled.
15430 Draw some statistics. By default is enabled.
15433 Draw scope. By default is enabled.
15436 @subsection Commands
15437 This filter supports same @ref{commands} as options.
15438 The command accepts the same syntax of the corresponding option.
15440 If the specified expression is not valid, it is kept at its current
15443 @subsection Examples
15447 Inspect full first row of video frame.
15449 oscilloscope=x=0.5:y=0:s=1
15453 Inspect full last row of video frame.
15455 oscilloscope=x=0.5:y=1:s=1
15459 Inspect full 5th line of video frame of height 1080.
15461 oscilloscope=x=0.5:y=5/1080:s=1
15465 Inspect full last column of video frame.
15467 oscilloscope=x=1:y=0.5:s=1:t=1
15475 Overlay one video on top of another.
15477 It takes two inputs and has one output. The first input is the "main"
15478 video on which the second input is overlaid.
15480 It accepts the following parameters:
15482 A description of the accepted options follows.
15487 Set the expression for the x and y coordinates of the overlaid video
15488 on the main video. Default value is "0" for both expressions. In case
15489 the expression is invalid, it is set to a huge value (meaning that the
15490 overlay will not be displayed within the output visible area).
15493 See @ref{framesync}.
15496 Set when the expressions for @option{x}, and @option{y} are evaluated.
15498 It accepts the following values:
15501 only evaluate expressions once during the filter initialization or
15502 when a command is processed
15505 evaluate expressions for each incoming frame
15508 Default value is @samp{frame}.
15511 See @ref{framesync}.
15514 Set the format for the output video.
15516 It accepts the following values:
15519 force YUV420 output
15522 force YUV420p10 output
15525 force YUV422 output
15528 force YUV422p10 output
15531 force YUV444 output
15534 force packed RGB output
15537 force planar RGB output
15540 automatically pick format
15543 Default value is @samp{yuv420}.
15546 See @ref{framesync}.
15549 Set format of alpha of the overlaid video, it can be @var{straight} or
15550 @var{premultiplied}. Default is @var{straight}.
15553 The @option{x}, and @option{y} expressions can contain the following
15559 The main input width and height.
15563 The overlay input width and height.
15567 The computed values for @var{x} and @var{y}. They are evaluated for
15572 horizontal and vertical chroma subsample values of the output
15573 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15577 the number of input frame, starting from 0
15580 the position in the file of the input frame, NAN if unknown
15583 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15587 This filter also supports the @ref{framesync} options.
15589 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15590 when evaluation is done @emph{per frame}, and will evaluate to NAN
15591 when @option{eval} is set to @samp{init}.
15593 Be aware that frames are taken from each input video in timestamp
15594 order, hence, if their initial timestamps differ, it is a good idea
15595 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15596 have them begin in the same zero timestamp, as the example for
15597 the @var{movie} filter does.
15599 You can chain together more overlays but you should test the
15600 efficiency of such approach.
15602 @subsection Commands
15604 This filter supports the following commands:
15608 Modify the x and y of the overlay input.
15609 The command accepts the same syntax of the corresponding option.
15611 If the specified expression is not valid, it is kept at its current
15615 @subsection Examples
15619 Draw the overlay at 10 pixels from the bottom right corner of the main
15622 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15625 Using named options the example above becomes:
15627 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15631 Insert a transparent PNG logo in the bottom left corner of the input,
15632 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15634 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15638 Insert 2 different transparent PNG logos (second logo on bottom
15639 right corner) using the @command{ffmpeg} tool:
15641 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
15645 Add a transparent color layer on top of the main video; @code{WxH}
15646 must specify the size of the main input to the overlay filter:
15648 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15652 Play an original video and a filtered version (here with the deshake
15653 filter) side by side using the @command{ffplay} tool:
15655 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15658 The above command is the same as:
15660 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15664 Make a sliding overlay appearing from the left to the right top part of the
15665 screen starting since time 2:
15667 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15671 Compose output by putting two input videos side to side:
15673 ffmpeg -i left.avi -i right.avi -filter_complex "
15674 nullsrc=size=200x100 [background];
15675 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15676 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15677 [background][left] overlay=shortest=1 [background+left];
15678 [background+left][right] overlay=shortest=1:x=100 [left+right]
15683 Mask 10-20 seconds of a video by applying the delogo filter to a section
15685 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15686 -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]'
15691 Chain several overlays in cascade:
15693 nullsrc=s=200x200 [bg];
15694 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15695 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15696 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15697 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15698 [in3] null, [mid2] overlay=100:100 [out0]
15703 @anchor{overlay_cuda}
15704 @section overlay_cuda
15706 Overlay one video on top of another.
15708 This is the CUDA variant of the @ref{overlay} filter.
15709 It only accepts CUDA frames. The underlying input pixel formats have to match.
15711 It takes two inputs and has one output. The first input is the "main"
15712 video on which the second input is overlaid.
15714 It accepts the following parameters:
15719 Set the x and y coordinates of the overlaid video on the main video.
15720 Default value is "0" for both expressions.
15723 See @ref{framesync}.
15726 See @ref{framesync}.
15729 See @ref{framesync}.
15733 This filter also supports the @ref{framesync} options.
15737 Apply Overcomplete Wavelet denoiser.
15739 The filter accepts the following options:
15745 Larger depth values will denoise lower frequency components more, but
15746 slow down filtering.
15748 Must be an int in the range 8-16, default is @code{8}.
15750 @item luma_strength, ls
15753 Must be a double value in the range 0-1000, default is @code{1.0}.
15755 @item chroma_strength, cs
15756 Set chroma strength.
15758 Must be a double value in the range 0-1000, default is @code{1.0}.
15764 Add paddings to the input image, and place the original input at the
15765 provided @var{x}, @var{y} coordinates.
15767 It accepts the following parameters:
15772 Specify an expression for the size of the output image with the
15773 paddings added. If the value for @var{width} or @var{height} is 0, the
15774 corresponding input size is used for the output.
15776 The @var{width} expression can reference the value set by the
15777 @var{height} expression, and vice versa.
15779 The default value of @var{width} and @var{height} is 0.
15783 Specify the offsets to place the input image at within the padded area,
15784 with respect to the top/left border of the output image.
15786 The @var{x} expression can reference the value set by the @var{y}
15787 expression, and vice versa.
15789 The default value of @var{x} and @var{y} is 0.
15791 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15792 so the input image is centered on the padded area.
15795 Specify the color of the padded area. For the syntax of this option,
15796 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15797 manual,ffmpeg-utils}.
15799 The default value of @var{color} is "black".
15802 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15804 It accepts the following values:
15808 Only evaluate expressions once during the filter initialization or when
15809 a command is processed.
15812 Evaluate expressions for each incoming frame.
15816 Default value is @samp{init}.
15819 Pad to aspect instead to a resolution.
15823 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15824 options are expressions containing the following constants:
15829 The input video width and height.
15833 These are the same as @var{in_w} and @var{in_h}.
15837 The output width and height (the size of the padded area), as
15838 specified by the @var{width} and @var{height} expressions.
15842 These are the same as @var{out_w} and @var{out_h}.
15846 The x and y offsets as specified by the @var{x} and @var{y}
15847 expressions, or NAN if not yet specified.
15850 same as @var{iw} / @var{ih}
15853 input sample aspect ratio
15856 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15860 The horizontal and vertical chroma subsample values. For example for the
15861 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15864 @subsection Examples
15868 Add paddings with the color "violet" to the input video. The output video
15869 size is 640x480, and the top-left corner of the input video is placed at
15872 pad=640:480:0:40:violet
15875 The example above is equivalent to the following command:
15877 pad=width=640:height=480:x=0:y=40:color=violet
15881 Pad the input to get an output with dimensions increased by 3/2,
15882 and put the input video at the center of the padded area:
15884 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15888 Pad the input to get a squared output with size equal to the maximum
15889 value between the input width and height, and put the input video at
15890 the center of the padded area:
15892 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15896 Pad the input to get a final w/h ratio of 16:9:
15898 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15902 In case of anamorphic video, in order to set the output display aspect
15903 correctly, it is necessary to use @var{sar} in the expression,
15904 according to the relation:
15906 (ih * X / ih) * sar = output_dar
15907 X = output_dar / sar
15910 Thus the previous example needs to be modified to:
15912 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15916 Double the output size and put the input video in the bottom-right
15917 corner of the output padded area:
15919 pad="2*iw:2*ih:ow-iw:oh-ih"
15923 @anchor{palettegen}
15924 @section palettegen
15926 Generate one palette for a whole video stream.
15928 It accepts the following options:
15932 Set the maximum number of colors to quantize in the palette.
15933 Note: the palette will still contain 256 colors; the unused palette entries
15936 @item reserve_transparent
15937 Create a palette of 255 colors maximum and reserve the last one for
15938 transparency. Reserving the transparency color is useful for GIF optimization.
15939 If not set, the maximum of colors in the palette will be 256. You probably want
15940 to disable this option for a standalone image.
15943 @item transparency_color
15944 Set the color that will be used as background for transparency.
15947 Set statistics mode.
15949 It accepts the following values:
15952 Compute full frame histograms.
15954 Compute histograms only for the part that differs from previous frame. This
15955 might be relevant to give more importance to the moving part of your input if
15956 the background is static.
15958 Compute new histogram for each frame.
15961 Default value is @var{full}.
15964 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15965 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15966 color quantization of the palette. This information is also visible at
15967 @var{info} logging level.
15969 @subsection Examples
15973 Generate a representative palette of a given video using @command{ffmpeg}:
15975 ffmpeg -i input.mkv -vf palettegen palette.png
15979 @section paletteuse
15981 Use a palette to downsample an input video stream.
15983 The filter takes two inputs: one video stream and a palette. The palette must
15984 be a 256 pixels image.
15986 It accepts the following options:
15990 Select dithering mode. Available algorithms are:
15993 Ordered 8x8 bayer dithering (deterministic)
15995 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15996 Note: this dithering is sometimes considered "wrong" and is included as a
15998 @item floyd_steinberg
15999 Floyd and Steingberg dithering (error diffusion)
16001 Frankie Sierra dithering v2 (error diffusion)
16003 Frankie Sierra dithering v2 "Lite" (error diffusion)
16006 Default is @var{sierra2_4a}.
16009 When @var{bayer} dithering is selected, this option defines the scale of the
16010 pattern (how much the crosshatch pattern is visible). A low value means more
16011 visible pattern for less banding, and higher value means less visible pattern
16012 at the cost of more banding.
16014 The option must be an integer value in the range [0,5]. Default is @var{2}.
16017 If set, define the zone to process
16021 Only the changing rectangle will be reprocessed. This is similar to GIF
16022 cropping/offsetting compression mechanism. This option can be useful for speed
16023 if only a part of the image is changing, and has use cases such as limiting the
16024 scope of the error diffusal @option{dither} to the rectangle that bounds the
16025 moving scene (it leads to more deterministic output if the scene doesn't change
16026 much, and as a result less moving noise and better GIF compression).
16029 Default is @var{none}.
16032 Take new palette for each output frame.
16034 @item alpha_threshold
16035 Sets the alpha threshold for transparency. Alpha values above this threshold
16036 will be treated as completely opaque, and values below this threshold will be
16037 treated as completely transparent.
16039 The option must be an integer value in the range [0,255]. Default is @var{128}.
16042 @subsection Examples
16046 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
16047 using @command{ffmpeg}:
16049 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
16053 @section perspective
16055 Correct perspective of video not recorded perpendicular to the screen.
16057 A description of the accepted parameters follows.
16068 Set coordinates expression for top left, top right, bottom left and bottom right corners.
16069 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
16070 If the @code{sense} option is set to @code{source}, then the specified points will be sent
16071 to the corners of the destination. If the @code{sense} option is set to @code{destination},
16072 then the corners of the source will be sent to the specified coordinates.
16074 The expressions can use the following variables:
16079 the width and height of video frame.
16083 Output frame count.
16086 @item interpolation
16087 Set interpolation for perspective correction.
16089 It accepts the following values:
16095 Default value is @samp{linear}.
16098 Set interpretation of coordinate options.
16100 It accepts the following values:
16104 Send point in the source specified by the given coordinates to
16105 the corners of the destination.
16107 @item 1, destination
16109 Send the corners of the source to the point in the destination specified
16110 by the given coordinates.
16112 Default value is @samp{source}.
16116 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
16118 It accepts the following values:
16121 only evaluate expressions once during the filter initialization or
16122 when a command is processed
16125 evaluate expressions for each incoming frame
16128 Default value is @samp{init}.
16133 Delay interlaced video by one field time so that the field order changes.
16135 The intended use is to fix PAL movies that have been captured with the
16136 opposite field order to the film-to-video transfer.
16138 A description of the accepted parameters follows.
16144 It accepts the following values:
16147 Capture field order top-first, transfer bottom-first.
16148 Filter will delay the bottom field.
16151 Capture field order bottom-first, transfer top-first.
16152 Filter will delay the top field.
16155 Capture and transfer with the same field order. This mode only exists
16156 for the documentation of the other options to refer to, but if you
16157 actually select it, the filter will faithfully do nothing.
16160 Capture field order determined automatically by field flags, transfer
16162 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
16163 basis using field flags. If no field information is available,
16164 then this works just like @samp{u}.
16167 Capture unknown or varying, transfer opposite.
16168 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
16169 analyzing the images and selecting the alternative that produces best
16170 match between the fields.
16173 Capture top-first, transfer unknown or varying.
16174 Filter selects among @samp{t} and @samp{p} using image analysis.
16177 Capture bottom-first, transfer unknown or varying.
16178 Filter selects among @samp{b} and @samp{p} using image analysis.
16181 Capture determined by field flags, transfer unknown or varying.
16182 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
16183 image analysis. If no field information is available, then this works just
16184 like @samp{U}. This is the default mode.
16187 Both capture and transfer unknown or varying.
16188 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
16192 @subsection Commands
16194 This filter supports the all above options as @ref{commands}.
16196 @section photosensitivity
16197 Reduce various flashes in video, so to help users with epilepsy.
16199 It accepts the following options:
16202 Set how many frames to use when filtering. Default is 30.
16205 Set detection threshold factor. Default is 1.
16209 Set how many pixels to skip when sampling frames. Default is 1.
16210 Allowed range is from 1 to 1024.
16213 Leave frames unchanged. Default is disabled.
16216 @section pixdesctest
16218 Pixel format descriptor test filter, mainly useful for internal
16219 testing. The output video should be equal to the input video.
16223 format=monow, pixdesctest
16226 can be used to test the monowhite pixel format descriptor definition.
16230 Display sample values of color channels. Mainly useful for checking color
16231 and levels. Minimum supported resolution is 640x480.
16233 The filters accept the following options:
16237 Set scope X position, relative offset on X axis.
16240 Set scope Y position, relative offset on Y axis.
16249 Set window opacity. This window also holds statistics about pixel area.
16252 Set window X position, relative offset on X axis.
16255 Set window Y position, relative offset on Y axis.
16258 @subsection Commands
16260 This filter supports same @ref{commands} as options.
16264 Enable the specified chain of postprocessing subfilters using libpostproc. This
16265 library should be automatically selected with a GPL build (@code{--enable-gpl}).
16266 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
16267 Each subfilter and some options have a short and a long name that can be used
16268 interchangeably, i.e. dr/dering are the same.
16270 The filters accept the following options:
16274 Set postprocessing subfilters string.
16277 All subfilters share common options to determine their scope:
16281 Honor the quality commands for this subfilter.
16284 Do chrominance filtering, too (default).
16287 Do luminance filtering only (no chrominance).
16290 Do chrominance filtering only (no luminance).
16293 These options can be appended after the subfilter name, separated by a '|'.
16295 Available subfilters are:
16298 @item hb/hdeblock[|difference[|flatness]]
16299 Horizontal deblocking filter
16302 Difference factor where higher values mean more deblocking (default: @code{32}).
16304 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16307 @item vb/vdeblock[|difference[|flatness]]
16308 Vertical deblocking filter
16311 Difference factor where higher values mean more deblocking (default: @code{32}).
16313 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16316 @item ha/hadeblock[|difference[|flatness]]
16317 Accurate horizontal deblocking filter
16320 Difference factor where higher values mean more deblocking (default: @code{32}).
16322 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16325 @item va/vadeblock[|difference[|flatness]]
16326 Accurate vertical deblocking filter
16329 Difference factor where higher values mean more deblocking (default: @code{32}).
16331 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16335 The horizontal and vertical deblocking filters share the difference and
16336 flatness values so you cannot set different horizontal and vertical
16340 @item h1/x1hdeblock
16341 Experimental horizontal deblocking filter
16343 @item v1/x1vdeblock
16344 Experimental vertical deblocking filter
16349 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
16352 larger -> stronger filtering
16354 larger -> stronger filtering
16356 larger -> stronger filtering
16359 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
16362 Stretch luminance to @code{0-255}.
16365 @item lb/linblenddeint
16366 Linear blend deinterlacing filter that deinterlaces the given block by
16367 filtering all lines with a @code{(1 2 1)} filter.
16369 @item li/linipoldeint
16370 Linear interpolating deinterlacing filter that deinterlaces the given block by
16371 linearly interpolating every second line.
16373 @item ci/cubicipoldeint
16374 Cubic interpolating deinterlacing filter deinterlaces the given block by
16375 cubically interpolating every second line.
16377 @item md/mediandeint
16378 Median deinterlacing filter that deinterlaces the given block by applying a
16379 median filter to every second line.
16381 @item fd/ffmpegdeint
16382 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
16383 second line with a @code{(-1 4 2 4 -1)} filter.
16386 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
16387 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
16389 @item fq/forceQuant[|quantizer]
16390 Overrides the quantizer table from the input with the constant quantizer you
16398 Default pp filter combination (@code{hb|a,vb|a,dr|a})
16401 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
16404 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
16407 @subsection Examples
16411 Apply horizontal and vertical deblocking, deringing and automatic
16412 brightness/contrast:
16418 Apply default filters without brightness/contrast correction:
16424 Apply default filters and temporal denoiser:
16426 pp=default/tmpnoise|1|2|3
16430 Apply deblocking on luminance only, and switch vertical deblocking on or off
16431 automatically depending on available CPU time:
16438 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
16439 similar to spp = 6 with 7 point DCT, where only the center sample is
16442 The filter accepts the following options:
16446 Force a constant quantization parameter. It accepts an integer in range
16447 0 to 63. If not set, the filter will use the QP from the video stream
16451 Set thresholding mode. Available modes are:
16455 Set hard thresholding.
16457 Set soft thresholding (better de-ringing effect, but likely blurrier).
16459 Set medium thresholding (good results, default).
16463 @section premultiply
16464 Apply alpha premultiply effect to input video stream using first plane
16465 of second stream as alpha.
16467 Both streams must have same dimensions and same pixel format.
16469 The filter accepts the following option:
16473 Set which planes will be processed, unprocessed planes will be copied.
16474 By default value 0xf, all planes will be processed.
16477 Do not require 2nd input for processing, instead use alpha plane from input stream.
16481 Apply prewitt operator to input video stream.
16483 The filter accepts the following option:
16487 Set which planes will be processed, unprocessed planes will be copied.
16488 By default value 0xf, all planes will be processed.
16491 Set value which will be multiplied with filtered result.
16494 Set value which will be added to filtered result.
16497 @subsection Commands
16499 This filter supports the all above options as @ref{commands}.
16501 @section pseudocolor
16503 Alter frame colors in video with pseudocolors.
16505 This filter accepts the following options:
16509 set pixel first component expression
16512 set pixel second component expression
16515 set pixel third component expression
16518 set pixel fourth component expression, corresponds to the alpha component
16521 set component to use as base for altering colors
16524 Pick one of built-in LUTs. By default is set to none.
16541 Set opacity of output colors. Allowed range is from 0 to 1.
16542 Default value is set to 1.
16545 Each of the expression options specifies the expression to use for computing
16546 the lookup table for the corresponding pixel component values.
16548 The expressions can contain the following constants and functions:
16553 The input width and height.
16556 The input value for the pixel component.
16558 @item ymin, umin, vmin, amin
16559 The minimum allowed component value.
16561 @item ymax, umax, vmax, amax
16562 The maximum allowed component value.
16565 All expressions default to "val".
16567 @subsection Commands
16569 This filter supports the all above options as @ref{commands}.
16571 @subsection Examples
16575 Change too high luma values to gradient:
16577 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'"
16583 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16584 Ratio) between two input videos.
16586 This filter takes in input two input videos, the first input is
16587 considered the "main" source and is passed unchanged to the
16588 output. The second input is used as a "reference" video for computing
16591 Both video inputs must have the same resolution and pixel format for
16592 this filter to work correctly. Also it assumes that both inputs
16593 have the same number of frames, which are compared one by one.
16595 The obtained average PSNR is printed through the logging system.
16597 The filter stores the accumulated MSE (mean squared error) of each
16598 frame, and at the end of the processing it is averaged across all frames
16599 equally, and the following formula is applied to obtain the PSNR:
16602 PSNR = 10*log10(MAX^2/MSE)
16605 Where MAX is the average of the maximum values of each component of the
16608 The description of the accepted parameters follows.
16611 @item stats_file, f
16612 If specified the filter will use the named file to save the PSNR of
16613 each individual frame. When filename equals "-" the data is sent to
16616 @item stats_version
16617 Specifies which version of the stats file format to use. Details of
16618 each format are written below.
16619 Default value is 1.
16621 @item stats_add_max
16622 Determines whether the max value is output to the stats log.
16623 Default value is 0.
16624 Requires stats_version >= 2. If this is set and stats_version < 2,
16625 the filter will return an error.
16628 This filter also supports the @ref{framesync} options.
16630 The file printed if @var{stats_file} is selected, contains a sequence of
16631 key/value pairs of the form @var{key}:@var{value} for each compared
16634 If a @var{stats_version} greater than 1 is specified, a header line precedes
16635 the list of per-frame-pair stats, with key value pairs following the frame
16636 format with the following parameters:
16639 @item psnr_log_version
16640 The version of the log file format. Will match @var{stats_version}.
16643 A comma separated list of the per-frame-pair parameters included in
16647 A description of each shown per-frame-pair parameter follows:
16651 sequential number of the input frame, starting from 1
16654 Mean Square Error pixel-by-pixel average difference of the compared
16655 frames, averaged over all the image components.
16657 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16658 Mean Square Error pixel-by-pixel average difference of the compared
16659 frames for the component specified by the suffix.
16661 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16662 Peak Signal to Noise ratio of the compared frames for the component
16663 specified by the suffix.
16665 @item max_avg, max_y, max_u, max_v
16666 Maximum allowed value for each channel, and average over all
16670 @subsection Examples
16675 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16676 [main][ref] psnr="stats_file=stats.log" [out]
16679 On this example the input file being processed is compared with the
16680 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16681 is stored in @file{stats.log}.
16684 Another example with different containers:
16686 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 -
16693 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16694 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16697 The pullup filter is designed to take advantage of future context in making
16698 its decisions. This filter is stateless in the sense that it does not lock
16699 onto a pattern to follow, but it instead looks forward to the following
16700 fields in order to identify matches and rebuild progressive frames.
16702 To produce content with an even framerate, insert the fps filter after
16703 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16704 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16706 The filter accepts the following options:
16713 These options set the amount of "junk" to ignore at the left, right, top, and
16714 bottom of the image, respectively. Left and right are in units of 8 pixels,
16715 while top and bottom are in units of 2 lines.
16716 The default is 8 pixels on each side.
16719 Set the strict breaks. Setting this option to 1 will reduce the chances of
16720 filter generating an occasional mismatched frame, but it may also cause an
16721 excessive number of frames to be dropped during high motion sequences.
16722 Conversely, setting it to -1 will make filter match fields more easily.
16723 This may help processing of video where there is slight blurring between
16724 the fields, but may also cause there to be interlaced frames in the output.
16725 Default value is @code{0}.
16728 Set the metric plane to use. It accepts the following values:
16734 Use chroma blue plane.
16737 Use chroma red plane.
16740 This option may be set to use chroma plane instead of the default luma plane
16741 for doing filter's computations. This may improve accuracy on very clean
16742 source material, but more likely will decrease accuracy, especially if there
16743 is chroma noise (rainbow effect) or any grayscale video.
16744 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16745 load and make pullup usable in realtime on slow machines.
16748 For best results (without duplicated frames in the output file) it is
16749 necessary to change the output frame rate. For example, to inverse
16750 telecine NTSC input:
16752 ffmpeg -i input -vf pullup -r 24000/1001 ...
16757 Change video quantization parameters (QP).
16759 The filter accepts the following option:
16763 Set expression for quantization parameter.
16766 The expression is evaluated through the eval API and can contain, among others,
16767 the following constants:
16771 1 if index is not 129, 0 otherwise.
16774 Sequential index starting from -129 to 128.
16777 @subsection Examples
16781 Some equation like:
16789 Flush video frames from internal cache of frames into a random order.
16790 No frame is discarded.
16791 Inspired by @ref{frei0r} nervous filter.
16795 Set size in number of frames of internal cache, in range from @code{2} to
16796 @code{512}. Default is @code{30}.
16799 Set seed for random number generator, must be an integer included between
16800 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16801 less than @code{0}, the filter will try to use a good random seed on a
16805 @section readeia608
16807 Read closed captioning (EIA-608) information from the top lines of a video frame.
16809 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16810 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16811 with EIA-608 data (starting from 0). A description of each metadata value follows:
16814 @item lavfi.readeia608.X.cc
16815 The two bytes stored as EIA-608 data (printed in hexadecimal).
16817 @item lavfi.readeia608.X.line
16818 The number of the line on which the EIA-608 data was identified and read.
16821 This filter accepts the following options:
16825 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16828 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16831 Set the ratio of width reserved for sync code detection.
16832 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16835 Enable checking the parity bit. In the event of a parity error, the filter will output
16836 @code{0x00} for that character. Default is false.
16839 Lowpass lines prior to further processing. Default is enabled.
16842 @subsection Commands
16844 This filter supports the all above options as @ref{commands}.
16846 @subsection Examples
16850 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16852 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
16858 Read vertical interval timecode (VITC) information from the top lines of a
16861 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16862 timecode value, if a valid timecode has been detected. Further metadata key
16863 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16864 timecode data has been found or not.
16866 This filter accepts the following options:
16870 Set the maximum number of lines to scan for VITC data. If the value is set to
16871 @code{-1} the full video frame is scanned. Default is @code{45}.
16874 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16875 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16878 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16879 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16882 @subsection Examples
16886 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16887 draw @code{--:--:--:--} as a placeholder:
16889 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16895 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16897 Destination pixel at position (X, Y) will be picked from source (x, y) position
16898 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16899 value for pixel will be used for destination pixel.
16901 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16902 will have Xmap/Ymap video stream dimensions.
16903 Xmap and Ymap input video streams are 16bit depth, single channel.
16907 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16908 Default is @code{color}.
16911 Specify the color of the unmapped pixels. For the syntax of this option,
16912 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16913 manual,ffmpeg-utils}. Default color is @code{black}.
16916 @section removegrain
16918 The removegrain filter is a spatial denoiser for progressive video.
16922 Set mode for the first plane.
16925 Set mode for the second plane.
16928 Set mode for the third plane.
16931 Set mode for the fourth plane.
16934 Range of mode is from 0 to 24. Description of each mode follows:
16938 Leave input plane unchanged. Default.
16941 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16944 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16947 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16950 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16951 This is equivalent to a median filter.
16954 Line-sensitive clipping giving the minimal change.
16957 Line-sensitive clipping, intermediate.
16960 Line-sensitive clipping, intermediate.
16963 Line-sensitive clipping, intermediate.
16966 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16969 Replaces the target pixel with the closest neighbour.
16972 [1 2 1] horizontal and vertical kernel blur.
16978 Bob mode, interpolates top field from the line where the neighbours
16979 pixels are the closest.
16982 Bob mode, interpolates bottom field from the line where the neighbours
16983 pixels are the closest.
16986 Bob mode, interpolates top field. Same as 13 but with a more complicated
16987 interpolation formula.
16990 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16991 interpolation formula.
16994 Clips the pixel with the minimum and maximum of respectively the maximum and
16995 minimum of each pair of opposite neighbour pixels.
16998 Line-sensitive clipping using opposite neighbours whose greatest distance from
16999 the current pixel is minimal.
17002 Replaces the pixel with the average of its 8 neighbours.
17005 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
17008 Clips pixels using the averages of opposite neighbour.
17011 Same as mode 21 but simpler and faster.
17014 Small edge and halo removal, but reputed useless.
17020 @section removelogo
17022 Suppress a TV station logo, using an image file to determine which
17023 pixels comprise the logo. It works by filling in the pixels that
17024 comprise the logo with neighboring pixels.
17026 The filter accepts the following options:
17030 Set the filter bitmap file, which can be any image format supported by
17031 libavformat. The width and height of the image file must match those of the
17032 video stream being processed.
17035 Pixels in the provided bitmap image with a value of zero are not
17036 considered part of the logo, non-zero pixels are considered part of
17037 the logo. If you use white (255) for the logo and black (0) for the
17038 rest, you will be safe. For making the filter bitmap, it is
17039 recommended to take a screen capture of a black frame with the logo
17040 visible, and then using a threshold filter followed by the erode
17041 filter once or twice.
17043 If needed, little splotches can be fixed manually. Remember that if
17044 logo pixels are not covered, the filter quality will be much
17045 reduced. Marking too many pixels as part of the logo does not hurt as
17046 much, but it will increase the amount of blurring needed to cover over
17047 the image and will destroy more information than necessary, and extra
17048 pixels will slow things down on a large logo.
17050 @section repeatfields
17052 This filter uses the repeat_field flag from the Video ES headers and hard repeats
17053 fields based on its value.
17057 Reverse a video clip.
17059 Warning: This filter requires memory to buffer the entire clip, so trimming
17062 @subsection Examples
17066 Take the first 5 seconds of a clip, and reverse it.
17073 Shift R/G/B/A pixels horizontally and/or vertically.
17075 The filter accepts the following options:
17078 Set amount to shift red horizontally.
17080 Set amount to shift red vertically.
17082 Set amount to shift green horizontally.
17084 Set amount to shift green vertically.
17086 Set amount to shift blue horizontally.
17088 Set amount to shift blue vertically.
17090 Set amount to shift alpha horizontally.
17092 Set amount to shift alpha vertically.
17094 Set edge mode, can be @var{smear}, default, or @var{warp}.
17097 @subsection Commands
17099 This filter supports the all above options as @ref{commands}.
17102 Apply roberts cross operator to input video stream.
17104 The filter accepts the following option:
17108 Set which planes will be processed, unprocessed planes will be copied.
17109 By default value 0xf, all planes will be processed.
17112 Set value which will be multiplied with filtered result.
17115 Set value which will be added to filtered result.
17118 @subsection Commands
17120 This filter supports the all above options as @ref{commands}.
17124 Rotate video by an arbitrary angle expressed in radians.
17126 The filter accepts the following options:
17128 A description of the optional parameters follows.
17131 Set an expression for the angle by which to rotate the input video
17132 clockwise, expressed as a number of radians. A negative value will
17133 result in a counter-clockwise rotation. By default it is set to "0".
17135 This expression is evaluated for each frame.
17138 Set the output width expression, default value is "iw".
17139 This expression is evaluated just once during configuration.
17142 Set the output height expression, default value is "ih".
17143 This expression is evaluated just once during configuration.
17146 Enable bilinear interpolation if set to 1, a value of 0 disables
17147 it. Default value is 1.
17150 Set the color used to fill the output area not covered by the rotated
17151 image. For the general syntax of this option, check the
17152 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
17153 If the special value "none" is selected then no
17154 background is printed (useful for example if the background is never shown).
17156 Default value is "black".
17159 The expressions for the angle and the output size can contain the
17160 following constants and functions:
17164 sequential number of the input frame, starting from 0. It is always NAN
17165 before the first frame is filtered.
17168 time in seconds of the input frame, it is set to 0 when the filter is
17169 configured. It is always NAN before the first frame is filtered.
17173 horizontal and vertical chroma subsample values. For example for the
17174 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17178 the input video width and height
17182 the output width and height, that is the size of the padded area as
17183 specified by the @var{width} and @var{height} expressions
17187 the minimal width/height required for completely containing the input
17188 video rotated by @var{a} radians.
17190 These are only available when computing the @option{out_w} and
17191 @option{out_h} expressions.
17194 @subsection Examples
17198 Rotate the input by PI/6 radians clockwise:
17204 Rotate the input by PI/6 radians counter-clockwise:
17210 Rotate the input by 45 degrees clockwise:
17216 Apply a constant rotation with period T, starting from an angle of PI/3:
17218 rotate=PI/3+2*PI*t/T
17222 Make the input video rotation oscillating with a period of T
17223 seconds and an amplitude of A radians:
17225 rotate=A*sin(2*PI/T*t)
17229 Rotate the video, output size is chosen so that the whole rotating
17230 input video is always completely contained in the output:
17232 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
17236 Rotate the video, reduce the output size so that no background is ever
17239 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
17243 @subsection Commands
17245 The filter supports the following commands:
17249 Set the angle expression.
17250 The command accepts the same syntax of the corresponding option.
17252 If the specified expression is not valid, it is kept at its current
17258 Apply Shape Adaptive Blur.
17260 The filter accepts the following options:
17263 @item luma_radius, lr
17264 Set luma blur filter strength, must be a value in range 0.1-4.0, default
17265 value is 1.0. A greater value will result in a more blurred image, and
17266 in slower processing.
17268 @item luma_pre_filter_radius, lpfr
17269 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
17272 @item luma_strength, ls
17273 Set luma maximum difference between pixels to still be considered, must
17274 be a value in the 0.1-100.0 range, default value is 1.0.
17276 @item chroma_radius, cr
17277 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
17278 greater value will result in a more blurred image, and in slower
17281 @item chroma_pre_filter_radius, cpfr
17282 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
17284 @item chroma_strength, cs
17285 Set chroma maximum difference between pixels to still be considered,
17286 must be a value in the -0.9-100.0 range.
17289 Each chroma option value, if not explicitly specified, is set to the
17290 corresponding luma option value.
17295 Scale (resize) the input video, using the libswscale library.
17297 The scale filter forces the output display aspect ratio to be the same
17298 of the input, by changing the output sample aspect ratio.
17300 If the input image format is different from the format requested by
17301 the next filter, the scale filter will convert the input to the
17304 @subsection Options
17305 The filter accepts the following options, or any of the options
17306 supported by the libswscale scaler.
17308 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
17309 the complete list of scaler options.
17314 Set the output video dimension expression. Default value is the input
17317 If the @var{width} or @var{w} value is 0, the input width is used for
17318 the output. If the @var{height} or @var{h} value is 0, the input height
17319 is used for the output.
17321 If one and only one of the values is -n with n >= 1, the scale filter
17322 will use a value that maintains the aspect ratio of the input image,
17323 calculated from the other specified dimension. After that it will,
17324 however, make sure that the calculated dimension is divisible by n and
17325 adjust the value if necessary.
17327 If both values are -n with n >= 1, the behavior will be identical to
17328 both values being set to 0 as previously detailed.
17330 See below for the list of accepted constants for use in the dimension
17334 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
17338 Only evaluate expressions once during the filter initialization or when a command is processed.
17341 Evaluate expressions for each incoming frame.
17345 Default value is @samp{init}.
17349 Set the interlacing mode. It accepts the following values:
17353 Force interlaced aware scaling.
17356 Do not apply interlaced scaling.
17359 Select interlaced aware scaling depending on whether the source frames
17360 are flagged as interlaced or not.
17363 Default value is @samp{0}.
17366 Set libswscale scaling flags. See
17367 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
17368 complete list of values. If not explicitly specified the filter applies
17372 @item param0, param1
17373 Set libswscale input parameters for scaling algorithms that need them. See
17374 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
17375 complete documentation. If not explicitly specified the filter applies
17381 Set the video size. For the syntax of this option, check the
17382 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17384 @item in_color_matrix
17385 @item out_color_matrix
17386 Set in/output YCbCr color space type.
17388 This allows the autodetected value to be overridden as well as allows forcing
17389 a specific value used for the output and encoder.
17391 If not specified, the color space type depends on the pixel format.
17397 Choose automatically.
17400 Format conforming to International Telecommunication Union (ITU)
17401 Recommendation BT.709.
17404 Set color space conforming to the United States Federal Communications
17405 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
17410 Set color space conforming to:
17414 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
17417 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
17420 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
17425 Set color space conforming to SMPTE ST 240:1999.
17428 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
17433 Set in/output YCbCr sample range.
17435 This allows the autodetected value to be overridden as well as allows forcing
17436 a specific value used for the output and encoder. If not specified, the
17437 range depends on the pixel format. Possible values:
17441 Choose automatically.
17444 Set full range (0-255 in case of 8-bit luma).
17446 @item mpeg/limited/tv
17447 Set "MPEG" range (16-235 in case of 8-bit luma).
17450 @item force_original_aspect_ratio
17451 Enable decreasing or increasing output video width or height if necessary to
17452 keep the original aspect ratio. Possible values:
17456 Scale the video as specified and disable this feature.
17459 The output video dimensions will automatically be decreased if needed.
17462 The output video dimensions will automatically be increased if needed.
17466 One useful instance of this option is that when you know a specific device's
17467 maximum allowed resolution, you can use this to limit the output video to
17468 that, while retaining the aspect ratio. For example, device A allows
17469 1280x720 playback, and your video is 1920x800. Using this option (set it to
17470 decrease) and specifying 1280x720 to the command line makes the output
17473 Please note that this is a different thing than specifying -1 for @option{w}
17474 or @option{h}, you still need to specify the output resolution for this option
17477 @item force_divisible_by
17478 Ensures that both the output dimensions, width and height, are divisible by the
17479 given integer when used together with @option{force_original_aspect_ratio}. This
17480 works similar to using @code{-n} in the @option{w} and @option{h} options.
17482 This option respects the value set for @option{force_original_aspect_ratio},
17483 increasing or decreasing the resolution accordingly. The video's aspect ratio
17484 may be slightly modified.
17486 This option can be handy if you need to have a video fit within or exceed
17487 a defined resolution using @option{force_original_aspect_ratio} but also have
17488 encoder restrictions on width or height divisibility.
17492 The values of the @option{w} and @option{h} options are expressions
17493 containing the following constants:
17498 The input width and height
17502 These are the same as @var{in_w} and @var{in_h}.
17506 The output (scaled) width and height
17510 These are the same as @var{out_w} and @var{out_h}
17513 The same as @var{iw} / @var{ih}
17516 input sample aspect ratio
17519 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17523 horizontal and vertical input chroma subsample values. For example for the
17524 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17528 horizontal and vertical output chroma subsample values. For example for the
17529 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17532 The (sequential) number of the input frame, starting from 0.
17533 Only available with @code{eval=frame}.
17536 The presentation timestamp of the input frame, expressed as a number of
17537 seconds. Only available with @code{eval=frame}.
17540 The position (byte offset) of the frame in the input stream, or NaN if
17541 this information is unavailable and/or meaningless (for example in case of synthetic video).
17542 Only available with @code{eval=frame}.
17545 @subsection Examples
17549 Scale the input video to a size of 200x100
17554 This is equivalent to:
17565 Specify a size abbreviation for the output size:
17570 which can also be written as:
17576 Scale the input to 2x:
17578 scale=w=2*iw:h=2*ih
17582 The above is the same as:
17584 scale=2*in_w:2*in_h
17588 Scale the input to 2x with forced interlaced scaling:
17590 scale=2*iw:2*ih:interl=1
17594 Scale the input to half size:
17596 scale=w=iw/2:h=ih/2
17600 Increase the width, and set the height to the same size:
17606 Seek Greek harmony:
17613 Increase the height, and set the width to 3/2 of the height:
17615 scale=w=3/2*oh:h=3/5*ih
17619 Increase the size, making the size a multiple of the chroma
17622 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17626 Increase the width to a maximum of 500 pixels,
17627 keeping the same aspect ratio as the input:
17629 scale=w='min(500\, iw*3/2):h=-1'
17633 Make pixels square by combining scale and setsar:
17635 scale='trunc(ih*dar):ih',setsar=1/1
17639 Make pixels square by combining scale and setsar,
17640 making sure the resulting resolution is even (required by some codecs):
17642 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17646 @subsection Commands
17648 This filter supports the following commands:
17652 Set the output video dimension expression.
17653 The command accepts the same syntax of the corresponding option.
17655 If the specified expression is not valid, it is kept at its current
17661 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17662 format conversion on CUDA video frames. Setting the output width and height
17663 works in the same way as for the @var{scale} filter.
17665 The following additional options are accepted:
17668 The pixel format of the output CUDA frames. If set to the string "same" (the
17669 default), the input format will be kept. Note that automatic format negotiation
17670 and conversion is not yet supported for hardware frames
17673 The interpolation algorithm used for resizing. One of the following:
17680 @item cubic2p_bspline
17681 2-parameter cubic (B=1, C=0)
17683 @item cubic2p_catmullrom
17684 2-parameter cubic (B=0, C=1/2)
17686 @item cubic2p_b05c03
17687 2-parameter cubic (B=1/2, C=3/10)
17695 @item force_original_aspect_ratio
17696 Enable decreasing or increasing output video width or height if necessary to
17697 keep the original aspect ratio. Possible values:
17701 Scale the video as specified and disable this feature.
17704 The output video dimensions will automatically be decreased if needed.
17707 The output video dimensions will automatically be increased if needed.
17711 One useful instance of this option is that when you know a specific device's
17712 maximum allowed resolution, you can use this to limit the output video to
17713 that, while retaining the aspect ratio. For example, device A allows
17714 1280x720 playback, and your video is 1920x800. Using this option (set it to
17715 decrease) and specifying 1280x720 to the command line makes the output
17718 Please note that this is a different thing than specifying -1 for @option{w}
17719 or @option{h}, you still need to specify the output resolution for this option
17722 @item force_divisible_by
17723 Ensures that both the output dimensions, width and height, are divisible by the
17724 given integer when used together with @option{force_original_aspect_ratio}. This
17725 works similar to using @code{-n} in the @option{w} and @option{h} options.
17727 This option respects the value set for @option{force_original_aspect_ratio},
17728 increasing or decreasing the resolution accordingly. The video's aspect ratio
17729 may be slightly modified.
17731 This option can be handy if you need to have a video fit within or exceed
17732 a defined resolution using @option{force_original_aspect_ratio} but also have
17733 encoder restrictions on width or height divisibility.
17739 Scale (resize) the input video, based on a reference video.
17741 See the scale filter for available options, scale2ref supports the same but
17742 uses the reference video instead of the main input as basis. scale2ref also
17743 supports the following additional constants for the @option{w} and
17744 @option{h} options:
17749 The main input video's width and height
17752 The same as @var{main_w} / @var{main_h}
17755 The main input video's sample aspect ratio
17757 @item main_dar, mdar
17758 The main input video's display aspect ratio. Calculated from
17759 @code{(main_w / main_h) * main_sar}.
17763 The main input video's horizontal and vertical chroma subsample values.
17764 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17768 The (sequential) number of the main input frame, starting from 0.
17769 Only available with @code{eval=frame}.
17772 The presentation timestamp of the main input frame, expressed as a number of
17773 seconds. Only available with @code{eval=frame}.
17776 The position (byte offset) of the frame in the main input stream, or NaN if
17777 this information is unavailable and/or meaningless (for example in case of synthetic video).
17778 Only available with @code{eval=frame}.
17781 @subsection Examples
17785 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17787 'scale2ref[b][a];[a][b]overlay'
17791 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17793 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17797 @subsection Commands
17799 This filter supports the following commands:
17803 Set the output video dimension expression.
17804 The command accepts the same syntax of the corresponding option.
17806 If the specified expression is not valid, it is kept at its current
17811 Scroll input video horizontally and/or vertically by constant speed.
17813 The filter accepts the following options:
17815 @item horizontal, h
17816 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17817 Negative values changes scrolling direction.
17820 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17821 Negative values changes scrolling direction.
17824 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17827 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17830 @subsection Commands
17832 This filter supports the following @ref{commands}:
17834 @item horizontal, h
17835 Set the horizontal scrolling speed.
17837 Set the vertical scrolling speed.
17843 Detect video scene change.
17845 This filter sets frame metadata with mafd between frame, the scene score, and
17846 forward the frame to the next filter, so they can use these metadata to detect
17847 scene change or others.
17849 In addition, this filter logs a message and sets frame metadata when it detects
17850 a scene change by @option{threshold}.
17852 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17854 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17855 to detect scene change.
17857 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17858 detect scene change with @option{threshold}.
17860 The filter accepts the following options:
17864 Set the scene change detection threshold as a percentage of maximum change. Good
17865 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17868 Default value is @code{10.}.
17871 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17872 You can enable it if you want to get snapshot of scene change frames only.
17875 @anchor{selectivecolor}
17876 @section selectivecolor
17878 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17879 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17880 by the "purity" of the color (that is, how saturated it already is).
17882 This filter is similar to the Adobe Photoshop Selective Color tool.
17884 The filter accepts the following options:
17887 @item correction_method
17888 Select color correction method.
17890 Available values are:
17893 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17896 Specified adjustments are relative to the original component value.
17898 Default is @code{absolute}.
17900 Adjustments for red pixels (pixels where the red component is the maximum)
17902 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17904 Adjustments for green pixels (pixels where the green component is the maximum)
17906 Adjustments for cyan pixels (pixels where the red component is the minimum)
17908 Adjustments for blue pixels (pixels where the blue component is the maximum)
17910 Adjustments for magenta pixels (pixels where the green component is the minimum)
17912 Adjustments for white pixels (pixels where all components are greater than 128)
17914 Adjustments for all pixels except pure black and pure white
17916 Adjustments for black pixels (pixels where all components are lesser than 128)
17918 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17921 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17922 4 space separated floating point adjustment values in the [-1,1] range,
17923 respectively to adjust the amount of cyan, magenta, yellow and black for the
17924 pixels of its range.
17926 @subsection Examples
17930 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17931 increase magenta by 27% in blue areas:
17933 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17937 Use a Photoshop selective color preset:
17939 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17943 @anchor{separatefields}
17944 @section separatefields
17946 The @code{separatefields} takes a frame-based video input and splits
17947 each frame into its components fields, producing a new half height clip
17948 with twice the frame rate and twice the frame count.
17950 This filter use field-dominance information in frame to decide which
17951 of each pair of fields to place first in the output.
17952 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17954 @section setdar, setsar
17956 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17959 This is done by changing the specified Sample (aka Pixel) Aspect
17960 Ratio, according to the following equation:
17962 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17965 Keep in mind that the @code{setdar} filter does not modify the pixel
17966 dimensions of the video frame. Also, the display aspect ratio set by
17967 this filter may be changed by later filters in the filterchain,
17968 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17971 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17972 the filter output video.
17974 Note that as a consequence of the application of this filter, the
17975 output display aspect ratio will change according to the equation
17978 Keep in mind that the sample aspect ratio set by the @code{setsar}
17979 filter may be changed by later filters in the filterchain, e.g. if
17980 another "setsar" or a "setdar" filter is applied.
17982 It accepts the following parameters:
17985 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17986 Set the aspect ratio used by the filter.
17988 The parameter can be a floating point number string, an expression, or
17989 a string of the form @var{num}:@var{den}, where @var{num} and
17990 @var{den} are the numerator and denominator of the aspect ratio. If
17991 the parameter is not specified, it is assumed the value "0".
17992 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17996 Set the maximum integer value to use for expressing numerator and
17997 denominator when reducing the expressed aspect ratio to a rational.
17998 Default value is @code{100}.
18002 The parameter @var{sar} is an expression containing
18003 the following constants:
18007 These are approximated values for the mathematical constants e
18008 (Euler's number), pi (Greek pi), and phi (the golden ratio).
18011 The input width and height.
18014 These are the same as @var{w} / @var{h}.
18017 The input sample aspect ratio.
18020 The input display aspect ratio. It is the same as
18021 (@var{w} / @var{h}) * @var{sar}.
18024 Horizontal and vertical chroma subsample values. For example, for the
18025 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
18028 @subsection Examples
18033 To change the display aspect ratio to 16:9, specify one of the following:
18040 To change the sample aspect ratio to 10:11, specify:
18046 To set a display aspect ratio of 16:9, and specify a maximum integer value of
18047 1000 in the aspect ratio reduction, use the command:
18049 setdar=ratio=16/9:max=1000
18057 Force field for the output video frame.
18059 The @code{setfield} filter marks the interlace type field for the
18060 output frames. It does not change the input frame, but only sets the
18061 corresponding property, which affects how the frame is treated by
18062 following filters (e.g. @code{fieldorder} or @code{yadif}).
18064 The filter accepts the following options:
18069 Available values are:
18073 Keep the same field property.
18076 Mark the frame as bottom-field-first.
18079 Mark the frame as top-field-first.
18082 Mark the frame as progressive.
18089 Force frame parameter for the output video frame.
18091 The @code{setparams} filter marks interlace and color range for the
18092 output frames. It does not change the input frame, but only sets the
18093 corresponding property, which affects how the frame is treated by
18098 Available values are:
18102 Keep the same field property (default).
18105 Mark the frame as bottom-field-first.
18108 Mark the frame as top-field-first.
18111 Mark the frame as progressive.
18115 Available values are:
18119 Keep the same color range property (default).
18121 @item unspecified, unknown
18122 Mark the frame as unspecified color range.
18124 @item limited, tv, mpeg
18125 Mark the frame as limited range.
18127 @item full, pc, jpeg
18128 Mark the frame as full range.
18131 @item color_primaries
18132 Set the color primaries.
18133 Available values are:
18137 Keep the same color primaries property (default).
18154 Set the color transfer.
18155 Available values are:
18159 Keep the same color trc property (default).
18181 Set the colorspace.
18182 Available values are:
18186 Keep the same colorspace property (default).
18199 @item chroma-derived-nc
18200 @item chroma-derived-c
18206 Apply shear transform to input video.
18208 This filter supports the following options:
18212 Shear factor in X-direction. Default value is 0.
18213 Allowed range is from -2 to 2.
18216 Shear factor in Y-direction. Default value is 0.
18217 Allowed range is from -2 to 2.
18220 Set the color used to fill the output area not covered by the transformed
18221 video. For the general syntax of this option, check the
18222 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18223 If the special value "none" is selected then no
18224 background is printed (useful for example if the background is never shown).
18226 Default value is "black".
18229 Set interpolation type. Can be @code{bilinear} or @code{nearest}. Default is @code{bilinear}.
18232 @subsection Commands
18234 This filter supports the all above options as @ref{commands}.
18238 Show a line containing various information for each input video frame.
18239 The input video is not modified.
18241 This filter supports the following options:
18245 Calculate checksums of each plane. By default enabled.
18248 The shown line contains a sequence of key/value pairs of the form
18249 @var{key}:@var{value}.
18251 The following values are shown in the output:
18255 The (sequential) number of the input frame, starting from 0.
18258 The Presentation TimeStamp of the input frame, expressed as a number of
18259 time base units. The time base unit depends on the filter input pad.
18262 The Presentation TimeStamp of the input frame, expressed as a number of
18266 The position of the frame in the input stream, or -1 if this information is
18267 unavailable and/or meaningless (for example in case of synthetic video).
18270 The pixel format name.
18273 The sample aspect ratio of the input frame, expressed in the form
18274 @var{num}/@var{den}.
18277 The size of the input frame. For the syntax of this option, check the
18278 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18281 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
18282 for bottom field first).
18285 This is 1 if the frame is a key frame, 0 otherwise.
18288 The picture type of the input frame ("I" for an I-frame, "P" for a
18289 P-frame, "B" for a B-frame, or "?" for an unknown type).
18290 Also refer to the documentation of the @code{AVPictureType} enum and of
18291 the @code{av_get_picture_type_char} function defined in
18292 @file{libavutil/avutil.h}.
18295 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
18297 @item plane_checksum
18298 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
18299 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
18302 The mean value of pixels in each plane of the input frame, expressed in the form
18303 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
18306 The standard deviation of pixel values in each plane of the input frame, expressed
18307 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
18311 @section showpalette
18313 Displays the 256 colors palette of each frame. This filter is only relevant for
18314 @var{pal8} pixel format frames.
18316 It accepts the following option:
18320 Set the size of the box used to represent one palette color entry. Default is
18321 @code{30} (for a @code{30x30} pixel box).
18324 @section shuffleframes
18326 Reorder and/or duplicate and/or drop video frames.
18328 It accepts the following parameters:
18332 Set the destination indexes of input frames.
18333 This is space or '|' separated list of indexes that maps input frames to output
18334 frames. Number of indexes also sets maximal value that each index may have.
18335 '-1' index have special meaning and that is to drop frame.
18338 The first frame has the index 0. The default is to keep the input unchanged.
18340 @subsection Examples
18344 Swap second and third frame of every three frames of the input:
18346 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
18350 Swap 10th and 1st frame of every ten frames of the input:
18352 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
18356 @section shufflepixels
18358 Reorder pixels in video frames.
18360 This filter accepts the following options:
18364 Set shuffle direction. Can be forward or inverse direction.
18365 Default direction is forward.
18368 Set shuffle mode. Can be horizontal, vertical or block mode.
18372 Set shuffle block_size. In case of horizontal shuffle mode only width
18373 part of size is used, and in case of vertical shuffle mode only height
18374 part of size is used.
18377 Set random seed used with shuffling pixels. Mainly useful to set to be able
18378 to reverse filtering process to get original input.
18379 For example, to reverse forward shuffle you need to use same parameters
18380 and exact same seed and to set direction to inverse.
18383 @section shuffleplanes
18385 Reorder and/or duplicate video planes.
18387 It accepts the following parameters:
18392 The index of the input plane to be used as the first output plane.
18395 The index of the input plane to be used as the second output plane.
18398 The index of the input plane to be used as the third output plane.
18401 The index of the input plane to be used as the fourth output plane.
18405 The first plane has the index 0. The default is to keep the input unchanged.
18407 @subsection Examples
18411 Swap the second and third planes of the input:
18413 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
18417 @anchor{signalstats}
18418 @section signalstats
18419 Evaluate various visual metrics that assist in determining issues associated
18420 with the digitization of analog video media.
18422 By default the filter will log these metadata values:
18426 Display the minimal Y value contained within the input frame. Expressed in
18430 Display the Y value at the 10% percentile within the input frame. Expressed in
18434 Display the average Y value within the input frame. Expressed in range of
18438 Display the Y value at the 90% percentile within the input frame. Expressed in
18442 Display the maximum Y value contained within the input frame. Expressed in
18446 Display the minimal U value contained within the input frame. Expressed in
18450 Display the U value at the 10% percentile within the input frame. Expressed in
18454 Display the average U value within the input frame. Expressed in range of
18458 Display the U value at the 90% percentile within the input frame. Expressed in
18462 Display the maximum U value contained within the input frame. Expressed in
18466 Display the minimal V value contained within the input frame. Expressed in
18470 Display the V value at the 10% percentile within the input frame. Expressed in
18474 Display the average V value within the input frame. Expressed in range of
18478 Display the V value at the 90% percentile within the input frame. Expressed in
18482 Display the maximum V value contained within the input frame. Expressed in
18486 Display the minimal saturation value contained within the input frame.
18487 Expressed in range of [0-~181.02].
18490 Display the saturation value at the 10% percentile within the input frame.
18491 Expressed in range of [0-~181.02].
18494 Display the average saturation value within the input frame. Expressed in range
18498 Display the saturation value at the 90% percentile within the input frame.
18499 Expressed in range of [0-~181.02].
18502 Display the maximum saturation value contained within the input frame.
18503 Expressed in range of [0-~181.02].
18506 Display the median value for hue within the input frame. Expressed in range of
18510 Display the average value for hue within the input frame. Expressed in range of
18514 Display the average of sample value difference between all values of the Y
18515 plane in the current frame and corresponding values of the previous input frame.
18516 Expressed in range of [0-255].
18519 Display the average of sample value difference between all values of the U
18520 plane in the current frame and corresponding values of the previous input frame.
18521 Expressed in range of [0-255].
18524 Display the average of sample value difference between all values of the V
18525 plane in the current frame and corresponding values of the previous input frame.
18526 Expressed in range of [0-255].
18529 Display bit depth of Y plane in current frame.
18530 Expressed in range of [0-16].
18533 Display bit depth of U plane in current frame.
18534 Expressed in range of [0-16].
18537 Display bit depth of V plane in current frame.
18538 Expressed in range of [0-16].
18541 The filter accepts the following options:
18547 @option{stat} specify an additional form of image analysis.
18548 @option{out} output video with the specified type of pixel highlighted.
18550 Both options accept the following values:
18554 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18555 unlike the neighboring pixels of the same field. Examples of temporal outliers
18556 include the results of video dropouts, head clogs, or tape tracking issues.
18559 Identify @var{vertical line repetition}. Vertical line repetition includes
18560 similar rows of pixels within a frame. In born-digital video vertical line
18561 repetition is common, but this pattern is uncommon in video digitized from an
18562 analog source. When it occurs in video that results from the digitization of an
18563 analog source it can indicate concealment from a dropout compensator.
18566 Identify pixels that fall outside of legal broadcast range.
18570 Set the highlight color for the @option{out} option. The default color is
18574 @subsection Examples
18578 Output data of various video metrics:
18580 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18584 Output specific data about the minimum and maximum values of the Y plane per frame:
18586 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18590 Playback video while highlighting pixels that are outside of broadcast range in red.
18592 ffplay example.mov -vf signalstats="out=brng:color=red"
18596 Playback video with signalstats metadata drawn over the frame.
18598 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18601 The contents of signalstat_drawtext.txt used in the command are:
18604 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18605 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18606 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18607 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18615 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18616 input. In this case the matching between the inputs can be calculated additionally.
18617 The filter always passes through the first input. The signature of each stream can
18618 be written into a file.
18620 It accepts the following options:
18624 Enable or disable the matching process.
18626 Available values are:
18630 Disable the calculation of a matching (default).
18632 Calculate the matching for the whole video and output whether the whole video
18633 matches or only parts.
18635 Calculate only until a matching is found or the video ends. Should be faster in
18640 Set the number of inputs. The option value must be a non negative integer.
18641 Default value is 1.
18644 Set the path to which the output is written. If there is more than one input,
18645 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18646 integer), that will be replaced with the input number. If no filename is
18647 specified, no output will be written. This is the default.
18650 Choose the output format.
18652 Available values are:
18656 Use the specified binary representation (default).
18658 Use the specified xml representation.
18662 Set threshold to detect one word as similar. The option value must be an integer
18663 greater than zero. The default value is 9000.
18666 Set threshold to detect all words as similar. The option value must be an integer
18667 greater than zero. The default value is 60000.
18670 Set threshold to detect frames as similar. The option value must be an integer
18671 greater than zero. The default value is 116.
18674 Set the minimum length of a sequence in frames to recognize it as matching
18675 sequence. The option value must be a non negative integer value.
18676 The default value is 0.
18679 Set the minimum relation, that matching frames to all frames must have.
18680 The option value must be a double value between 0 and 1. The default value is 0.5.
18683 @subsection Examples
18687 To calculate the signature of an input video and store it in signature.bin:
18689 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18693 To detect whether two videos match and store the signatures in XML format in
18694 signature0.xml and signature1.xml:
18696 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 -
18704 Blur the input video without impacting the outlines.
18706 It accepts the following options:
18709 @item luma_radius, lr
18710 Set the luma radius. The option value must be a float number in
18711 the range [0.1,5.0] that specifies the variance of the gaussian filter
18712 used to blur the image (slower if larger). Default value is 1.0.
18714 @item luma_strength, ls
18715 Set the luma strength. The option value must be a float number
18716 in the range [-1.0,1.0] that configures the blurring. A value included
18717 in [0.0,1.0] will blur the image whereas a value included in
18718 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18720 @item luma_threshold, lt
18721 Set the luma threshold used as a coefficient to determine
18722 whether a pixel should be blurred or not. The option value must be an
18723 integer in the range [-30,30]. A value of 0 will filter all the image,
18724 a value included in [0,30] will filter flat areas and a value included
18725 in [-30,0] will filter edges. Default value is 0.
18727 @item chroma_radius, cr
18728 Set the chroma radius. The option value must be a float number in
18729 the range [0.1,5.0] that specifies the variance of the gaussian filter
18730 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18732 @item chroma_strength, cs
18733 Set the chroma strength. The option value must be a float number
18734 in the range [-1.0,1.0] that configures the blurring. A value included
18735 in [0.0,1.0] will blur the image whereas a value included in
18736 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18738 @item chroma_threshold, ct
18739 Set the chroma threshold used as a coefficient to determine
18740 whether a pixel should be blurred or not. The option value must be an
18741 integer in the range [-30,30]. A value of 0 will filter all the image,
18742 a value included in [0,30] will filter flat areas and a value included
18743 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18746 If a chroma option is not explicitly set, the corresponding luma value
18750 Apply sobel operator to input video stream.
18752 The filter accepts the following option:
18756 Set which planes will be processed, unprocessed planes will be copied.
18757 By default value 0xf, all planes will be processed.
18760 Set value which will be multiplied with filtered result.
18763 Set value which will be added to filtered result.
18766 @subsection Commands
18768 This filter supports the all above options as @ref{commands}.
18773 Apply a simple postprocessing filter that compresses and decompresses the image
18774 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18775 and average the results.
18777 The filter accepts the following options:
18781 Set quality. This option defines the number of levels for averaging. It accepts
18782 an integer in the range 0-6. If set to @code{0}, the filter will have no
18783 effect. A value of @code{6} means the higher quality. For each increment of
18784 that value the speed drops by a factor of approximately 2. Default value is
18788 Force a constant quantization parameter. If not set, the filter will use the QP
18789 from the video stream (if available).
18792 Set thresholding mode. Available modes are:
18796 Set hard thresholding (default).
18798 Set soft thresholding (better de-ringing effect, but likely blurrier).
18801 @item use_bframe_qp
18802 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18803 option may cause flicker since the B-Frames have often larger QP. Default is
18804 @code{0} (not enabled).
18807 @subsection Commands
18809 This filter supports the following commands:
18811 @item quality, level
18812 Set quality level. The value @code{max} can be used to set the maximum level,
18813 currently @code{6}.
18819 Scale the input by applying one of the super-resolution methods based on
18820 convolutional neural networks. Supported models:
18824 Super-Resolution Convolutional Neural Network model (SRCNN).
18825 See @url{https://arxiv.org/abs/1501.00092}.
18828 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18829 See @url{https://arxiv.org/abs/1609.05158}.
18832 Training scripts as well as scripts for model file (.pb) saving can be found at
18833 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18834 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18836 Native model files (.model) can be generated from TensorFlow model
18837 files (.pb) by using tools/python/convert.py
18839 The filter accepts the following options:
18843 Specify which DNN backend to use for model loading and execution. This option accepts
18844 the following values:
18848 Native implementation of DNN loading and execution.
18851 TensorFlow backend. To enable this backend you
18852 need to install the TensorFlow for C library (see
18853 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18854 @code{--enable-libtensorflow}
18857 Default value is @samp{native}.
18860 Set path to model file specifying network architecture and its parameters.
18861 Note that different backends use different file formats. TensorFlow backend
18862 can load files for both formats, while native backend can load files for only
18866 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18867 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18868 input upscaled using bicubic upscaling with proper scale factor.
18871 This feature can also be finished with @ref{dnn_processing} filter.
18875 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18877 This filter takes in input two input videos, the first input is
18878 considered the "main" source and is passed unchanged to the
18879 output. The second input is used as a "reference" video for computing
18882 Both video inputs must have the same resolution and pixel format for
18883 this filter to work correctly. Also it assumes that both inputs
18884 have the same number of frames, which are compared one by one.
18886 The filter stores the calculated SSIM of each frame.
18888 The description of the accepted parameters follows.
18891 @item stats_file, f
18892 If specified the filter will use the named file to save the SSIM of
18893 each individual frame. When filename equals "-" the data is sent to
18897 The file printed if @var{stats_file} is selected, contains a sequence of
18898 key/value pairs of the form @var{key}:@var{value} for each compared
18901 A description of each shown parameter follows:
18905 sequential number of the input frame, starting from 1
18907 @item Y, U, V, R, G, B
18908 SSIM of the compared frames for the component specified by the suffix.
18911 SSIM of the compared frames for the whole frame.
18914 Same as above but in dB representation.
18917 This filter also supports the @ref{framesync} options.
18919 @subsection Examples
18924 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18925 [main][ref] ssim="stats_file=stats.log" [out]
18928 On this example the input file being processed is compared with the
18929 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18930 is stored in @file{stats.log}.
18933 Another example with both psnr and ssim at same time:
18935 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18939 Another example with different containers:
18941 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 -
18947 Convert between different stereoscopic image formats.
18949 The filters accept the following options:
18953 Set stereoscopic image format of input.
18955 Available values for input image formats are:
18958 side by side parallel (left eye left, right eye right)
18961 side by side crosseye (right eye left, left eye right)
18964 side by side parallel with half width resolution
18965 (left eye left, right eye right)
18968 side by side crosseye with half width resolution
18969 (right eye left, left eye right)
18973 above-below (left eye above, right eye below)
18977 above-below (right eye above, left eye below)
18981 above-below with half height resolution
18982 (left eye above, right eye below)
18986 above-below with half height resolution
18987 (right eye above, left eye below)
18990 alternating frames (left eye first, right eye second)
18993 alternating frames (right eye first, left eye second)
18996 interleaved rows (left eye has top row, right eye starts on next row)
18999 interleaved rows (right eye has top row, left eye starts on next row)
19002 interleaved columns, left eye first
19005 interleaved columns, right eye first
19007 Default value is @samp{sbsl}.
19011 Set stereoscopic image format of output.
19015 side by side parallel (left eye left, right eye right)
19018 side by side crosseye (right eye left, left eye right)
19021 side by side parallel with half width resolution
19022 (left eye left, right eye right)
19025 side by side crosseye with half width resolution
19026 (right eye left, left eye right)
19030 above-below (left eye above, right eye below)
19034 above-below (right eye above, left eye below)
19038 above-below with half height resolution
19039 (left eye above, right eye below)
19043 above-below with half height resolution
19044 (right eye above, left eye below)
19047 alternating frames (left eye first, right eye second)
19050 alternating frames (right eye first, left eye second)
19053 interleaved rows (left eye has top row, right eye starts on next row)
19056 interleaved rows (right eye has top row, left eye starts on next row)
19059 anaglyph red/blue gray
19060 (red filter on left eye, blue filter on right eye)
19063 anaglyph red/green gray
19064 (red filter on left eye, green filter on right eye)
19067 anaglyph red/cyan gray
19068 (red filter on left eye, cyan filter on right eye)
19071 anaglyph red/cyan half colored
19072 (red filter on left eye, cyan filter on right eye)
19075 anaglyph red/cyan color
19076 (red filter on left eye, cyan filter on right eye)
19079 anaglyph red/cyan color optimized with the least squares projection of dubois
19080 (red filter on left eye, cyan filter on right eye)
19083 anaglyph green/magenta gray
19084 (green filter on left eye, magenta filter on right eye)
19087 anaglyph green/magenta half colored
19088 (green filter on left eye, magenta filter on right eye)
19091 anaglyph green/magenta colored
19092 (green filter on left eye, magenta filter on right eye)
19095 anaglyph green/magenta color optimized with the least squares projection of dubois
19096 (green filter on left eye, magenta filter on right eye)
19099 anaglyph yellow/blue gray
19100 (yellow filter on left eye, blue filter on right eye)
19103 anaglyph yellow/blue half colored
19104 (yellow filter on left eye, blue filter on right eye)
19107 anaglyph yellow/blue colored
19108 (yellow filter on left eye, blue filter on right eye)
19111 anaglyph yellow/blue color optimized with the least squares projection of dubois
19112 (yellow filter on left eye, blue filter on right eye)
19115 mono output (left eye only)
19118 mono output (right eye only)
19121 checkerboard, left eye first
19124 checkerboard, right eye first
19127 interleaved columns, left eye first
19130 interleaved columns, right eye first
19136 Default value is @samp{arcd}.
19139 @subsection Examples
19143 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
19149 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
19155 @section streamselect, astreamselect
19156 Select video or audio streams.
19158 The filter accepts the following options:
19162 Set number of inputs. Default is 2.
19165 Set input indexes to remap to outputs.
19168 @subsection Commands
19170 The @code{streamselect} and @code{astreamselect} filter supports the following
19175 Set input indexes to remap to outputs.
19178 @subsection Examples
19182 Select first 5 seconds 1st stream and rest of time 2nd stream:
19184 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
19188 Same as above, but for audio:
19190 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
19197 Draw subtitles on top of input video using the libass library.
19199 To enable compilation of this filter you need to configure FFmpeg with
19200 @code{--enable-libass}. This filter also requires a build with libavcodec and
19201 libavformat to convert the passed subtitles file to ASS (Advanced Substation
19202 Alpha) subtitles format.
19204 The filter accepts the following options:
19208 Set the filename of the subtitle file to read. It must be specified.
19210 @item original_size
19211 Specify the size of the original video, the video for which the ASS file
19212 was composed. For the syntax of this option, check the
19213 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19214 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
19215 correctly scale the fonts if the aspect ratio has been changed.
19218 Set a directory path containing fonts that can be used by the filter.
19219 These fonts will be used in addition to whatever the font provider uses.
19222 Process alpha channel, by default alpha channel is untouched.
19225 Set subtitles input character encoding. @code{subtitles} filter only. Only
19226 useful if not UTF-8.
19228 @item stream_index, si
19229 Set subtitles stream index. @code{subtitles} filter only.
19232 Override default style or script info parameters of the subtitles. It accepts a
19233 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
19236 If the first key is not specified, it is assumed that the first value
19237 specifies the @option{filename}.
19239 For example, to render the file @file{sub.srt} on top of the input
19240 video, use the command:
19245 which is equivalent to:
19247 subtitles=filename=sub.srt
19250 To render the default subtitles stream from file @file{video.mkv}, use:
19252 subtitles=video.mkv
19255 To render the second subtitles stream from that file, use:
19257 subtitles=video.mkv:si=1
19260 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
19261 @code{DejaVu Serif}, use:
19263 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
19266 @section super2xsai
19268 Scale the input by 2x and smooth using the Super2xSaI (Scale and
19269 Interpolate) pixel art scaling algorithm.
19271 Useful for enlarging pixel art images without reducing sharpness.
19275 Swap two rectangular objects in video.
19277 This filter accepts the following options:
19287 Set 1st rect x coordinate.
19290 Set 1st rect y coordinate.
19293 Set 2nd rect x coordinate.
19296 Set 2nd rect y coordinate.
19298 All expressions are evaluated once for each frame.
19301 The all options are expressions containing the following constants:
19306 The input width and height.
19309 same as @var{w} / @var{h}
19312 input sample aspect ratio
19315 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
19318 The number of the input frame, starting from 0.
19321 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
19324 the position in the file of the input frame, NAN if unknown
19327 @subsection Commands
19329 This filter supports the all above options as @ref{commands}.
19335 Blend successive video frames.
19341 Apply telecine process to the video.
19343 This filter accepts the following options:
19352 The default value is @code{top}.
19356 A string of numbers representing the pulldown pattern you wish to apply.
19357 The default value is @code{23}.
19361 Some typical patterns:
19366 24p: 2332 (preferred)
19373 24p: 222222222223 ("Euro pulldown")
19378 @section thistogram
19380 Compute and draw a color distribution histogram for the input video across time.
19382 Unlike @ref{histogram} video filter which only shows histogram of single input frame
19383 at certain time, this filter shows also past histograms of number of frames defined
19384 by @code{width} option.
19386 The computed histogram is a representation of the color component
19387 distribution in an image.
19389 The filter accepts the following options:
19393 Set width of single color component output. Default value is @code{0}.
19394 Value of @code{0} means width will be picked from input video.
19395 This also set number of passed histograms to keep.
19396 Allowed range is [0, 8192].
19398 @item display_mode, d
19400 It accepts the following values:
19403 Per color component graphs are placed below each other.
19406 Per color component graphs are placed side by side.
19409 Presents information identical to that in the @code{parade}, except
19410 that the graphs representing color components are superimposed directly
19413 Default is @code{stack}.
19415 @item levels_mode, m
19416 Set mode. Can be either @code{linear}, or @code{logarithmic}.
19417 Default is @code{linear}.
19419 @item components, c
19420 Set what color components to display.
19421 Default is @code{7}.
19424 Set background opacity. Default is @code{0.9}.
19427 Show envelope. Default is disabled.
19430 Set envelope color. Default is @code{gold}.
19435 Available values for slide is:
19438 Draw new frame when right border is reached.
19441 Replace old columns with new ones.
19444 Scroll from right to left.
19447 Scroll from left to right.
19450 Draw single picture.
19453 Default is @code{replace}.
19458 Apply threshold effect to video stream.
19460 This filter needs four video streams to perform thresholding.
19461 First stream is stream we are filtering.
19462 Second stream is holding threshold values, third stream is holding min values,
19463 and last, fourth stream is holding max values.
19465 The filter accepts the following option:
19469 Set which planes will be processed, unprocessed planes will be copied.
19470 By default value 0xf, all planes will be processed.
19473 For example if first stream pixel's component value is less then threshold value
19474 of pixel component from 2nd threshold stream, third stream value will picked,
19475 otherwise fourth stream pixel component value will be picked.
19477 Using color source filter one can perform various types of thresholding:
19479 @subsection Examples
19483 Binary threshold, using gray color as threshold:
19485 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
19489 Inverted binary threshold, using gray color as threshold:
19491 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
19495 Truncate binary threshold, using gray color as threshold:
19497 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
19501 Threshold to zero, using gray color as threshold:
19503 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
19507 Inverted threshold to zero, using gray color as threshold:
19509 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
19514 Select the most representative frame in a given sequence of consecutive frames.
19516 The filter accepts the following options:
19520 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
19521 will pick one of them, and then handle the next batch of @var{n} frames until
19522 the end. Default is @code{100}.
19525 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
19526 value will result in a higher memory usage, so a high value is not recommended.
19528 @subsection Examples
19532 Extract one picture each 50 frames:
19538 Complete example of a thumbnail creation with @command{ffmpeg}:
19540 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
19547 Tile several successive frames together.
19549 The @ref{untile} filter can do the reverse.
19551 The filter accepts the following options:
19556 Set the grid size (i.e. the number of lines and columns). For the syntax of
19557 this option, check the
19558 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19561 Set the maximum number of frames to render in the given area. It must be less
19562 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19563 the area will be used.
19566 Set the outer border margin in pixels.
19569 Set the inner border thickness (i.e. the number of pixels between frames). For
19570 more advanced padding options (such as having different values for the edges),
19571 refer to the pad video filter.
19574 Specify the color of the unused area. For the syntax of this option, check the
19575 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19576 The default value of @var{color} is "black".
19579 Set the number of frames to overlap when tiling several successive frames together.
19580 The value must be between @code{0} and @var{nb_frames - 1}.
19583 Set the number of frames to initially be empty before displaying first output frame.
19584 This controls how soon will one get first output frame.
19585 The value must be between @code{0} and @var{nb_frames - 1}.
19588 @subsection Examples
19592 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19594 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19596 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19597 duplicating each output frame to accommodate the originally detected frame
19601 Display @code{5} pictures in an area of @code{3x2} frames,
19602 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19603 mixed flat and named options:
19605 tile=3x2:nb_frames=5:padding=7:margin=2
19609 @section tinterlace
19611 Perform various types of temporal field interlacing.
19613 Frames are counted starting from 1, so the first input frame is
19616 The filter accepts the following options:
19621 Specify the mode of the interlacing. This option can also be specified
19622 as a value alone. See below for a list of values for this option.
19624 Available values are:
19628 Move odd frames into the upper field, even into the lower field,
19629 generating a double height frame at half frame rate.
19633 Frame 1 Frame 2 Frame 3 Frame 4
19635 11111 22222 33333 44444
19636 11111 22222 33333 44444
19637 11111 22222 33333 44444
19638 11111 22222 33333 44444
19652 Only output odd frames, even frames are dropped, generating a frame with
19653 unchanged height at half frame rate.
19658 Frame 1 Frame 2 Frame 3 Frame 4
19660 11111 22222 33333 44444
19661 11111 22222 33333 44444
19662 11111 22222 33333 44444
19663 11111 22222 33333 44444
19673 Only output even frames, odd frames are dropped, generating a frame with
19674 unchanged height at half frame rate.
19679 Frame 1 Frame 2 Frame 3 Frame 4
19681 11111 22222 33333 44444
19682 11111 22222 33333 44444
19683 11111 22222 33333 44444
19684 11111 22222 33333 44444
19694 Expand each frame to full height, but pad alternate lines with black,
19695 generating a frame with double height at the same input frame rate.
19700 Frame 1 Frame 2 Frame 3 Frame 4
19702 11111 22222 33333 44444
19703 11111 22222 33333 44444
19704 11111 22222 33333 44444
19705 11111 22222 33333 44444
19708 11111 ..... 33333 .....
19709 ..... 22222 ..... 44444
19710 11111 ..... 33333 .....
19711 ..... 22222 ..... 44444
19712 11111 ..... 33333 .....
19713 ..... 22222 ..... 44444
19714 11111 ..... 33333 .....
19715 ..... 22222 ..... 44444
19719 @item interleave_top, 4
19720 Interleave the upper field from odd frames with the lower field from
19721 even frames, generating a frame with unchanged height at half frame rate.
19726 Frame 1 Frame 2 Frame 3 Frame 4
19728 11111<- 22222 33333<- 44444
19729 11111 22222<- 33333 44444<-
19730 11111<- 22222 33333<- 44444
19731 11111 22222<- 33333 44444<-
19741 @item interleave_bottom, 5
19742 Interleave the lower field from odd frames with the upper field from
19743 even frames, generating a frame with unchanged height at half frame rate.
19748 Frame 1 Frame 2 Frame 3 Frame 4
19750 11111 22222<- 33333 44444<-
19751 11111<- 22222 33333<- 44444
19752 11111 22222<- 33333 44444<-
19753 11111<- 22222 33333<- 44444
19763 @item interlacex2, 6
19764 Double frame rate with unchanged height. Frames are inserted each
19765 containing the second temporal field from the previous input frame and
19766 the first temporal field from the next input frame. This mode relies on
19767 the top_field_first flag. Useful for interlaced video displays with no
19768 field synchronisation.
19773 Frame 1 Frame 2 Frame 3 Frame 4
19775 11111 22222 33333 44444
19776 11111 22222 33333 44444
19777 11111 22222 33333 44444
19778 11111 22222 33333 44444
19781 11111 22222 22222 33333 33333 44444 44444
19782 11111 11111 22222 22222 33333 33333 44444
19783 11111 22222 22222 33333 33333 44444 44444
19784 11111 11111 22222 22222 33333 33333 44444
19789 Move odd frames into the upper field, even into the lower field,
19790 generating a double height frame at same frame rate.
19795 Frame 1 Frame 2 Frame 3 Frame 4
19797 11111 22222 33333 44444
19798 11111 22222 33333 44444
19799 11111 22222 33333 44444
19800 11111 22222 33333 44444
19803 11111 33333 33333 55555
19804 22222 22222 44444 44444
19805 11111 33333 33333 55555
19806 22222 22222 44444 44444
19807 11111 33333 33333 55555
19808 22222 22222 44444 44444
19809 11111 33333 33333 55555
19810 22222 22222 44444 44444
19815 Numeric values are deprecated but are accepted for backward
19816 compatibility reasons.
19818 Default mode is @code{merge}.
19821 Specify flags influencing the filter process.
19823 Available value for @var{flags} is:
19826 @item low_pass_filter, vlpf
19827 Enable linear vertical low-pass filtering in the filter.
19828 Vertical low-pass filtering is required when creating an interlaced
19829 destination from a progressive source which contains high-frequency
19830 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19833 @item complex_filter, cvlpf
19834 Enable complex vertical low-pass filtering.
19835 This will slightly less reduce interlace 'twitter' and Moire
19836 patterning but better retain detail and subjective sharpness impression.
19839 Bypass already interlaced frames, only adjust the frame rate.
19842 Vertical low-pass filtering and bypassing already interlaced frames can only be
19843 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19848 Pick median pixels from several successive input video frames.
19850 The filter accepts the following options:
19854 Set radius of median filter.
19855 Default is 1. Allowed range is from 1 to 127.
19858 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19861 Set median percentile. Default value is @code{0.5}.
19862 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19863 minimum values, and @code{1} maximum values.
19866 @subsection Commands
19868 This filter supports all above options as @ref{commands}, excluding option @code{radius}.
19870 @section tmidequalizer
19872 Apply Temporal Midway Video Equalization effect.
19874 Midway Video Equalization adjusts a sequence of video frames to have the same
19875 histograms, while maintaining their dynamics as much as possible. It's
19876 useful for e.g. matching exposures from a video frames sequence.
19878 This filter accepts the following option:
19882 Set filtering radius. Default is @code{5}. Allowed range is from 1 to 127.
19885 Set filtering sigma. Default is @code{0.5}. This controls strength of filtering.
19886 Setting this option to 0 effectively does nothing.
19889 Set which planes to process. Default is @code{15}, which is all available planes.
19894 Mix successive video frames.
19896 A description of the accepted options follows.
19900 The number of successive frames to mix. If unspecified, it defaults to 3.
19903 Specify weight of each input video frame.
19904 Each weight is separated by space. If number of weights is smaller than
19905 number of @var{frames} last specified weight will be used for all remaining
19909 Specify scale, if it is set it will be multiplied with sum
19910 of each weight multiplied with pixel values to give final destination
19911 pixel value. By default @var{scale} is auto scaled to sum of weights.
19914 @subsection Examples
19918 Average 7 successive frames:
19920 tmix=frames=7:weights="1 1 1 1 1 1 1"
19924 Apply simple temporal convolution:
19926 tmix=frames=3:weights="-1 3 -1"
19930 Similar as above but only showing temporal differences:
19932 tmix=frames=3:weights="-1 2 -1":scale=1
19936 @subsection Commands
19938 This filter supports the following commands:
19942 Syntax is same as option with same name.
19947 Tone map colors from different dynamic ranges.
19949 This filter expects data in single precision floating point, as it needs to
19950 operate on (and can output) out-of-range values. Another filter, such as
19951 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19953 The tonemapping algorithms implemented only work on linear light, so input
19954 data should be linearized beforehand (and possibly correctly tagged).
19957 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19960 @subsection Options
19961 The filter accepts the following options.
19965 Set the tone map algorithm to use.
19967 Possible values are:
19970 Do not apply any tone map, only desaturate overbright pixels.
19973 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19974 in-range values, while distorting out-of-range values.
19977 Stretch the entire reference gamut to a linear multiple of the display.
19980 Fit a logarithmic transfer between the tone curves.
19983 Preserve overall image brightness with a simple curve, using nonlinear
19984 contrast, which results in flattening details and degrading color accuracy.
19987 Preserve both dark and bright details better than @var{reinhard}, at the cost
19988 of slightly darkening everything. Use it when detail preservation is more
19989 important than color and brightness accuracy.
19992 Smoothly map out-of-range values, while retaining contrast and colors for
19993 in-range material as much as possible. Use it when color accuracy is more
19994 important than detail preservation.
20000 Tune the tone mapping algorithm.
20002 This affects the following algorithms:
20008 Specifies the scale factor to use while stretching.
20012 Specifies the exponent of the function.
20016 Specify an extra linear coefficient to multiply into the signal before clipping.
20020 Specify the local contrast coefficient at the display peak.
20021 Default to 0.5, which means that in-gamut values will be about half as bright
20028 Specify the transition point from linear to mobius transform. Every value
20029 below this point is guaranteed to be mapped 1:1. The higher the value, the
20030 more accurate the result will be, at the cost of losing bright details.
20031 Default to 0.3, which due to the steep initial slope still preserves in-range
20032 colors fairly accurately.
20036 Apply desaturation for highlights that exceed this level of brightness. The
20037 higher the parameter, the more color information will be preserved. This
20038 setting helps prevent unnaturally blown-out colors for super-highlights, by
20039 (smoothly) turning into white instead. This makes images feel more natural,
20040 at the cost of reducing information about out-of-range colors.
20042 The default of 2.0 is somewhat conservative and will mostly just apply to
20043 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
20045 This option works only if the input frame has a supported color tag.
20048 Override signal/nominal/reference peak with this value. Useful when the
20049 embedded peak information in display metadata is not reliable or when tone
20050 mapping from a lower range to a higher range.
20055 Temporarily pad video frames.
20057 The filter accepts the following options:
20061 Specify number of delay frames before input video stream. Default is 0.
20064 Specify number of padding frames after input video stream.
20065 Set to -1 to pad indefinitely. Default is 0.
20068 Set kind of frames added to beginning of stream.
20069 Can be either @var{add} or @var{clone}.
20070 With @var{add} frames of solid-color are added.
20071 With @var{clone} frames are clones of first frame.
20072 Default is @var{add}.
20075 Set kind of frames added to end of stream.
20076 Can be either @var{add} or @var{clone}.
20077 With @var{add} frames of solid-color are added.
20078 With @var{clone} frames are clones of last frame.
20079 Default is @var{add}.
20081 @item start_duration, stop_duration
20082 Specify the duration of the start/stop delay. See
20083 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20084 for the accepted syntax.
20085 These options override @var{start} and @var{stop}. Default is 0.
20088 Specify the color of the padded area. For the syntax of this option,
20089 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
20090 manual,ffmpeg-utils}.
20092 The default value of @var{color} is "black".
20098 Transpose rows with columns in the input video and optionally flip it.
20100 It accepts the following parameters:
20105 Specify the transposition direction.
20107 Can assume the following values:
20109 @item 0, 4, cclock_flip
20110 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
20118 Rotate by 90 degrees clockwise, that is:
20126 Rotate by 90 degrees counterclockwise, that is:
20133 @item 3, 7, clock_flip
20134 Rotate by 90 degrees clockwise and vertically flip, that is:
20142 For values between 4-7, the transposition is only done if the input
20143 video geometry is portrait and not landscape. These values are
20144 deprecated, the @code{passthrough} option should be used instead.
20146 Numerical values are deprecated, and should be dropped in favor of
20147 symbolic constants.
20150 Do not apply the transposition if the input geometry matches the one
20151 specified by the specified value. It accepts the following values:
20154 Always apply transposition.
20156 Preserve portrait geometry (when @var{height} >= @var{width}).
20158 Preserve landscape geometry (when @var{width} >= @var{height}).
20161 Default value is @code{none}.
20164 For example to rotate by 90 degrees clockwise and preserve portrait
20167 transpose=dir=1:passthrough=portrait
20170 The command above can also be specified as:
20172 transpose=1:portrait
20175 @section transpose_npp
20177 Transpose rows with columns in the input video and optionally flip it.
20178 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
20180 It accepts the following parameters:
20185 Specify the transposition direction.
20187 Can assume the following values:
20190 Rotate by 90 degrees counterclockwise and vertically flip. (default)
20193 Rotate by 90 degrees clockwise.
20196 Rotate by 90 degrees counterclockwise.
20199 Rotate by 90 degrees clockwise and vertically flip.
20203 Do not apply the transposition if the input geometry matches the one
20204 specified by the specified value. It accepts the following values:
20207 Always apply transposition. (default)
20209 Preserve portrait geometry (when @var{height} >= @var{width}).
20211 Preserve landscape geometry (when @var{width} >= @var{height}).
20217 Trim the input so that the output contains one continuous subpart of the input.
20219 It accepts the following parameters:
20222 Specify the time of the start of the kept section, i.e. the frame with the
20223 timestamp @var{start} will be the first frame in the output.
20226 Specify the time of the first frame that will be dropped, i.e. the frame
20227 immediately preceding the one with the timestamp @var{end} will be the last
20228 frame in the output.
20231 This is the same as @var{start}, except this option sets the start timestamp
20232 in timebase units instead of seconds.
20235 This is the same as @var{end}, except this option sets the end timestamp
20236 in timebase units instead of seconds.
20239 The maximum duration of the output in seconds.
20242 The number of the first frame that should be passed to the output.
20245 The number of the first frame that should be dropped.
20248 @option{start}, @option{end}, and @option{duration} are expressed as time
20249 duration specifications; see
20250 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20251 for the accepted syntax.
20253 Note that the first two sets of the start/end options and the @option{duration}
20254 option look at the frame timestamp, while the _frame variants simply count the
20255 frames that pass through the filter. Also note that this filter does not modify
20256 the timestamps. If you wish for the output timestamps to start at zero, insert a
20257 setpts filter after the trim filter.
20259 If multiple start or end options are set, this filter tries to be greedy and
20260 keep all the frames that match at least one of the specified constraints. To keep
20261 only the part that matches all the constraints at once, chain multiple trim
20264 The defaults are such that all the input is kept. So it is possible to set e.g.
20265 just the end values to keep everything before the specified time.
20270 Drop everything except the second minute of input:
20272 ffmpeg -i INPUT -vf trim=60:120
20276 Keep only the first second:
20278 ffmpeg -i INPUT -vf trim=duration=1
20283 @section unpremultiply
20284 Apply alpha unpremultiply effect to input video stream using first plane
20285 of second stream as alpha.
20287 Both streams must have same dimensions and same pixel format.
20289 The filter accepts the following option:
20293 Set which planes will be processed, unprocessed planes will be copied.
20294 By default value 0xf, all planes will be processed.
20296 If the format has 1 or 2 components, then luma is bit 0.
20297 If the format has 3 or 4 components:
20298 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
20299 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
20300 If present, the alpha channel is always the last bit.
20303 Do not require 2nd input for processing, instead use alpha plane from input stream.
20309 Sharpen or blur the input video.
20311 It accepts the following parameters:
20314 @item luma_msize_x, lx
20315 Set the luma matrix horizontal size. It must be an odd integer between
20316 3 and 23. The default value is 5.
20318 @item luma_msize_y, ly
20319 Set the luma matrix vertical size. It must be an odd integer between 3
20320 and 23. The default value is 5.
20322 @item luma_amount, la
20323 Set the luma effect strength. It must be a floating point number, reasonable
20324 values lay between -1.5 and 1.5.
20326 Negative values will blur the input video, while positive values will
20327 sharpen it, a value of zero will disable the effect.
20329 Default value is 1.0.
20331 @item chroma_msize_x, cx
20332 Set the chroma matrix horizontal size. It must be an odd integer
20333 between 3 and 23. The default value is 5.
20335 @item chroma_msize_y, cy
20336 Set the chroma matrix vertical size. It must be an odd integer
20337 between 3 and 23. The default value is 5.
20339 @item chroma_amount, ca
20340 Set the chroma effect strength. It must be a floating point number, reasonable
20341 values lay between -1.5 and 1.5.
20343 Negative values will blur the input video, while positive values will
20344 sharpen it, a value of zero will disable the effect.
20346 Default value is 0.0.
20350 All parameters are optional and default to the equivalent of the
20351 string '5:5:1.0:5:5:0.0'.
20353 @subsection Examples
20357 Apply strong luma sharpen effect:
20359 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
20363 Apply a strong blur of both luma and chroma parameters:
20365 unsharp=7:7:-2:7:7:-2
20372 Decompose a video made of tiled images into the individual images.
20374 The frame rate of the output video is the frame rate of the input video
20375 multiplied by the number of tiles.
20377 This filter does the reverse of @ref{tile}.
20379 The filter accepts the following options:
20384 Set the grid size (i.e. the number of lines and columns). For the syntax of
20385 this option, check the
20386 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20389 @subsection Examples
20393 Produce a 1-second video from a still image file made of 25 frames stacked
20394 vertically, like an analogic film reel:
20396 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
20402 Apply ultra slow/simple postprocessing filter that compresses and decompresses
20403 the image at several (or - in the case of @option{quality} level @code{8} - all)
20404 shifts and average the results.
20406 The way this differs from the behavior of spp is that uspp actually encodes &
20407 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
20408 DCT similar to MJPEG.
20410 The filter accepts the following options:
20414 Set quality. This option defines the number of levels for averaging. It accepts
20415 an integer in the range 0-8. If set to @code{0}, the filter will have no
20416 effect. A value of @code{8} means the higher quality. For each increment of
20417 that value the speed drops by a factor of approximately 2. Default value is
20421 Force a constant quantization parameter. If not set, the filter will use the QP
20422 from the video stream (if available).
20427 Convert 360 videos between various formats.
20429 The filter accepts the following options:
20435 Set format of the input/output video.
20443 Equirectangular projection.
20448 Cubemap with 3x2/6x1/1x6 layout.
20450 Format specific options:
20455 Set padding proportion for the input/output cubemap. Values in decimals.
20462 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)
20465 Default value is @b{@samp{0}}.
20466 Maximum value is @b{@samp{0.1}}.
20470 Set fixed padding for the input/output cubemap. Values in pixels.
20472 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
20476 Set order of faces for the input/output cubemap. Choose one direction for each position.
20478 Designation of directions:
20494 Default value is @b{@samp{rludfb}}.
20498 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
20500 Designation of angles:
20503 0 degrees clockwise
20505 90 degrees clockwise
20507 180 degrees clockwise
20509 270 degrees clockwise
20512 Default value is @b{@samp{000000}}.
20516 Equi-Angular Cubemap.
20523 Format specific options:
20528 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20530 If diagonal field of view is set it overrides horizontal and vertical field of view.
20535 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20537 If diagonal field of view is set it overrides horizontal and vertical field of view.
20543 Format specific options:
20548 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20550 If diagonal field of view is set it overrides horizontal and vertical field of view.
20555 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20557 If diagonal field of view is set it overrides horizontal and vertical field of view.
20563 Facebook's 360 formats.
20566 Stereographic format.
20568 Format specific options:
20573 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20575 If diagonal field of view is set it overrides horizontal and vertical field of view.
20580 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20582 If diagonal field of view is set it overrides horizontal and vertical field of view.
20589 Ball format, gives significant distortion toward the back.
20592 Hammer-Aitoff map projection format.
20595 Sinusoidal map projection format.
20598 Fisheye projection.
20600 Format specific options:
20605 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20607 If diagonal field of view is set it overrides horizontal and vertical field of view.
20612 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20614 If diagonal field of view is set it overrides horizontal and vertical field of view.
20618 Pannini projection.
20620 Format specific options:
20623 Set output pannini parameter.
20626 Set input pannini parameter.
20630 Cylindrical projection.
20632 Format specific options:
20637 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20639 If diagonal field of view is set it overrides horizontal and vertical field of view.
20644 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20646 If diagonal field of view is set it overrides horizontal and vertical field of view.
20650 Perspective projection. @i{(output only)}
20652 Format specific options:
20655 Set perspective parameter.
20659 Tetrahedron projection.
20662 Truncated square pyramid projection.
20666 Half equirectangular projection.
20671 Format specific options:
20676 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20678 If diagonal field of view is set it overrides horizontal and vertical field of view.
20683 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20685 If diagonal field of view is set it overrides horizontal and vertical field of view.
20689 Orthographic format.
20691 Format specific options:
20696 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20698 If diagonal field of view is set it overrides horizontal and vertical field of view.
20703 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20705 If diagonal field of view is set it overrides horizontal and vertical field of view.
20709 Octahedron projection.
20713 Set interpolation method.@*
20714 @i{Note: more complex interpolation methods require much more memory to run.}
20724 Bilinear interpolation.
20726 Lagrange9 interpolation.
20729 Bicubic interpolation.
20732 Lanczos interpolation.
20735 Spline16 interpolation.
20738 Gaussian interpolation.
20740 Mitchell interpolation.
20743 Default value is @b{@samp{line}}.
20747 Set the output video resolution.
20749 Default resolution depends on formats.
20753 Set the input/output stereo format.
20764 Default value is @b{@samp{2d}} for input and output format.
20769 Set rotation for the output video. Values in degrees.
20772 Set rotation order for the output video. Choose one item for each position.
20783 Default value is @b{@samp{ypr}}.
20788 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20792 Set if input video is flipped horizontally/vertically. Boolean values.
20795 Set if input video is transposed. Boolean value, by default disabled.
20798 Set if output video needs to be transposed. Boolean value, by default disabled.
20801 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20804 @subsection Examples
20808 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20810 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20813 Extract back view of Equi-Angular Cubemap:
20815 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20818 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20820 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20824 @subsection Commands
20826 This filter supports subset of above options as @ref{commands}.
20828 @section vaguedenoiser
20830 Apply a wavelet based denoiser.
20832 It transforms each frame from the video input into the wavelet domain,
20833 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20834 the obtained coefficients. It does an inverse wavelet transform after.
20835 Due to wavelet properties, it should give a nice smoothed result, and
20836 reduced noise, without blurring picture features.
20838 This filter accepts the following options:
20842 The filtering strength. The higher, the more filtered the video will be.
20843 Hard thresholding can use a higher threshold than soft thresholding
20844 before the video looks overfiltered. Default value is 2.
20847 The filtering method the filter will use.
20849 It accepts the following values:
20852 All values under the threshold will be zeroed.
20855 All values under the threshold will be zeroed. All values above will be
20856 reduced by the threshold.
20859 Scales or nullifies coefficients - intermediary between (more) soft and
20860 (less) hard thresholding.
20863 Default is garrote.
20866 Number of times, the wavelet will decompose the picture. Picture can't
20867 be decomposed beyond a particular point (typically, 8 for a 640x480
20868 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20871 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20874 A list of the planes to process. By default all planes are processed.
20877 The threshold type the filter will use.
20879 It accepts the following values:
20882 Threshold used is same for all decompositions.
20885 Threshold used depends also on each decomposition coefficients.
20888 Default is universal.
20891 @section vectorscope
20893 Display 2 color component values in the two dimensional graph (which is called
20896 This filter accepts the following options:
20900 Set vectorscope mode.
20902 It accepts the following values:
20906 Gray values are displayed on graph, higher brightness means more pixels have
20907 same component color value on location in graph. This is the default mode.
20910 Gray values are displayed on graph. Surrounding pixels values which are not
20911 present in video frame are drawn in gradient of 2 color components which are
20912 set by option @code{x} and @code{y}. The 3rd color component is static.
20915 Actual color components values present in video frame are displayed on graph.
20918 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20919 on graph increases value of another color component, which is luminance by
20920 default values of @code{x} and @code{y}.
20923 Actual colors present in video frame are displayed on graph. If two different
20924 colors map to same position on graph then color with higher value of component
20925 not present in graph is picked.
20928 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20929 component picked from radial gradient.
20933 Set which color component will be represented on X-axis. Default is @code{1}.
20936 Set which color component will be represented on Y-axis. Default is @code{2}.
20939 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20940 of color component which represents frequency of (X, Y) location in graph.
20945 No envelope, this is default.
20948 Instant envelope, even darkest single pixel will be clearly highlighted.
20951 Hold maximum and minimum values presented in graph over time. This way you
20952 can still spot out of range values without constantly looking at vectorscope.
20955 Peak and instant envelope combined together.
20959 Set what kind of graticule to draw.
20968 Set graticule opacity.
20971 Set graticule flags.
20975 Draw graticule for white point.
20978 Draw graticule for black point.
20981 Draw color points short names.
20985 Set background opacity.
20987 @item lthreshold, l
20988 Set low threshold for color component not represented on X or Y axis.
20989 Values lower than this value will be ignored. Default is 0.
20990 Note this value is multiplied with actual max possible value one pixel component
20991 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20994 @item hthreshold, h
20995 Set high threshold for color component not represented on X or Y axis.
20996 Values higher than this value will be ignored. Default is 1.
20997 Note this value is multiplied with actual max possible value one pixel component
20998 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20999 is 0.9 * 255 = 230.
21001 @item colorspace, c
21002 Set what kind of colorspace to use when drawing graticule.
21012 Set color tint for gray/tint vectorscope mode. By default both options are zero.
21013 This means no tint, and output will remain gray.
21016 @anchor{vidstabdetect}
21017 @section vidstabdetect
21019 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
21020 @ref{vidstabtransform} for pass 2.
21022 This filter generates a file with relative translation and rotation
21023 transform information about subsequent frames, which is then used by
21024 the @ref{vidstabtransform} filter.
21026 To enable compilation of this filter you need to configure FFmpeg with
21027 @code{--enable-libvidstab}.
21029 This filter accepts the following options:
21033 Set the path to the file used to write the transforms information.
21034 Default value is @file{transforms.trf}.
21037 Set how shaky the video is and how quick the camera is. It accepts an
21038 integer in the range 1-10, a value of 1 means little shakiness, a
21039 value of 10 means strong shakiness. Default value is 5.
21042 Set the accuracy of the detection process. It must be a value in the
21043 range 1-15. A value of 1 means low accuracy, a value of 15 means high
21044 accuracy. Default value is 15.
21047 Set stepsize of the search process. The region around minimum is
21048 scanned with 1 pixel resolution. Default value is 6.
21051 Set minimum contrast. Below this value a local measurement field is
21052 discarded. Must be a floating point value in the range 0-1. Default
21056 Set reference frame number for tripod mode.
21058 If enabled, the motion of the frames is compared to a reference frame
21059 in the filtered stream, identified by the specified number. The idea
21060 is to compensate all movements in a more-or-less static scene and keep
21061 the camera view absolutely still.
21063 If set to 0, it is disabled. The frames are counted starting from 1.
21066 Show fields and transforms in the resulting frames. It accepts an
21067 integer in the range 0-2. Default value is 0, which disables any
21071 @subsection Examples
21075 Use default values:
21081 Analyze strongly shaky movie and put the results in file
21082 @file{mytransforms.trf}:
21084 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
21088 Visualize the result of internal transformations in the resulting
21091 vidstabdetect=show=1
21095 Analyze a video with medium shakiness using @command{ffmpeg}:
21097 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
21101 @anchor{vidstabtransform}
21102 @section vidstabtransform
21104 Video stabilization/deshaking: pass 2 of 2,
21105 see @ref{vidstabdetect} for pass 1.
21107 Read a file with transform information for each frame and
21108 apply/compensate them. Together with the @ref{vidstabdetect}
21109 filter this can be used to deshake videos. See also
21110 @url{http://public.hronopik.de/vid.stab}. It is important to also use
21111 the @ref{unsharp} filter, see below.
21113 To enable compilation of this filter you need to configure FFmpeg with
21114 @code{--enable-libvidstab}.
21116 @subsection Options
21120 Set path to the file used to read the transforms. Default value is
21121 @file{transforms.trf}.
21124 Set the number of frames (value*2 + 1) used for lowpass filtering the
21125 camera movements. Default value is 10.
21127 For example a number of 10 means that 21 frames are used (10 in the
21128 past and 10 in the future) to smoothen the motion in the video. A
21129 larger value leads to a smoother video, but limits the acceleration of
21130 the camera (pan/tilt movements). 0 is a special case where a static
21131 camera is simulated.
21134 Set the camera path optimization algorithm.
21136 Accepted values are:
21139 gaussian kernel low-pass filter on camera motion (default)
21141 averaging on transformations
21145 Set maximal number of pixels to translate frames. Default value is -1,
21149 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
21150 value is -1, meaning no limit.
21153 Specify how to deal with borders that may be visible due to movement
21156 Available values are:
21159 keep image information from previous frame (default)
21161 fill the border black
21165 Invert transforms if set to 1. Default value is 0.
21168 Consider transforms as relative to previous frame if set to 1,
21169 absolute if set to 0. Default value is 0.
21172 Set percentage to zoom. A positive value will result in a zoom-in
21173 effect, a negative value in a zoom-out effect. Default value is 0 (no
21177 Set optimal zooming to avoid borders.
21179 Accepted values are:
21184 optimal static zoom value is determined (only very strong movements
21185 will lead to visible borders) (default)
21187 optimal adaptive zoom value is determined (no borders will be
21188 visible), see @option{zoomspeed}
21191 Note that the value given at zoom is added to the one calculated here.
21194 Set percent to zoom maximally each frame (enabled when
21195 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
21199 Specify type of interpolation.
21201 Available values are:
21206 linear only horizontal
21208 linear in both directions (default)
21210 cubic in both directions (slow)
21214 Enable virtual tripod mode if set to 1, which is equivalent to
21215 @code{relative=0:smoothing=0}. Default value is 0.
21217 Use also @code{tripod} option of @ref{vidstabdetect}.
21220 Increase log verbosity if set to 1. Also the detected global motions
21221 are written to the temporary file @file{global_motions.trf}. Default
21225 @subsection Examples
21229 Use @command{ffmpeg} for a typical stabilization with default values:
21231 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
21234 Note the use of the @ref{unsharp} filter which is always recommended.
21237 Zoom in a bit more and load transform data from a given file:
21239 vidstabtransform=zoom=5:input="mytransforms.trf"
21243 Smoothen the video even more:
21245 vidstabtransform=smoothing=30
21251 Flip the input video vertically.
21253 For example, to vertically flip a video with @command{ffmpeg}:
21255 ffmpeg -i in.avi -vf "vflip" out.avi
21260 Detect variable frame rate video.
21262 This filter tries to detect if the input is variable or constant frame rate.
21264 At end it will output number of frames detected as having variable delta pts,
21265 and ones with constant delta pts.
21266 If there was frames with variable delta, than it will also show min, max and
21267 average delta encountered.
21271 Boost or alter saturation.
21273 The filter accepts the following options:
21276 Set strength of boost if positive value or strength of alter if negative value.
21277 Default is 0. Allowed range is from -2 to 2.
21280 Set the red balance. Default is 1. Allowed range is from -10 to 10.
21283 Set the green balance. Default is 1. Allowed range is from -10 to 10.
21286 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
21289 Set the red luma coefficient.
21292 Set the green luma coefficient.
21295 Set the blue luma coefficient.
21298 If @code{intensity} is negative and this is set to 1, colors will change,
21299 otherwise colors will be less saturated, more towards gray.
21302 @subsection Commands
21304 This filter supports the all above options as @ref{commands}.
21308 Obtain the average VIF (Visual Information Fidelity) between two input videos.
21310 This filter takes two input videos.
21312 Both input videos must have the same resolution and pixel format for
21313 this filter to work correctly. Also it assumes that both inputs
21314 have the same number of frames, which are compared one by one.
21316 The obtained average VIF score is printed through the logging system.
21318 The filter stores the calculated VIF score of each frame.
21320 In the below example the input file @file{main.mpg} being processed is compared
21321 with the reference file @file{ref.mpg}.
21324 ffmpeg -i main.mpg -i ref.mpg -lavfi vif -f null -
21330 Make or reverse a natural vignetting effect.
21332 The filter accepts the following options:
21336 Set lens angle expression as a number of radians.
21338 The value is clipped in the @code{[0,PI/2]} range.
21340 Default value: @code{"PI/5"}
21344 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
21348 Set forward/backward mode.
21350 Available modes are:
21353 The larger the distance from the central point, the darker the image becomes.
21356 The larger the distance from the central point, the brighter the image becomes.
21357 This can be used to reverse a vignette effect, though there is no automatic
21358 detection to extract the lens @option{angle} and other settings (yet). It can
21359 also be used to create a burning effect.
21362 Default value is @samp{forward}.
21365 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
21367 It accepts the following values:
21370 Evaluate expressions only once during the filter initialization.
21373 Evaluate expressions for each incoming frame. This is way slower than the
21374 @samp{init} mode since it requires all the scalers to be re-computed, but it
21375 allows advanced dynamic expressions.
21378 Default value is @samp{init}.
21381 Set dithering to reduce the circular banding effects. Default is @code{1}
21385 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
21386 Setting this value to the SAR of the input will make a rectangular vignetting
21387 following the dimensions of the video.
21389 Default is @code{1/1}.
21392 @subsection Expressions
21394 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
21395 following parameters.
21400 input width and height
21403 the number of input frame, starting from 0
21406 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
21407 @var{TB} units, NAN if undefined
21410 frame rate of the input video, NAN if the input frame rate is unknown
21413 the PTS (Presentation TimeStamp) of the filtered video frame,
21414 expressed in seconds, NAN if undefined
21417 time base of the input video
21421 @subsection Examples
21425 Apply simple strong vignetting effect:
21431 Make a flickering vignetting:
21433 vignette='PI/4+random(1)*PI/50':eval=frame
21438 @section vmafmotion
21440 Obtain the average VMAF motion score of a video.
21441 It is one of the component metrics of VMAF.
21443 The obtained average motion score is printed through the logging system.
21445 The filter accepts the following options:
21449 If specified, the filter will use the named file to save the motion score of
21450 each frame with respect to the previous frame.
21451 When filename equals "-" the data is sent to standard output.
21456 ffmpeg -i ref.mpg -vf vmafmotion -f null -
21460 Stack input videos vertically.
21462 All streams must be of same pixel format and of same width.
21464 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
21465 to create same output.
21467 The filter accepts the following options:
21471 Set number of input streams. Default is 2.
21474 If set to 1, force the output to terminate when the shortest input
21475 terminates. Default value is 0.
21480 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
21481 Deinterlacing Filter").
21483 Based on the process described by Martin Weston for BBC R&D, and
21484 implemented based on the de-interlace algorithm written by Jim
21485 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
21486 uses filter coefficients calculated by BBC R&D.
21488 This filter uses field-dominance information in frame to decide which
21489 of each pair of fields to place first in the output.
21490 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
21492 There are two sets of filter coefficients, so called "simple"
21493 and "complex". Which set of filter coefficients is used can
21494 be set by passing an optional parameter:
21498 Set the interlacing filter coefficients. Accepts one of the following values:
21502 Simple filter coefficient set.
21504 More-complex filter coefficient set.
21506 Default value is @samp{complex}.
21509 The interlacing mode to adopt. It accepts one of the following values:
21513 Output one frame for each frame.
21515 Output one frame for each field.
21518 The default value is @code{field}.
21521 The picture field parity assumed for the input interlaced video. It accepts one
21522 of the following values:
21526 Assume the top field is first.
21528 Assume the bottom field is first.
21530 Enable automatic detection of field parity.
21533 The default value is @code{auto}.
21534 If the interlacing is unknown or the decoder does not export this information,
21535 top field first will be assumed.
21538 Specify which frames to deinterlace. Accepts one of the following values:
21542 Deinterlace all frames,
21544 Only deinterlace frames marked as interlaced.
21547 Default value is @samp{all}.
21550 @subsection Commands
21551 This filter supports same @ref{commands} as options.
21554 Video waveform monitor.
21556 The waveform monitor plots color component intensity. By default luminance
21557 only. Each column of the waveform corresponds to a column of pixels in the
21560 It accepts the following options:
21564 Can be either @code{row}, or @code{column}. Default is @code{column}.
21565 In row mode, the graph on the left side represents color component value 0 and
21566 the right side represents value = 255. In column mode, the top side represents
21567 color component value = 0 and bottom side represents value = 255.
21570 Set intensity. Smaller values are useful to find out how many values of the same
21571 luminance are distributed across input rows/columns.
21572 Default value is @code{0.04}. Allowed range is [0, 1].
21575 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
21576 In mirrored mode, higher values will be represented on the left
21577 side for @code{row} mode and at the top for @code{column} mode. Default is
21578 @code{1} (mirrored).
21582 It accepts the following values:
21585 Presents information identical to that in the @code{parade}, except
21586 that the graphs representing color components are superimposed directly
21589 This display mode makes it easier to spot relative differences or similarities
21590 in overlapping areas of the color components that are supposed to be identical,
21591 such as neutral whites, grays, or blacks.
21594 Display separate graph for the color components side by side in
21595 @code{row} mode or one below the other in @code{column} mode.
21598 Display separate graph for the color components side by side in
21599 @code{column} mode or one below the other in @code{row} mode.
21601 Using this display mode makes it easy to spot color casts in the highlights
21602 and shadows of an image, by comparing the contours of the top and the bottom
21603 graphs of each waveform. Since whites, grays, and blacks are characterized
21604 by exactly equal amounts of red, green, and blue, neutral areas of the picture
21605 should display three waveforms of roughly equal width/height. If not, the
21606 correction is easy to perform by making level adjustments the three waveforms.
21608 Default is @code{stack}.
21610 @item components, c
21611 Set which color components to display. Default is 1, which means only luminance
21612 or red color component if input is in RGB colorspace. If is set for example to
21613 7 it will display all 3 (if) available color components.
21618 No envelope, this is default.
21621 Instant envelope, minimum and maximum values presented in graph will be easily
21622 visible even with small @code{step} value.
21625 Hold minimum and maximum values presented in graph across time. This way you
21626 can still spot out of range values without constantly looking at waveforms.
21629 Peak and instant envelope combined together.
21635 No filtering, this is default.
21638 Luma and chroma combined together.
21641 Similar as above, but shows difference between blue and red chroma.
21644 Similar as above, but use different colors.
21647 Similar as above, but again with different colors.
21650 Displays only chroma.
21653 Displays actual color value on waveform.
21656 Similar as above, but with luma showing frequency of chroma values.
21660 Set which graticule to display.
21664 Do not display graticule.
21667 Display green graticule showing legal broadcast ranges.
21670 Display orange graticule showing legal broadcast ranges.
21673 Display invert graticule showing legal broadcast ranges.
21677 Set graticule opacity.
21680 Set graticule flags.
21684 Draw numbers above lines. By default enabled.
21687 Draw dots instead of lines.
21691 Set scale used for displaying graticule.
21698 Default is digital.
21701 Set background opacity.
21705 Set tint for output.
21706 Only used with lowpass filter and when display is not overlay and input
21707 pixel formats are not RGB.
21710 @section weave, doubleweave
21712 The @code{weave} takes a field-based video input and join
21713 each two sequential fields into single frame, producing a new double
21714 height clip with half the frame rate and half the frame count.
21716 The @code{doubleweave} works same as @code{weave} but without
21717 halving frame rate and frame count.
21719 It accepts the following option:
21723 Set first field. Available values are:
21727 Set the frame as top-field-first.
21730 Set the frame as bottom-field-first.
21734 @subsection Examples
21738 Interlace video using @ref{select} and @ref{separatefields} filter:
21740 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21745 Apply the xBR high-quality magnification filter which is designed for pixel
21746 art. It follows a set of edge-detection rules, see
21747 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21749 It accepts the following option:
21753 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21754 @code{3xBR} and @code{4} for @code{4xBR}.
21755 Default is @code{3}.
21760 Apply cross fade from one input video stream to another input video stream.
21761 The cross fade is applied for specified duration.
21763 The filter accepts the following options:
21767 Set one of available transition effects:
21815 Default transition effect is fade.
21818 Set cross fade duration in seconds.
21819 Default duration is 1 second.
21822 Set cross fade start relative to first input stream in seconds.
21823 Default offset is 0.
21826 Set expression for custom transition effect.
21828 The expressions can use the following variables and functions:
21833 The coordinates of the current sample.
21837 The width and height of the image.
21840 Progress of transition effect.
21843 Currently processed plane.
21846 Return value of first input at current location and plane.
21849 Return value of second input at current location and plane.
21855 Return the value of the pixel at location (@var{x},@var{y}) of the
21856 first/second/third/fourth component of first input.
21862 Return the value of the pixel at location (@var{x},@var{y}) of the
21863 first/second/third/fourth component of second input.
21867 @subsection Examples
21871 Cross fade from one input video to another input video, with fade transition and duration of transition
21872 of 2 seconds starting at offset of 5 seconds:
21874 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21879 Pick median pixels from several input videos.
21881 The filter accepts the following options:
21885 Set number of inputs.
21886 Default is 3. Allowed range is from 3 to 255.
21887 If number of inputs is even number, than result will be mean value between two median values.
21890 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21893 Set median percentile. Default value is @code{0.5}.
21894 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21895 minimum values, and @code{1} maximum values.
21898 @subsection Commands
21900 This filter supports all above options as @ref{commands}, excluding option @code{inputs}.
21903 Stack video inputs into custom layout.
21905 All streams must be of same pixel format.
21907 The filter accepts the following options:
21911 Set number of input streams. Default is 2.
21914 Specify layout of inputs.
21915 This option requires the desired layout configuration to be explicitly set by the user.
21916 This sets position of each video input in output. Each input
21917 is separated by '|'.
21918 The first number represents the column, and the second number represents the row.
21919 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21920 where X is video input from which to take width or height.
21921 Multiple values can be used when separated by '+'. In such
21922 case values are summed together.
21924 Note that if inputs are of different sizes gaps may appear, as not all of
21925 the output video frame will be filled. Similarly, videos can overlap each
21926 other if their position doesn't leave enough space for the full frame of
21929 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21930 a layout must be set by the user.
21933 If set to 1, force the output to terminate when the shortest input
21934 terminates. Default value is 0.
21937 If set to valid color, all unused pixels will be filled with that color.
21938 By default fill is set to none, so it is disabled.
21941 @subsection Examples
21945 Display 4 inputs into 2x2 grid.
21949 input1(0, 0) | input3(w0, 0)
21950 input2(0, h0) | input4(w0, h0)
21954 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21957 Note that if inputs are of different sizes, gaps or overlaps may occur.
21960 Display 4 inputs into 1x4 grid.
21967 input4(0, h0+h1+h2)
21971 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21974 Note that if inputs are of different widths, unused space will appear.
21977 Display 9 inputs into 3x3 grid.
21981 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21982 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21983 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21987 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
21990 Note that if inputs are of different sizes, gaps or overlaps may occur.
21993 Display 16 inputs into 4x4 grid.
21997 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21998 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21999 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
22000 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
22004 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|
22005 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
22008 Note that if inputs are of different sizes, gaps or overlaps may occur.
22015 Deinterlace the input video ("yadif" means "yet another deinterlacing
22018 It accepts the following parameters:
22024 The interlacing mode to adopt. It accepts one of the following values:
22027 @item 0, send_frame
22028 Output one frame for each frame.
22029 @item 1, send_field
22030 Output one frame for each field.
22031 @item 2, send_frame_nospatial
22032 Like @code{send_frame}, but it skips the spatial interlacing check.
22033 @item 3, send_field_nospatial
22034 Like @code{send_field}, but it skips the spatial interlacing check.
22037 The default value is @code{send_frame}.
22040 The picture field parity assumed for the input interlaced video. It accepts one
22041 of the following values:
22045 Assume the top field is first.
22047 Assume the bottom field is first.
22049 Enable automatic detection of field parity.
22052 The default value is @code{auto}.
22053 If the interlacing is unknown or the decoder does not export this information,
22054 top field first will be assumed.
22057 Specify which frames to deinterlace. Accepts one of the following
22062 Deinterlace all frames.
22063 @item 1, interlaced
22064 Only deinterlace frames marked as interlaced.
22067 The default value is @code{all}.
22070 @section yadif_cuda
22072 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
22073 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
22076 It accepts the following parameters:
22082 The interlacing mode to adopt. It accepts one of the following values:
22085 @item 0, send_frame
22086 Output one frame for each frame.
22087 @item 1, send_field
22088 Output one frame for each field.
22089 @item 2, send_frame_nospatial
22090 Like @code{send_frame}, but it skips the spatial interlacing check.
22091 @item 3, send_field_nospatial
22092 Like @code{send_field}, but it skips the spatial interlacing check.
22095 The default value is @code{send_frame}.
22098 The picture field parity assumed for the input interlaced video. It accepts one
22099 of the following values:
22103 Assume the top field is first.
22105 Assume the bottom field is first.
22107 Enable automatic detection of field parity.
22110 The default value is @code{auto}.
22111 If the interlacing is unknown or the decoder does not export this information,
22112 top field first will be assumed.
22115 Specify which frames to deinterlace. Accepts one of the following
22120 Deinterlace all frames.
22121 @item 1, interlaced
22122 Only deinterlace frames marked as interlaced.
22125 The default value is @code{all}.
22130 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
22131 The algorithm is described in
22132 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
22134 It accepts the following parameters:
22138 Set the window radius. Default value is 3.
22141 Set which planes to filter. Default is only the first plane.
22144 Set blur strength. Default value is 128.
22147 @subsection Commands
22148 This filter supports same @ref{commands} as options.
22152 Apply Zoom & Pan effect.
22154 This filter accepts the following options:
22158 Set the zoom expression. Range is 1-10. Default is 1.
22162 Set the x and y expression. Default is 0.
22165 Set the duration expression in number of frames.
22166 This sets for how many number of frames effect will last for
22167 single input image. Default is 90.
22170 Set the output image size, default is 'hd720'.
22173 Set the output frame rate, default is '25'.
22176 Each expression can contain the following constants:
22195 Output frame count.
22198 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
22200 @item out_time, time, ot
22201 The output timestamp expressed in seconds.
22205 Last calculated 'x' and 'y' position from 'x' and 'y' expression
22206 for current input frame.
22210 'x' and 'y' of last output frame of previous input frame or 0 when there was
22211 not yet such frame (first input frame).
22214 Last calculated zoom from 'z' expression for current input frame.
22217 Last calculated zoom of last output frame of previous input frame.
22220 Number of output frames for current input frame. Calculated from 'd' expression
22221 for each input frame.
22224 number of output frames created for previous input frame
22227 Rational number: input width / input height
22230 sample aspect ratio
22233 display aspect ratio
22237 @subsection Examples
22241 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
22243 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
22247 Zoom in up to 1.5x and pan always at center of picture:
22249 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
22253 Same as above but without pausing:
22255 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
22259 Zoom in 2x into center of picture only for the first second of the input video:
22261 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
22268 Scale (resize) the input video, using the z.lib library:
22269 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
22270 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
22272 The zscale filter forces the output display aspect ratio to be the same
22273 as the input, by changing the output sample aspect ratio.
22275 If the input image format is different from the format requested by
22276 the next filter, the zscale filter will convert the input to the
22279 @subsection Options
22280 The filter accepts the following options.
22285 Set the output video dimension expression. Default value is the input
22288 If the @var{width} or @var{w} value is 0, the input width is used for
22289 the output. If the @var{height} or @var{h} value is 0, the input height
22290 is used for the output.
22292 If one and only one of the values is -n with n >= 1, the zscale filter
22293 will use a value that maintains the aspect ratio of the input image,
22294 calculated from the other specified dimension. After that it will,
22295 however, make sure that the calculated dimension is divisible by n and
22296 adjust the value if necessary.
22298 If both values are -n with n >= 1, the behavior will be identical to
22299 both values being set to 0 as previously detailed.
22301 See below for the list of accepted constants for use in the dimension
22305 Set the video size. For the syntax of this option, check the
22306 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22309 Set the dither type.
22311 Possible values are:
22316 @item error_diffusion
22322 Set the resize filter type.
22324 Possible values are:
22334 Default is bilinear.
22337 Set the color range.
22339 Possible values are:
22346 Default is same as input.
22349 Set the color primaries.
22351 Possible values are:
22361 Default is same as input.
22364 Set the transfer characteristics.
22366 Possible values are:
22380 Default is same as input.
22383 Set the colorspace matrix.
22385 Possible value are:
22396 Default is same as input.
22399 Set the input color range.
22401 Possible values are:
22408 Default is same as input.
22410 @item primariesin, pin
22411 Set the input color primaries.
22413 Possible values are:
22423 Default is same as input.
22425 @item transferin, tin
22426 Set the input transfer characteristics.
22428 Possible values are:
22439 Default is same as input.
22441 @item matrixin, min
22442 Set the input colorspace matrix.
22444 Possible value are:
22456 Set the output chroma location.
22458 Possible values are:
22469 @item chromalin, cin
22470 Set the input chroma location.
22472 Possible values are:
22484 Set the nominal peak luminance.
22487 Parameter A for scaling filters. Parameter "b" for bicubic, and the number of
22488 filter taps for lanczos.
22491 Parameter B for scaling filters. Parameter "c" for bicubic.
22494 The values of the @option{w} and @option{h} options are expressions
22495 containing the following constants:
22500 The input width and height
22504 These are the same as @var{in_w} and @var{in_h}.
22508 The output (scaled) width and height
22512 These are the same as @var{out_w} and @var{out_h}
22515 The same as @var{iw} / @var{ih}
22518 input sample aspect ratio
22521 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
22525 horizontal and vertical input chroma subsample values. For example for the
22526 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22530 horizontal and vertical output chroma subsample values. For example for the
22531 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22534 @subsection Commands
22536 This filter supports the following commands:
22540 Set the output video dimension expression.
22541 The command accepts the same syntax of the corresponding option.
22543 If the specified expression is not valid, it is kept at its current
22547 @c man end VIDEO FILTERS
22549 @chapter OpenCL Video Filters
22550 @c man begin OPENCL VIDEO FILTERS
22552 Below is a description of the currently available OpenCL video filters.
22554 To enable compilation of these filters you need to configure FFmpeg with
22555 @code{--enable-opencl}.
22557 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
22560 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
22561 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
22562 given device parameters.
22564 @item -filter_hw_device @var{name}
22565 Pass the hardware device called @var{name} to all filters in any filter graph.
22569 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
22573 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
22575 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
22579 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.
22581 @section avgblur_opencl
22583 Apply average blur filter.
22585 The filter accepts the following options:
22589 Set horizontal radius size.
22590 Range is @code{[1, 1024]} and default value is @code{1}.
22593 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22596 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
22599 @subsection Example
22603 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.
22605 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
22609 @section boxblur_opencl
22611 Apply a boxblur algorithm to the input video.
22613 It accepts the following parameters:
22617 @item luma_radius, lr
22618 @item luma_power, lp
22619 @item chroma_radius, cr
22620 @item chroma_power, cp
22621 @item alpha_radius, ar
22622 @item alpha_power, ap
22626 A description of the accepted options follows.
22629 @item luma_radius, lr
22630 @item chroma_radius, cr
22631 @item alpha_radius, ar
22632 Set an expression for the box radius in pixels used for blurring the
22633 corresponding input plane.
22635 The radius value must be a non-negative number, and must not be
22636 greater than the value of the expression @code{min(w,h)/2} for the
22637 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
22640 Default value for @option{luma_radius} is "2". If not specified,
22641 @option{chroma_radius} and @option{alpha_radius} default to the
22642 corresponding value set for @option{luma_radius}.
22644 The expressions can contain the following constants:
22648 The input width and height in pixels.
22652 The input chroma image width and height in pixels.
22656 The horizontal and vertical chroma subsample values. For example, for the
22657 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22660 @item luma_power, lp
22661 @item chroma_power, cp
22662 @item alpha_power, ap
22663 Specify how many times the boxblur filter is applied to the
22664 corresponding plane.
22666 Default value for @option{luma_power} is 2. If not specified,
22667 @option{chroma_power} and @option{alpha_power} default to the
22668 corresponding value set for @option{luma_power}.
22670 A value of 0 will disable the effect.
22673 @subsection Examples
22675 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.
22679 Apply a boxblur filter with the luma, chroma, and alpha radius
22680 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.
22682 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22683 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22687 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.
22689 For the luma plane, a 2x2 box radius will be run once.
22691 For the chroma plane, a 4x4 box radius will be run 5 times.
22693 For the alpha plane, a 3x3 box radius will be run 7 times.
22695 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22699 @section colorkey_opencl
22700 RGB colorspace color keying.
22702 The filter accepts the following options:
22706 The color which will be replaced with transparency.
22709 Similarity percentage with the key color.
22711 0.01 matches only the exact key color, while 1.0 matches everything.
22716 0.0 makes pixels either fully transparent, or not transparent at all.
22718 Higher values result in semi-transparent pixels, with a higher transparency
22719 the more similar the pixels color is to the key color.
22722 @subsection Examples
22726 Make every semi-green pixel in the input transparent with some slight blending:
22728 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22732 @section convolution_opencl
22734 Apply convolution of 3x3, 5x5, 7x7 matrix.
22736 The filter accepts the following options:
22743 Set matrix for each plane.
22744 Matrix is sequence of 9, 25 or 49 signed numbers.
22745 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22751 Set multiplier for calculated value for each plane.
22752 If unset or 0, it will be sum of all matrix elements.
22753 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22759 Set bias for each plane. This value is added to the result of the multiplication.
22760 Useful for making the overall image brighter or darker.
22761 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22765 @subsection Examples
22771 -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
22777 -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
22781 Apply edge enhance:
22783 -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
22789 -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
22793 Apply laplacian edge detector which includes diagonals:
22795 -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
22801 -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
22805 @section erosion_opencl
22807 Apply erosion effect to the video.
22809 This filter replaces the pixel by the local(3x3) minimum.
22811 It accepts the following options:
22818 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22819 If @code{0}, plane will remain unchanged.
22822 Flag which specifies the pixel to refer to.
22823 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22825 Flags to local 3x3 coordinates region centered on @code{x}:
22834 @subsection Example
22838 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.
22840 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22844 @section deshake_opencl
22845 Feature-point based video stabilization filter.
22847 The filter accepts the following options:
22851 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22854 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22856 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22858 Viewing point matches in the output video is only supported for RGB input.
22860 Defaults to @code{0}.
22862 @item adaptive_crop
22863 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22865 Defaults to @code{1}.
22867 @item refine_features
22868 Whether or not feature points should be refined at a sub-pixel level.
22870 This can be turned off for a slight performance gain at the cost of precision.
22872 Defaults to @code{1}.
22874 @item smooth_strength
22875 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22877 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22879 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22881 Defaults to @code{0.0}.
22883 @item smooth_window_multiplier
22884 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22886 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22888 Acceptable values range from @code{0.1} to @code{10.0}.
22890 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22891 potentially improving smoothness, but also increase latency and memory usage.
22893 Defaults to @code{2.0}.
22897 @subsection Examples
22901 Stabilize a video with a fixed, medium smoothing strength:
22903 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22907 Stabilize a video with debugging (both in console and in rendered video):
22909 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22913 @section dilation_opencl
22915 Apply dilation effect to the video.
22917 This filter replaces the pixel by the local(3x3) maximum.
22919 It accepts the following options:
22926 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22927 If @code{0}, plane will remain unchanged.
22930 Flag which specifies the pixel to refer to.
22931 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22933 Flags to local 3x3 coordinates region centered on @code{x}:
22942 @subsection Example
22946 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.
22948 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22952 @section nlmeans_opencl
22954 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22956 @section overlay_opencl
22958 Overlay one video on top of another.
22960 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22961 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22963 The filter accepts the following options:
22968 Set the x coordinate of the overlaid video on the main video.
22969 Default value is @code{0}.
22972 Set the y coordinate of the overlaid video on the main video.
22973 Default value is @code{0}.
22977 @subsection Examples
22981 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22983 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22986 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22988 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22993 @section pad_opencl
22995 Add paddings to the input image, and place the original input at the
22996 provided @var{x}, @var{y} coordinates.
22998 It accepts the following options:
23003 Specify an expression for the size of the output image with the
23004 paddings added. If the value for @var{width} or @var{height} is 0, the
23005 corresponding input size is used for the output.
23007 The @var{width} expression can reference the value set by the
23008 @var{height} expression, and vice versa.
23010 The default value of @var{width} and @var{height} is 0.
23014 Specify the offsets to place the input image at within the padded area,
23015 with respect to the top/left border of the output image.
23017 The @var{x} expression can reference the value set by the @var{y}
23018 expression, and vice versa.
23020 The default value of @var{x} and @var{y} is 0.
23022 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
23023 so the input image is centered on the padded area.
23026 Specify the color of the padded area. For the syntax of this option,
23027 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
23028 manual,ffmpeg-utils}.
23031 Pad to an aspect instead to a resolution.
23034 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
23035 options are expressions containing the following constants:
23040 The input video width and height.
23044 These are the same as @var{in_w} and @var{in_h}.
23048 The output width and height (the size of the padded area), as
23049 specified by the @var{width} and @var{height} expressions.
23053 These are the same as @var{out_w} and @var{out_h}.
23057 The x and y offsets as specified by the @var{x} and @var{y}
23058 expressions, or NAN if not yet specified.
23061 same as @var{iw} / @var{ih}
23064 input sample aspect ratio
23067 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
23070 @section prewitt_opencl
23072 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
23074 The filter accepts the following option:
23078 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
23081 Set value which will be multiplied with filtered result.
23082 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
23085 Set value which will be added to filtered result.
23086 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
23089 @subsection Example
23093 Apply the Prewitt operator with scale set to 2 and delta set to 10.
23095 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
23099 @anchor{program_opencl}
23100 @section program_opencl
23102 Filter video using an OpenCL program.
23107 OpenCL program source file.
23110 Kernel name in program.
23113 Number of inputs to the filter. Defaults to 1.
23116 Size of output frames. Defaults to the same as the first input.
23120 The @code{program_opencl} filter also supports the @ref{framesync} options.
23122 The program source file must contain a kernel function with the given name,
23123 which will be run once for each plane of the output. Each run on a plane
23124 gets enqueued as a separate 2D global NDRange with one work-item for each
23125 pixel to be generated. The global ID offset for each work-item is therefore
23126 the coordinates of a pixel in the destination image.
23128 The kernel function needs to take the following arguments:
23131 Destination image, @var{__write_only image2d_t}.
23133 This image will become the output; the kernel should write all of it.
23135 Frame index, @var{unsigned int}.
23137 This is a counter starting from zero and increasing by one for each frame.
23139 Source images, @var{__read_only image2d_t}.
23141 These are the most recent images on each input. The kernel may read from
23142 them to generate the output, but they can't be written to.
23149 Copy the input to the output (output must be the same size as the input).
23151 __kernel void copy(__write_only image2d_t destination,
23152 unsigned int index,
23153 __read_only image2d_t source)
23155 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
23157 int2 location = (int2)(get_global_id(0), get_global_id(1));
23159 float4 value = read_imagef(source, sampler, location);
23161 write_imagef(destination, location, value);
23166 Apply a simple transformation, rotating the input by an amount increasing
23167 with the index counter. Pixel values are linearly interpolated by the
23168 sampler, and the output need not have the same dimensions as the input.
23170 __kernel void rotate_image(__write_only image2d_t dst,
23171 unsigned int index,
23172 __read_only image2d_t src)
23174 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23175 CLK_FILTER_LINEAR);
23177 float angle = (float)index / 100.0f;
23179 float2 dst_dim = convert_float2(get_image_dim(dst));
23180 float2 src_dim = convert_float2(get_image_dim(src));
23182 float2 dst_cen = dst_dim / 2.0f;
23183 float2 src_cen = src_dim / 2.0f;
23185 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
23187 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
23189 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
23190 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
23192 src_pos = src_pos * src_dim / dst_dim;
23194 float2 src_loc = src_pos + src_cen;
23196 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
23197 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
23198 write_imagef(dst, dst_loc, 0.5f);
23200 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
23205 Blend two inputs together, with the amount of each input used varying
23206 with the index counter.
23208 __kernel void blend_images(__write_only image2d_t dst,
23209 unsigned int index,
23210 __read_only image2d_t src1,
23211 __read_only image2d_t src2)
23213 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23214 CLK_FILTER_LINEAR);
23216 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
23218 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
23219 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
23220 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
23222 float4 val1 = read_imagef(src1, sampler, src1_loc);
23223 float4 val2 = read_imagef(src2, sampler, src2_loc);
23225 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
23231 @section roberts_opencl
23232 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
23234 The filter accepts the following option:
23238 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
23241 Set value which will be multiplied with filtered result.
23242 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
23245 Set value which will be added to filtered result.
23246 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
23249 @subsection Example
23253 Apply the Roberts cross operator with scale set to 2 and delta set to 10
23255 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
23259 @section sobel_opencl
23261 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
23263 The filter accepts the following option:
23267 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
23270 Set value which will be multiplied with filtered result.
23271 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
23274 Set value which will be added to filtered result.
23275 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
23278 @subsection Example
23282 Apply sobel operator with scale set to 2 and delta set to 10
23284 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
23288 @section tonemap_opencl
23290 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
23292 It accepts the following parameters:
23296 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
23299 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
23302 Apply desaturation for highlights that exceed this level of brightness. The
23303 higher the parameter, the more color information will be preserved. This
23304 setting helps prevent unnaturally blown-out colors for super-highlights, by
23305 (smoothly) turning into white instead. This makes images feel more natural,
23306 at the cost of reducing information about out-of-range colors.
23308 The default value is 0.5, and the algorithm here is a little different from
23309 the cpu version tonemap currently. A setting of 0.0 disables this option.
23312 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
23313 is used to detect whether the scene has changed or not. If the distance between
23314 the current frame average brightness and the current running average exceeds
23315 a threshold value, we would re-calculate scene average and peak brightness.
23316 The default value is 0.2.
23319 Specify the output pixel format.
23321 Currently supported formats are:
23328 Set the output color range.
23330 Possible values are:
23336 Default is same as input.
23339 Set the output color primaries.
23341 Possible values are:
23347 Default is same as input.
23350 Set the output transfer characteristics.
23352 Possible values are:
23361 Set the output colorspace matrix.
23363 Possible value are:
23369 Default is same as input.
23373 @subsection Example
23377 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
23379 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
23383 @section unsharp_opencl
23385 Sharpen or blur the input video.
23387 It accepts the following parameters:
23390 @item luma_msize_x, lx
23391 Set the luma matrix horizontal size.
23392 Range is @code{[1, 23]} and default value is @code{5}.
23394 @item luma_msize_y, ly
23395 Set the luma matrix vertical size.
23396 Range is @code{[1, 23]} and default value is @code{5}.
23398 @item luma_amount, la
23399 Set the luma effect strength.
23400 Range is @code{[-10, 10]} and default value is @code{1.0}.
23402 Negative values will blur the input video, while positive values will
23403 sharpen it, a value of zero will disable the effect.
23405 @item chroma_msize_x, cx
23406 Set the chroma matrix horizontal size.
23407 Range is @code{[1, 23]} and default value is @code{5}.
23409 @item chroma_msize_y, cy
23410 Set the chroma matrix vertical size.
23411 Range is @code{[1, 23]} and default value is @code{5}.
23413 @item chroma_amount, ca
23414 Set the chroma effect strength.
23415 Range is @code{[-10, 10]} and default value is @code{0.0}.
23417 Negative values will blur the input video, while positive values will
23418 sharpen it, a value of zero will disable the effect.
23422 All parameters are optional and default to the equivalent of the
23423 string '5:5:1.0:5:5:0.0'.
23425 @subsection Examples
23429 Apply strong luma sharpen effect:
23431 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
23435 Apply a strong blur of both luma and chroma parameters:
23437 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
23441 @section xfade_opencl
23443 Cross fade two videos with custom transition effect by using OpenCL.
23445 It accepts the following options:
23449 Set one of possible transition effects.
23453 Select custom transition effect, the actual transition description
23454 will be picked from source and kernel options.
23466 Default transition is fade.
23470 OpenCL program source file for custom transition.
23473 Set name of kernel to use for custom transition from program source file.
23476 Set duration of video transition.
23479 Set time of start of transition relative to first video.
23482 The program source file must contain a kernel function with the given name,
23483 which will be run once for each plane of the output. Each run on a plane
23484 gets enqueued as a separate 2D global NDRange with one work-item for each
23485 pixel to be generated. The global ID offset for each work-item is therefore
23486 the coordinates of a pixel in the destination image.
23488 The kernel function needs to take the following arguments:
23491 Destination image, @var{__write_only image2d_t}.
23493 This image will become the output; the kernel should write all of it.
23496 First Source image, @var{__read_only image2d_t}.
23497 Second Source image, @var{__read_only image2d_t}.
23499 These are the most recent images on each input. The kernel may read from
23500 them to generate the output, but they can't be written to.
23503 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
23510 Apply dots curtain transition effect:
23512 __kernel void blend_images(__write_only image2d_t dst,
23513 __read_only image2d_t src1,
23514 __read_only image2d_t src2,
23517 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23518 CLK_FILTER_LINEAR);
23519 int2 p = (int2)(get_global_id(0), get_global_id(1));
23520 float2 rp = (float2)(get_global_id(0), get_global_id(1));
23521 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
23524 float2 dots = (float2)(20.0, 20.0);
23525 float2 center = (float2)(0,0);
23528 float4 val1 = read_imagef(src1, sampler, p);
23529 float4 val2 = read_imagef(src2, sampler, p);
23530 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
23532 write_imagef(dst, p, next ? val1 : val2);
23538 @c man end OPENCL VIDEO FILTERS
23540 @chapter VAAPI Video Filters
23541 @c man begin VAAPI VIDEO FILTERS
23543 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
23545 To enable compilation of these filters you need to configure FFmpeg with
23546 @code{--enable-vaapi}.
23548 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}
23550 @section tonemap_vaapi
23552 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
23553 It maps the dynamic range of HDR10 content to the SDR content.
23554 It currently only accepts HDR10 as input.
23556 It accepts the following parameters:
23560 Specify the output pixel format.
23562 Currently supported formats are:
23571 Set the output color primaries.
23573 Default is same as input.
23576 Set the output transfer characteristics.
23581 Set the output colorspace matrix.
23583 Default is same as input.
23587 @subsection Example
23591 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
23593 tonemap_vaapi=format=p010:t=bt2020-10
23597 @c man end VAAPI VIDEO FILTERS
23599 @chapter Video Sources
23600 @c man begin VIDEO SOURCES
23602 Below is a description of the currently available video sources.
23606 Buffer video frames, and make them available to the filter chain.
23608 This source is mainly intended for a programmatic use, in particular
23609 through the interface defined in @file{libavfilter/buffersrc.h}.
23611 It accepts the following parameters:
23616 Specify the size (width and height) of the buffered video frames. For the
23617 syntax of this option, check the
23618 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23621 The input video width.
23624 The input video height.
23627 A string representing the pixel format of the buffered video frames.
23628 It may be a number corresponding to a pixel format, or a pixel format
23632 Specify the timebase assumed by the timestamps of the buffered frames.
23635 Specify the frame rate expected for the video stream.
23637 @item pixel_aspect, sar
23638 The sample (pixel) aspect ratio of the input video.
23641 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
23642 to the filtergraph description to specify swscale flags for automatically
23643 inserted scalers. See @ref{Filtergraph syntax}.
23645 @item hw_frames_ctx
23646 When using a hardware pixel format, this should be a reference to an
23647 AVHWFramesContext describing input frames.
23652 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23655 will instruct the source to accept video frames with size 320x240 and
23656 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23657 square pixels (1:1 sample aspect ratio).
23658 Since the pixel format with name "yuv410p" corresponds to the number 6
23659 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23660 this example corresponds to:
23662 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23665 Alternatively, the options can be specified as a flat string, but this
23666 syntax is deprecated:
23668 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23672 Create a pattern generated by an elementary cellular automaton.
23674 The initial state of the cellular automaton can be defined through the
23675 @option{filename} and @option{pattern} options. If such options are
23676 not specified an initial state is created randomly.
23678 At each new frame a new row in the video is filled with the result of
23679 the cellular automaton next generation. The behavior when the whole
23680 frame is filled is defined by the @option{scroll} option.
23682 This source accepts the following options:
23686 Read the initial cellular automaton state, i.e. the starting row, from
23687 the specified file.
23688 In the file, each non-whitespace character is considered an alive
23689 cell, a newline will terminate the row, and further characters in the
23690 file will be ignored.
23693 Read the initial cellular automaton state, i.e. the starting row, from
23694 the specified string.
23696 Each non-whitespace character in the string is considered an alive
23697 cell, a newline will terminate the row, and further characters in the
23698 string will be ignored.
23701 Set the video rate, that is the number of frames generated per second.
23704 @item random_fill_ratio, ratio
23705 Set the random fill ratio for the initial cellular automaton row. It
23706 is a floating point number value ranging from 0 to 1, defaults to
23709 This option is ignored when a file or a pattern is specified.
23711 @item random_seed, seed
23712 Set the seed for filling randomly the initial row, must be an integer
23713 included between 0 and UINT32_MAX. If not specified, or if explicitly
23714 set to -1, the filter will try to use a good random seed on a best
23718 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23719 Default value is 110.
23722 Set the size of the output video. For the syntax of this option, check the
23723 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23725 If @option{filename} or @option{pattern} is specified, the size is set
23726 by default to the width of the specified initial state row, and the
23727 height is set to @var{width} * PHI.
23729 If @option{size} is set, it must contain the width of the specified
23730 pattern string, and the specified pattern will be centered in the
23733 If a filename or a pattern string is not specified, the size value
23734 defaults to "320x518" (used for a randomly generated initial state).
23737 If set to 1, scroll the output upward when all the rows in the output
23738 have been already filled. If set to 0, the new generated row will be
23739 written over the top row just after the bottom row is filled.
23742 @item start_full, full
23743 If set to 1, completely fill the output with generated rows before
23744 outputting the first frame.
23745 This is the default behavior, for disabling set the value to 0.
23748 If set to 1, stitch the left and right row edges together.
23749 This is the default behavior, for disabling set the value to 0.
23752 @subsection Examples
23756 Read the initial state from @file{pattern}, and specify an output of
23759 cellauto=f=pattern:s=200x400
23763 Generate a random initial row with a width of 200 cells, with a fill
23766 cellauto=ratio=2/3:s=200x200
23770 Create a pattern generated by rule 18 starting by a single alive cell
23771 centered on an initial row with width 100:
23773 cellauto=p=@@:s=100x400:full=0:rule=18
23777 Specify a more elaborated initial pattern:
23779 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23784 @anchor{coreimagesrc}
23785 @section coreimagesrc
23786 Video source generated on GPU using Apple's CoreImage API on OSX.
23788 This video source is a specialized version of the @ref{coreimage} video filter.
23789 Use a core image generator at the beginning of the applied filterchain to
23790 generate the content.
23792 The coreimagesrc video source accepts the following options:
23794 @item list_generators
23795 List all available generators along with all their respective options as well as
23796 possible minimum and maximum values along with the default values.
23798 list_generators=true
23802 Specify the size of the sourced video. For the syntax of this option, check the
23803 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23804 The default value is @code{320x240}.
23807 Specify the frame rate of the sourced video, as the number of frames
23808 generated per second. It has to be a string in the format
23809 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23810 number or a valid video frame rate abbreviation. The default value is
23814 Set the sample aspect ratio of the sourced video.
23817 Set the duration of the sourced video. See
23818 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23819 for the accepted syntax.
23821 If not specified, or the expressed duration is negative, the video is
23822 supposed to be generated forever.
23825 Additionally, all options of the @ref{coreimage} video filter are accepted.
23826 A complete filterchain can be used for further processing of the
23827 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23828 and examples for details.
23830 @subsection Examples
23835 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23836 given as complete and escaped command-line for Apple's standard bash shell:
23838 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23840 This example is equivalent to the QRCode example of @ref{coreimage} without the
23841 need for a nullsrc video source.
23846 Generate several gradients.
23850 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23851 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23854 Set frame rate, expressed as number of frames per second. Default
23857 @item c0, c1, c2, c3, c4, c5, c6, c7
23858 Set 8 colors. Default values for colors is to pick random one.
23860 @item x0, y0, y0, y1
23861 Set gradient line source and destination points. If negative or out of range, random ones
23865 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23868 Set seed for picking gradient line points.
23871 Set the duration of the sourced video. See
23872 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23873 for the accepted syntax.
23875 If not specified, or the expressed duration is negative, the video is
23876 supposed to be generated forever.
23879 Set speed of gradients rotation.
23883 @section mandelbrot
23885 Generate a Mandelbrot set fractal, and progressively zoom towards the
23886 point specified with @var{start_x} and @var{start_y}.
23888 This source accepts the following options:
23893 Set the terminal pts value. Default value is 400.
23896 Set the terminal scale value.
23897 Must be a floating point value. Default value is 0.3.
23900 Set the inner coloring mode, that is the algorithm used to draw the
23901 Mandelbrot fractal internal region.
23903 It shall assume one of the following values:
23908 Show time until convergence.
23910 Set color based on point closest to the origin of the iterations.
23915 Default value is @var{mincol}.
23918 Set the bailout value. Default value is 10.0.
23921 Set the maximum of iterations performed by the rendering
23922 algorithm. Default value is 7189.
23925 Set outer coloring mode.
23926 It shall assume one of following values:
23928 @item iteration_count
23929 Set iteration count mode.
23930 @item normalized_iteration_count
23931 set normalized iteration count mode.
23933 Default value is @var{normalized_iteration_count}.
23936 Set frame rate, expressed as number of frames per second. Default
23940 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23941 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23944 Set the initial scale value. Default value is 3.0.
23947 Set the initial x position. Must be a floating point value between
23948 -100 and 100. Default value is -0.743643887037158704752191506114774.
23951 Set the initial y position. Must be a floating point value between
23952 -100 and 100. Default value is -0.131825904205311970493132056385139.
23957 Generate various test patterns, as generated by the MPlayer test filter.
23959 The size of the generated video is fixed, and is 256x256.
23960 This source is useful in particular for testing encoding features.
23962 This source accepts the following options:
23967 Specify the frame rate of the sourced video, as the number of frames
23968 generated per second. It has to be a string in the format
23969 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23970 number or a valid video frame rate abbreviation. The default value is
23974 Set the duration of the sourced video. See
23975 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23976 for the accepted syntax.
23978 If not specified, or the expressed duration is negative, the video is
23979 supposed to be generated forever.
23983 Set the number or the name of the test to perform. Supported tests are:
23997 @item max_frames, m
23998 Set the maximum number of frames generated for each test, default value is 30.
24002 Default value is "all", which will cycle through the list of all tests.
24007 mptestsrc=t=dc_luma
24010 will generate a "dc_luma" test pattern.
24012 @section frei0r_src
24014 Provide a frei0r source.
24016 To enable compilation of this filter you need to install the frei0r
24017 header and configure FFmpeg with @code{--enable-frei0r}.
24019 This source accepts the following parameters:
24024 The size of the video to generate. For the syntax of this option, check the
24025 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24028 The framerate of the generated video. It may be a string of the form
24029 @var{num}/@var{den} or a frame rate abbreviation.
24032 The name to the frei0r source to load. For more information regarding frei0r and
24033 how to set the parameters, read the @ref{frei0r} section in the video filters
24036 @item filter_params
24037 A '|'-separated list of parameters to pass to the frei0r source.
24041 For example, to generate a frei0r partik0l source with size 200x200
24042 and frame rate 10 which is overlaid on the overlay filter main input:
24044 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
24049 Generate a life pattern.
24051 This source is based on a generalization of John Conway's life game.
24053 The sourced input represents a life grid, each pixel represents a cell
24054 which can be in one of two possible states, alive or dead. Every cell
24055 interacts with its eight neighbours, which are the cells that are
24056 horizontally, vertically, or diagonally adjacent.
24058 At each interaction the grid evolves according to the adopted rule,
24059 which specifies the number of neighbor alive cells which will make a
24060 cell stay alive or born. The @option{rule} option allows one to specify
24063 This source accepts the following options:
24067 Set the file from which to read the initial grid state. In the file,
24068 each non-whitespace character is considered an alive cell, and newline
24069 is used to delimit the end of each row.
24071 If this option is not specified, the initial grid is generated
24075 Set the video rate, that is the number of frames generated per second.
24078 @item random_fill_ratio, ratio
24079 Set the random fill ratio for the initial random grid. It is a
24080 floating point number value ranging from 0 to 1, defaults to 1/PHI.
24081 It is ignored when a file is specified.
24083 @item random_seed, seed
24084 Set the seed for filling the initial random grid, must be an integer
24085 included between 0 and UINT32_MAX. If not specified, or if explicitly
24086 set to -1, the filter will try to use a good random seed on a best
24092 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
24093 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
24094 @var{NS} specifies the number of alive neighbor cells which make a
24095 live cell stay alive, and @var{NB} the number of alive neighbor cells
24096 which make a dead cell to become alive (i.e. to "born").
24097 "s" and "b" can be used in place of "S" and "B", respectively.
24099 Alternatively a rule can be specified by an 18-bits integer. The 9
24100 high order bits are used to encode the next cell state if it is alive
24101 for each number of neighbor alive cells, the low order bits specify
24102 the rule for "borning" new cells. Higher order bits encode for an
24103 higher number of neighbor cells.
24104 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
24105 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
24107 Default value is "S23/B3", which is the original Conway's game of life
24108 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
24109 cells, and will born a new cell if there are three alive cells around
24113 Set the size of the output video. For the syntax of this option, check the
24114 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24116 If @option{filename} is specified, the size is set by default to the
24117 same size of the input file. If @option{size} is set, it must contain
24118 the size specified in the input file, and the initial grid defined in
24119 that file is centered in the larger resulting area.
24121 If a filename is not specified, the size value defaults to "320x240"
24122 (used for a randomly generated initial grid).
24125 If set to 1, stitch the left and right grid edges together, and the
24126 top and bottom edges also. Defaults to 1.
24129 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
24130 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
24131 value from 0 to 255.
24134 Set the color of living (or new born) cells.
24137 Set the color of dead cells. If @option{mold} is set, this is the first color
24138 used to represent a dead cell.
24141 Set mold color, for definitely dead and moldy cells.
24143 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
24144 ffmpeg-utils manual,ffmpeg-utils}.
24147 @subsection Examples
24151 Read a grid from @file{pattern}, and center it on a grid of size
24154 life=f=pattern:s=300x300
24158 Generate a random grid of size 200x200, with a fill ratio of 2/3:
24160 life=ratio=2/3:s=200x200
24164 Specify a custom rule for evolving a randomly generated grid:
24170 Full example with slow death effect (mold) using @command{ffplay}:
24172 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
24179 @anchor{haldclutsrc}
24182 @anchor{pal100bars}
24183 @anchor{rgbtestsrc}
24185 @anchor{smptehdbars}
24188 @anchor{yuvtestsrc}
24189 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
24191 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
24193 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
24195 The @code{color} source provides an uniformly colored input.
24197 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
24198 @ref{haldclut} filter.
24200 The @code{nullsrc} source returns unprocessed video frames. It is
24201 mainly useful to be employed in analysis / debugging tools, or as the
24202 source for filters which ignore the input data.
24204 The @code{pal75bars} source generates a color bars pattern, based on
24205 EBU PAL recommendations with 75% color levels.
24207 The @code{pal100bars} source generates a color bars pattern, based on
24208 EBU PAL recommendations with 100% color levels.
24210 The @code{rgbtestsrc} source generates an RGB test pattern useful for
24211 detecting RGB vs BGR issues. You should see a red, green and blue
24212 stripe from top to bottom.
24214 The @code{smptebars} source generates a color bars pattern, based on
24215 the SMPTE Engineering Guideline EG 1-1990.
24217 The @code{smptehdbars} source generates a color bars pattern, based on
24218 the SMPTE RP 219-2002.
24220 The @code{testsrc} source generates a test video pattern, showing a
24221 color pattern, a scrolling gradient and a timestamp. This is mainly
24222 intended for testing purposes.
24224 The @code{testsrc2} source is similar to testsrc, but supports more
24225 pixel formats instead of just @code{rgb24}. This allows using it as an
24226 input for other tests without requiring a format conversion.
24228 The @code{yuvtestsrc} source generates an YUV test pattern. You should
24229 see a y, cb and cr stripe from top to bottom.
24231 The sources accept the following parameters:
24236 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
24237 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
24238 pixels to be used as identity matrix for 3D lookup tables. Each component is
24239 coded on a @code{1/(N*N)} scale.
24242 Specify the color of the source, only available in the @code{color}
24243 source. For the syntax of this option, check the
24244 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
24247 Specify the size of the sourced video. For the syntax of this option, check the
24248 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24249 The default value is @code{320x240}.
24251 This option is not available with the @code{allrgb}, @code{allyuv}, and
24252 @code{haldclutsrc} filters.
24255 Specify the frame rate of the sourced video, as the number of frames
24256 generated per second. It has to be a string in the format
24257 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
24258 number or a valid video frame rate abbreviation. The default value is
24262 Set the duration of the sourced video. See
24263 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
24264 for the accepted syntax.
24266 If not specified, or the expressed duration is negative, the video is
24267 supposed to be generated forever.
24269 Since the frame rate is used as time base, all frames including the last one
24270 will have their full duration. If the specified duration is not a multiple
24271 of the frame duration, it will be rounded up.
24274 Set the sample aspect ratio of the sourced video.
24277 Specify the alpha (opacity) of the background, only available in the
24278 @code{testsrc2} source. The value must be between 0 (fully transparent) and
24279 255 (fully opaque, the default).
24282 Set the number of decimals to show in the timestamp, only available in the
24283 @code{testsrc} source.
24285 The displayed timestamp value will correspond to the original
24286 timestamp value multiplied by the power of 10 of the specified
24287 value. Default value is 0.
24290 @subsection Examples
24294 Generate a video with a duration of 5.3 seconds, with size
24295 176x144 and a frame rate of 10 frames per second:
24297 testsrc=duration=5.3:size=qcif:rate=10
24301 The following graph description will generate a red source
24302 with an opacity of 0.2, with size "qcif" and a frame rate of 10
24305 color=c=red@@0.2:s=qcif:r=10
24309 If the input content is to be ignored, @code{nullsrc} can be used. The
24310 following command generates noise in the luminance plane by employing
24311 the @code{geq} filter:
24313 nullsrc=s=256x256, geq=random(1)*255:128:128
24317 @subsection Commands
24319 The @code{color} source supports the following commands:
24323 Set the color of the created image. Accepts the same syntax of the
24324 corresponding @option{color} option.
24329 Generate video using an OpenCL program.
24334 OpenCL program source file.
24337 Kernel name in program.
24340 Size of frames to generate. This must be set.
24343 Pixel format to use for the generated frames. This must be set.
24346 Number of frames generated every second. Default value is '25'.
24350 For details of how the program loading works, see the @ref{program_opencl}
24357 Generate a colour ramp by setting pixel values from the position of the pixel
24358 in the output image. (Note that this will work with all pixel formats, but
24359 the generated output will not be the same.)
24361 __kernel void ramp(__write_only image2d_t dst,
24362 unsigned int index)
24364 int2 loc = (int2)(get_global_id(0), get_global_id(1));
24367 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
24369 write_imagef(dst, loc, val);
24374 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
24376 __kernel void sierpinski_carpet(__write_only image2d_t dst,
24377 unsigned int index)
24379 int2 loc = (int2)(get_global_id(0), get_global_id(1));
24381 float4 value = 0.0f;
24382 int x = loc.x + index;
24383 int y = loc.y + index;
24384 while (x > 0 || y > 0) {
24385 if (x % 3 == 1 && y % 3 == 1) {
24393 write_imagef(dst, loc, value);
24399 @section sierpinski
24401 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
24403 This source accepts the following options:
24407 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
24408 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
24411 Set frame rate, expressed as number of frames per second. Default
24415 Set seed which is used for random panning.
24418 Set max jump for single pan destination. Allowed range is from 1 to 10000.
24421 Set fractal type, can be default @code{carpet} or @code{triangle}.
24424 @c man end VIDEO SOURCES
24426 @chapter Video Sinks
24427 @c man begin VIDEO SINKS
24429 Below is a description of the currently available video sinks.
24431 @section buffersink
24433 Buffer video frames, and make them available to the end of the filter
24436 This sink is mainly intended for programmatic use, in particular
24437 through the interface defined in @file{libavfilter/buffersink.h}
24438 or the options system.
24440 It accepts a pointer to an AVBufferSinkContext structure, which
24441 defines the incoming buffers' formats, to be passed as the opaque
24442 parameter to @code{avfilter_init_filter} for initialization.
24446 Null video sink: do absolutely nothing with the input video. It is
24447 mainly useful as a template and for use in analysis / debugging
24450 @c man end VIDEO SINKS
24452 @chapter Multimedia Filters
24453 @c man begin MULTIMEDIA FILTERS
24455 Below is a description of the currently available multimedia filters.
24459 Convert input audio to a video output, displaying the audio bit scope.
24461 The filter accepts the following options:
24465 Set frame rate, expressed as number of frames per second. Default
24469 Specify the video size for the output. For the syntax of this option, check the
24470 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24471 Default value is @code{1024x256}.
24474 Specify list of colors separated by space or by '|' which will be used to
24475 draw channels. Unrecognized or missing colors will be replaced
24479 @section adrawgraph
24480 Draw a graph using input audio metadata.
24482 See @ref{drawgraph}
24484 @section agraphmonitor
24486 See @ref{graphmonitor}.
24488 @section ahistogram
24490 Convert input audio to a video output, displaying the volume histogram.
24492 The filter accepts the following options:
24496 Specify how histogram is calculated.
24498 It accepts the following values:
24501 Use single histogram for all channels.
24503 Use separate histogram for each channel.
24505 Default is @code{single}.
24508 Set frame rate, expressed as number of frames per second. Default
24512 Specify the video size for the output. For the syntax of this option, check the
24513 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24514 Default value is @code{hd720}.
24519 It accepts the following values:
24530 reverse logarithmic
24532 Default is @code{log}.
24535 Set amplitude scale.
24537 It accepts the following values:
24544 Default is @code{log}.
24547 Set how much frames to accumulate in histogram.
24548 Default is 1. Setting this to -1 accumulates all frames.
24551 Set histogram ratio of window height.
24554 Set sonogram sliding.
24556 It accepts the following values:
24559 replace old rows with new ones.
24561 scroll from top to bottom.
24563 Default is @code{replace}.
24566 @section aphasemeter
24568 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
24569 representing mean phase of current audio frame. A video output can also be produced and is
24570 enabled by default. The audio is passed through as first output.
24572 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
24573 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
24574 and @code{1} means channels are in phase.
24576 The filter accepts the following options, all related to its video output:
24580 Set the output frame rate. Default value is @code{25}.
24583 Set the video size for the output. For the syntax of this option, check the
24584 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24585 Default value is @code{800x400}.
24590 Specify the red, green, blue contrast. Default values are @code{2},
24591 @code{7} and @code{1}.
24592 Allowed range is @code{[0, 255]}.
24595 Set color which will be used for drawing median phase. If color is
24596 @code{none} which is default, no median phase value will be drawn.
24599 Enable video output. Default is enabled.
24602 @subsection phasing detection
24604 The filter also detects out of phase and mono sequences in stereo streams.
24605 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
24607 The filter accepts the following options for this detection:
24611 Enable mono and out of phase detection. Default is disabled.
24614 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
24615 Allowed range is @code{[0, 1]}.
24618 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
24619 Allowed range is @code{[90, 180]}.
24622 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
24625 @subsection Examples
24629 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
24631 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
24635 @section avectorscope
24637 Convert input audio to a video output, representing the audio vector
24640 The filter is used to measure the difference between channels of stereo
24641 audio stream. A monaural signal, consisting of identical left and right
24642 signal, results in straight vertical line. Any stereo separation is visible
24643 as a deviation from this line, creating a Lissajous figure.
24644 If the straight (or deviation from it) but horizontal line appears this
24645 indicates that the left and right channels are out of phase.
24647 The filter accepts the following options:
24651 Set the vectorscope mode.
24653 Available values are:
24656 Lissajous rotated by 45 degrees.
24659 Same as above but not rotated.
24662 Shape resembling half of circle.
24665 Default value is @samp{lissajous}.
24668 Set the video size for the output. For the syntax of this option, check the
24669 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24670 Default value is @code{400x400}.
24673 Set the output frame rate. Default value is @code{25}.
24679 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24680 @code{160}, @code{80} and @code{255}.
24681 Allowed range is @code{[0, 255]}.
24687 Specify the red, green, blue and alpha fade. Default values are @code{15},
24688 @code{10}, @code{5} and @code{5}.
24689 Allowed range is @code{[0, 255]}.
24692 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24693 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24696 Set the vectorscope drawing mode.
24698 Available values are:
24701 Draw dot for each sample.
24704 Draw line between previous and current sample.
24707 Default value is @samp{dot}.
24710 Specify amplitude scale of audio samples.
24712 Available values are:
24728 Swap left channel axis with right channel axis.
24738 Mirror only x axis.
24741 Mirror only y axis.
24749 @subsection Examples
24753 Complete example using @command{ffplay}:
24755 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24756 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24760 @section bench, abench
24762 Benchmark part of a filtergraph.
24764 The filter accepts the following options:
24768 Start or stop a timer.
24770 Available values are:
24773 Get the current time, set it as frame metadata (using the key
24774 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24777 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24778 the input frame metadata to get the time difference. Time difference, average,
24779 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24780 @code{min}) are then printed. The timestamps are expressed in seconds.
24784 @subsection Examples
24788 Benchmark @ref{selectivecolor} filter:
24790 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24796 Concatenate audio and video streams, joining them together one after the
24799 The filter works on segments of synchronized video and audio streams. All
24800 segments must have the same number of streams of each type, and that will
24801 also be the number of streams at output.
24803 The filter accepts the following options:
24808 Set the number of segments. Default is 2.
24811 Set the number of output video streams, that is also the number of video
24812 streams in each segment. Default is 1.
24815 Set the number of output audio streams, that is also the number of audio
24816 streams in each segment. Default is 0.
24819 Activate unsafe mode: do not fail if segments have a different format.
24823 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24824 @var{a} audio outputs.
24826 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24827 segment, in the same order as the outputs, then the inputs for the second
24830 Related streams do not always have exactly the same duration, for various
24831 reasons including codec frame size or sloppy authoring. For that reason,
24832 related synchronized streams (e.g. a video and its audio track) should be
24833 concatenated at once. The concat filter will use the duration of the longest
24834 stream in each segment (except the last one), and if necessary pad shorter
24835 audio streams with silence.
24837 For this filter to work correctly, all segments must start at timestamp 0.
24839 All corresponding streams must have the same parameters in all segments; the
24840 filtering system will automatically select a common pixel format for video
24841 streams, and a common sample format, sample rate and channel layout for
24842 audio streams, but other settings, such as resolution, must be converted
24843 explicitly by the user.
24845 Different frame rates are acceptable but will result in variable frame rate
24846 at output; be sure to configure the output file to handle it.
24848 @subsection Examples
24852 Concatenate an opening, an episode and an ending, all in bilingual version
24853 (video in stream 0, audio in streams 1 and 2):
24855 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24856 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24857 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24858 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24862 Concatenate two parts, handling audio and video separately, using the
24863 (a)movie sources, and adjusting the resolution:
24865 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24866 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24867 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24869 Note that a desync will happen at the stitch if the audio and video streams
24870 do not have exactly the same duration in the first file.
24874 @subsection Commands
24876 This filter supports the following commands:
24879 Close the current segment and step to the next one
24885 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24886 level. By default, it logs a message at a frequency of 10Hz with the
24887 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24888 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24890 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24891 sample format is double-precision floating point. The input stream will be converted to
24892 this specification, if needed. Users may need to insert aformat and/or aresample filters
24893 after this filter to obtain the original parameters.
24895 The filter also has a video output (see the @var{video} option) with a real
24896 time graph to observe the loudness evolution. The graphic contains the logged
24897 message mentioned above, so it is not printed anymore when this option is set,
24898 unless the verbose logging is set. The main graphing area contains the
24899 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24900 the momentary loudness (400 milliseconds), but can optionally be configured
24901 to instead display short-term loudness (see @var{gauge}).
24903 The green area marks a +/- 1LU target range around the target loudness
24904 (-23LUFS by default, unless modified through @var{target}).
24906 More information about the Loudness Recommendation EBU R128 on
24907 @url{http://tech.ebu.ch/loudness}.
24909 The filter accepts the following options:
24914 Activate the video output. The audio stream is passed unchanged whether this
24915 option is set or no. The video stream will be the first output stream if
24916 activated. Default is @code{0}.
24919 Set the video size. This option is for video only. For the syntax of this
24921 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24922 Default and minimum resolution is @code{640x480}.
24925 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24926 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24927 other integer value between this range is allowed.
24930 Set metadata injection. If set to @code{1}, the audio input will be segmented
24931 into 100ms output frames, each of them containing various loudness information
24932 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24934 Default is @code{0}.
24937 Force the frame logging level.
24939 Available values are:
24942 information logging level
24944 verbose logging level
24947 By default, the logging level is set to @var{info}. If the @option{video} or
24948 the @option{metadata} options are set, it switches to @var{verbose}.
24953 Available modes can be cumulated (the option is a @code{flag} type). Possible
24957 Disable any peak mode (default).
24959 Enable sample-peak mode.
24961 Simple peak mode looking for the higher sample value. It logs a message
24962 for sample-peak (identified by @code{SPK}).
24964 Enable true-peak mode.
24966 If enabled, the peak lookup is done on an over-sampled version of the input
24967 stream for better peak accuracy. It logs a message for true-peak.
24968 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24969 This mode requires a build with @code{libswresample}.
24973 Treat mono input files as "dual mono". If a mono file is intended for playback
24974 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24975 If set to @code{true}, this option will compensate for this effect.
24976 Multi-channel input files are not affected by this option.
24979 Set a specific pan law to be used for the measurement of dual mono files.
24980 This parameter is optional, and has a default value of -3.01dB.
24983 Set a specific target level (in LUFS) used as relative zero in the visualization.
24984 This parameter is optional and has a default value of -23LUFS as specified
24985 by EBU R128. However, material published online may prefer a level of -16LUFS
24986 (e.g. for use with podcasts or video platforms).
24989 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24990 @code{shortterm}. By default the momentary value will be used, but in certain
24991 scenarios it may be more useful to observe the short term value instead (e.g.
24995 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24996 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24997 video output, not the summary or continuous log output.
25000 @subsection Examples
25004 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
25006 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
25010 Run an analysis with @command{ffmpeg}:
25012 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
25016 @section interleave, ainterleave
25018 Temporally interleave frames from several inputs.
25020 @code{interleave} works with video inputs, @code{ainterleave} with audio.
25022 These filters read frames from several inputs and send the oldest
25023 queued frame to the output.
25025 Input streams must have well defined, monotonically increasing frame
25028 In order to submit one frame to output, these filters need to enqueue
25029 at least one frame for each input, so they cannot work in case one
25030 input is not yet terminated and will not receive incoming frames.
25032 For example consider the case when one input is a @code{select} filter
25033 which always drops input frames. The @code{interleave} filter will keep
25034 reading from that input, but it will never be able to send new frames
25035 to output until the input sends an end-of-stream signal.
25037 Also, depending on inputs synchronization, the filters will drop
25038 frames in case one input receives more frames than the other ones, and
25039 the queue is already filled.
25041 These filters accept the following options:
25045 Set the number of different inputs, it is 2 by default.
25048 How to determine the end-of-stream.
25052 The duration of the longest input. (default)
25055 The duration of the shortest input.
25058 The duration of the first input.
25063 @subsection Examples
25067 Interleave frames belonging to different streams using @command{ffmpeg}:
25069 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
25073 Add flickering blur effect:
25075 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
25079 @section metadata, ametadata
25081 Manipulate frame metadata.
25083 This filter accepts the following options:
25087 Set mode of operation of the filter.
25089 Can be one of the following:
25093 If both @code{value} and @code{key} is set, select frames
25094 which have such metadata. If only @code{key} is set, select
25095 every frame that has such key in metadata.
25098 Add new metadata @code{key} and @code{value}. If key is already available
25102 Modify value of already present key.
25105 If @code{value} is set, delete only keys that have such value.
25106 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
25110 Print key and its value if metadata was found. If @code{key} is not set print all
25111 metadata values available in frame.
25115 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
25118 Set metadata value which will be used. This option is mandatory for
25119 @code{modify} and @code{add} mode.
25122 Which function to use when comparing metadata value and @code{value}.
25124 Can be one of following:
25128 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
25131 Values are interpreted as strings, returns true if metadata value starts with
25132 the @code{value} option string.
25135 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
25138 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
25141 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
25144 Values are interpreted as floats, returns true if expression from option @code{expr}
25148 Values are interpreted as strings, returns true if metadata value ends with
25149 the @code{value} option string.
25153 Set expression which is used when @code{function} is set to @code{expr}.
25154 The expression is evaluated through the eval API and can contain the following
25159 Float representation of @code{value} from metadata key.
25162 Float representation of @code{value} as supplied by user in @code{value} option.
25166 If specified in @code{print} mode, output is written to the named file. Instead of
25167 plain filename any writable url can be specified. Filename ``-'' is a shorthand
25168 for standard output. If @code{file} option is not set, output is written to the log
25169 with AV_LOG_INFO loglevel.
25172 Reduces buffering in print mode when output is written to a URL set using @var{file}.
25176 @subsection Examples
25180 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
25183 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
25186 Print silencedetect output to file @file{metadata.txt}.
25188 silencedetect,ametadata=mode=print:file=metadata.txt
25191 Direct all metadata to a pipe with file descriptor 4.
25193 metadata=mode=print:file='pipe\:4'
25197 @section perms, aperms
25199 Set read/write permissions for the output frames.
25201 These filters are mainly aimed at developers to test direct path in the
25202 following filter in the filtergraph.
25204 The filters accept the following options:
25208 Select the permissions mode.
25210 It accepts the following values:
25213 Do nothing. This is the default.
25215 Set all the output frames read-only.
25217 Set all the output frames directly writable.
25219 Make the frame read-only if writable, and writable if read-only.
25221 Set each output frame read-only or writable randomly.
25225 Set the seed for the @var{random} mode, must be an integer included between
25226 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
25227 @code{-1}, the filter will try to use a good random seed on a best effort
25231 Note: in case of auto-inserted filter between the permission filter and the
25232 following one, the permission might not be received as expected in that
25233 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
25234 perms/aperms filter can avoid this problem.
25236 @section realtime, arealtime
25238 Slow down filtering to match real time approximately.
25240 These filters will pause the filtering for a variable amount of time to
25241 match the output rate with the input timestamps.
25242 They are similar to the @option{re} option to @code{ffmpeg}.
25244 They accept the following options:
25248 Time limit for the pauses. Any pause longer than that will be considered
25249 a timestamp discontinuity and reset the timer. Default is 2 seconds.
25251 Speed factor for processing. The value must be a float larger than zero.
25252 Values larger than 1.0 will result in faster than realtime processing,
25253 smaller will slow processing down. The @var{limit} is automatically adapted
25254 accordingly. Default is 1.0.
25256 A processing speed faster than what is possible without these filters cannot
25261 @section select, aselect
25263 Select frames to pass in output.
25265 This filter accepts the following options:
25270 Set expression, which is evaluated for each input frame.
25272 If the expression is evaluated to zero, the frame is discarded.
25274 If the evaluation result is negative or NaN, the frame is sent to the
25275 first output; otherwise it is sent to the output with index
25276 @code{ceil(val)-1}, assuming that the input index starts from 0.
25278 For example a value of @code{1.2} corresponds to the output with index
25279 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
25282 Set the number of outputs. The output to which to send the selected
25283 frame is based on the result of the evaluation. Default value is 1.
25286 The expression can contain the following constants:
25290 The (sequential) number of the filtered frame, starting from 0.
25293 The (sequential) number of the selected frame, starting from 0.
25295 @item prev_selected_n
25296 The sequential number of the last selected frame. It's NAN if undefined.
25299 The timebase of the input timestamps.
25302 The PTS (Presentation TimeStamp) of the filtered video frame,
25303 expressed in @var{TB} units. It's NAN if undefined.
25306 The PTS of the filtered video frame,
25307 expressed in seconds. It's NAN if undefined.
25310 The PTS of the previously filtered video frame. It's NAN if undefined.
25312 @item prev_selected_pts
25313 The PTS of the last previously filtered video frame. It's NAN if undefined.
25315 @item prev_selected_t
25316 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
25319 The PTS of the first video frame in the video. It's NAN if undefined.
25322 The time of the first video frame in the video. It's NAN if undefined.
25324 @item pict_type @emph{(video only)}
25325 The type of the filtered frame. It can assume one of the following
25337 @item interlace_type @emph{(video only)}
25338 The frame interlace type. It can assume one of the following values:
25341 The frame is progressive (not interlaced).
25343 The frame is top-field-first.
25345 The frame is bottom-field-first.
25348 @item consumed_sample_n @emph{(audio only)}
25349 the number of selected samples before the current frame
25351 @item samples_n @emph{(audio only)}
25352 the number of samples in the current frame
25354 @item sample_rate @emph{(audio only)}
25355 the input sample rate
25358 This is 1 if the filtered frame is a key-frame, 0 otherwise.
25361 the position in the file of the filtered frame, -1 if the information
25362 is not available (e.g. for synthetic video)
25364 @item scene @emph{(video only)}
25365 value between 0 and 1 to indicate a new scene; a low value reflects a low
25366 probability for the current frame to introduce a new scene, while a higher
25367 value means the current frame is more likely to be one (see the example below)
25369 @item concatdec_select
25370 The concat demuxer can select only part of a concat input file by setting an
25371 inpoint and an outpoint, but the output packets may not be entirely contained
25372 in the selected interval. By using this variable, it is possible to skip frames
25373 generated by the concat demuxer which are not exactly contained in the selected
25376 This works by comparing the frame pts against the @var{lavf.concat.start_time}
25377 and the @var{lavf.concat.duration} packet metadata values which are also
25378 present in the decoded frames.
25380 The @var{concatdec_select} variable is -1 if the frame pts is at least
25381 start_time and either the duration metadata is missing or the frame pts is less
25382 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
25385 That basically means that an input frame is selected if its pts is within the
25386 interval set by the concat demuxer.
25390 The default value of the select expression is "1".
25392 @subsection Examples
25396 Select all frames in input:
25401 The example above is the same as:
25413 Select only I-frames:
25415 select='eq(pict_type\,I)'
25419 Select one frame every 100:
25421 select='not(mod(n\,100))'
25425 Select only frames contained in the 10-20 time interval:
25427 select=between(t\,10\,20)
25431 Select only I-frames contained in the 10-20 time interval:
25433 select=between(t\,10\,20)*eq(pict_type\,I)
25437 Select frames with a minimum distance of 10 seconds:
25439 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
25443 Use aselect to select only audio frames with samples number > 100:
25445 aselect='gt(samples_n\,100)'
25449 Create a mosaic of the first scenes:
25451 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
25454 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
25458 Send even and odd frames to separate outputs, and compose them:
25460 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
25464 Select useful frames from an ffconcat file which is using inpoints and
25465 outpoints but where the source files are not intra frame only.
25467 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
25471 @section sendcmd, asendcmd
25473 Send commands to filters in the filtergraph.
25475 These filters read commands to be sent to other filters in the
25478 @code{sendcmd} must be inserted between two video filters,
25479 @code{asendcmd} must be inserted between two audio filters, but apart
25480 from that they act the same way.
25482 The specification of commands can be provided in the filter arguments
25483 with the @var{commands} option, or in a file specified by the
25484 @var{filename} option.
25486 These filters accept the following options:
25489 Set the commands to be read and sent to the other filters.
25491 Set the filename of the commands to be read and sent to the other
25495 @subsection Commands syntax
25497 A commands description consists of a sequence of interval
25498 specifications, comprising a list of commands to be executed when a
25499 particular event related to that interval occurs. The occurring event
25500 is typically the current frame time entering or leaving a given time
25503 An interval is specified by the following syntax:
25505 @var{START}[-@var{END}] @var{COMMANDS};
25508 The time interval is specified by the @var{START} and @var{END} times.
25509 @var{END} is optional and defaults to the maximum time.
25511 The current frame time is considered within the specified interval if
25512 it is included in the interval [@var{START}, @var{END}), that is when
25513 the time is greater or equal to @var{START} and is lesser than
25516 @var{COMMANDS} consists of a sequence of one or more command
25517 specifications, separated by ",", relating to that interval. The
25518 syntax of a command specification is given by:
25520 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
25523 @var{FLAGS} is optional and specifies the type of events relating to
25524 the time interval which enable sending the specified command, and must
25525 be a non-null sequence of identifier flags separated by "+" or "|" and
25526 enclosed between "[" and "]".
25528 The following flags are recognized:
25531 The command is sent when the current frame timestamp enters the
25532 specified interval. In other words, the command is sent when the
25533 previous frame timestamp was not in the given interval, and the
25537 The command is sent when the current frame timestamp leaves the
25538 specified interval. In other words, the command is sent when the
25539 previous frame timestamp was in the given interval, and the
25543 The command @var{ARG} is interpreted as expression and result of
25544 expression is passed as @var{ARG}.
25546 The expression is evaluated through the eval API and can contain the following
25551 Original position in the file of the frame, or undefined if undefined
25552 for the current frame.
25555 The presentation timestamp in input.
25558 The count of the input frame for video or audio, starting from 0.
25561 The time in seconds of the current frame.
25564 The start time in seconds of the current command interval.
25567 The end time in seconds of the current command interval.
25570 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
25575 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
25578 @var{TARGET} specifies the target of the command, usually the name of
25579 the filter class or a specific filter instance name.
25581 @var{COMMAND} specifies the name of the command for the target filter.
25583 @var{ARG} is optional and specifies the optional list of argument for
25584 the given @var{COMMAND}.
25586 Between one interval specification and another, whitespaces, or
25587 sequences of characters starting with @code{#} until the end of line,
25588 are ignored and can be used to annotate comments.
25590 A simplified BNF description of the commands specification syntax
25593 @var{COMMAND_FLAG} ::= "enter" | "leave"
25594 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
25595 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
25596 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
25597 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
25598 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
25601 @subsection Examples
25605 Specify audio tempo change at second 4:
25607 asendcmd=c='4.0 atempo tempo 1.5',atempo
25611 Target a specific filter instance:
25613 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
25617 Specify a list of drawtext and hue commands in a file.
25619 # show text in the interval 5-10
25620 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
25621 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
25623 # desaturate the image in the interval 15-20
25624 15.0-20.0 [enter] hue s 0,
25625 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
25627 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
25629 # apply an exponential saturation fade-out effect, starting from time 25
25630 25 [enter] hue s exp(25-t)
25633 A filtergraph allowing to read and process the above command list
25634 stored in a file @file{test.cmd}, can be specified with:
25636 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
25641 @section setpts, asetpts
25643 Change the PTS (presentation timestamp) of the input frames.
25645 @code{setpts} works on video frames, @code{asetpts} on audio frames.
25647 This filter accepts the following options:
25652 The expression which is evaluated for each frame to construct its timestamp.
25656 The expression is evaluated through the eval API and can contain the following
25660 @item FRAME_RATE, FR
25661 frame rate, only defined for constant frame-rate video
25664 The presentation timestamp in input
25667 The count of the input frame for video or the number of consumed samples,
25668 not including the current frame for audio, starting from 0.
25670 @item NB_CONSUMED_SAMPLES
25671 The number of consumed samples, not including the current frame (only
25674 @item NB_SAMPLES, S
25675 The number of samples in the current frame (only audio)
25677 @item SAMPLE_RATE, SR
25678 The audio sample rate.
25681 The PTS of the first frame.
25684 the time in seconds of the first frame
25687 State whether the current frame is interlaced.
25690 the time in seconds of the current frame
25693 original position in the file of the frame, or undefined if undefined
25694 for the current frame
25697 The previous input PTS.
25700 previous input time in seconds
25703 The previous output PTS.
25706 previous output time in seconds
25709 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25713 The wallclock (RTC) time at the start of the movie in microseconds.
25716 The timebase of the input timestamps.
25720 @subsection Examples
25724 Start counting PTS from zero
25726 setpts=PTS-STARTPTS
25730 Apply fast motion effect:
25736 Apply slow motion effect:
25742 Set fixed rate of 25 frames per second:
25748 Set fixed rate 25 fps with some jitter:
25750 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25754 Apply an offset of 10 seconds to the input PTS:
25760 Generate timestamps from a "live source" and rebase onto the current timebase:
25762 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25766 Generate timestamps by counting samples:
25775 Force color range for the output video frame.
25777 The @code{setrange} filter marks the color range property for the
25778 output frames. It does not change the input frame, but only sets the
25779 corresponding property, which affects how the frame is treated by
25782 The filter accepts the following options:
25787 Available values are:
25791 Keep the same color range property.
25793 @item unspecified, unknown
25794 Set the color range as unspecified.
25796 @item limited, tv, mpeg
25797 Set the color range as limited.
25799 @item full, pc, jpeg
25800 Set the color range as full.
25804 @section settb, asettb
25806 Set the timebase to use for the output frames timestamps.
25807 It is mainly useful for testing timebase configuration.
25809 It accepts the following parameters:
25814 The expression which is evaluated into the output timebase.
25818 The value for @option{tb} is an arithmetic expression representing a
25819 rational. The expression can contain the constants "AVTB" (the default
25820 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25821 audio only). Default value is "intb".
25823 @subsection Examples
25827 Set the timebase to 1/25:
25833 Set the timebase to 1/10:
25839 Set the timebase to 1001/1000:
25845 Set the timebase to 2*intb:
25851 Set the default timebase value:
25858 Convert input audio to a video output representing frequency spectrum
25859 logarithmically using Brown-Puckette constant Q transform algorithm with
25860 direct frequency domain coefficient calculation (but the transform itself
25861 is not really constant Q, instead the Q factor is actually variable/clamped),
25862 with musical tone scale, from E0 to D#10.
25864 The filter accepts the following options:
25868 Specify the video size for the output. It must be even. For the syntax of this option,
25869 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25870 Default value is @code{1920x1080}.
25873 Set the output frame rate. Default value is @code{25}.
25876 Set the bargraph height. It must be even. Default value is @code{-1} which
25877 computes the bargraph height automatically.
25880 Set the axis height. It must be even. Default value is @code{-1} which computes
25881 the axis height automatically.
25884 Set the sonogram height. It must be even. Default value is @code{-1} which
25885 computes the sonogram height automatically.
25888 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25889 instead. Default value is @code{1}.
25891 @item sono_v, volume
25892 Specify the sonogram volume expression. It can contain variables:
25895 the @var{bar_v} evaluated expression
25896 @item frequency, freq, f
25897 the frequency where it is evaluated
25898 @item timeclamp, tc
25899 the value of @var{timeclamp} option
25903 @item a_weighting(f)
25904 A-weighting of equal loudness
25905 @item b_weighting(f)
25906 B-weighting of equal loudness
25907 @item c_weighting(f)
25908 C-weighting of equal loudness.
25910 Default value is @code{16}.
25912 @item bar_v, volume2
25913 Specify the bargraph volume expression. It can contain variables:
25916 the @var{sono_v} evaluated expression
25917 @item frequency, freq, f
25918 the frequency where it is evaluated
25919 @item timeclamp, tc
25920 the value of @var{timeclamp} option
25924 @item a_weighting(f)
25925 A-weighting of equal loudness
25926 @item b_weighting(f)
25927 B-weighting of equal loudness
25928 @item c_weighting(f)
25929 C-weighting of equal loudness.
25931 Default value is @code{sono_v}.
25933 @item sono_g, gamma
25934 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25935 higher gamma makes the spectrum having more range. Default value is @code{3}.
25936 Acceptable range is @code{[1, 7]}.
25938 @item bar_g, gamma2
25939 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25943 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25944 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25946 @item timeclamp, tc
25947 Specify the transform timeclamp. At low frequency, there is trade-off between
25948 accuracy in time domain and frequency domain. If timeclamp is lower,
25949 event in time domain is represented more accurately (such as fast bass drum),
25950 otherwise event in frequency domain is represented more accurately
25951 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25954 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25955 limits future samples by applying asymmetric windowing in time domain, useful
25956 when low latency is required. Accepted range is @code{[0, 1]}.
25959 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25960 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25963 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25964 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25967 This option is deprecated and ignored.
25970 Specify the transform length in time domain. Use this option to control accuracy
25971 trade-off between time domain and frequency domain at every frequency sample.
25972 It can contain variables:
25974 @item frequency, freq, f
25975 the frequency where it is evaluated
25976 @item timeclamp, tc
25977 the value of @var{timeclamp} option.
25979 Default value is @code{384*tc/(384+tc*f)}.
25982 Specify the transform count for every video frame. Default value is @code{6}.
25983 Acceptable range is @code{[1, 30]}.
25986 Specify the transform count for every single pixel. Default value is @code{0},
25987 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25990 Specify font file for use with freetype to draw the axis. If not specified,
25991 use embedded font. Note that drawing with font file or embedded font is not
25992 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25996 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25997 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
26001 Specify font color expression. This is arithmetic expression that should return
26002 integer value 0xRRGGBB. It can contain variables:
26004 @item frequency, freq, f
26005 the frequency where it is evaluated
26006 @item timeclamp, tc
26007 the value of @var{timeclamp} option
26012 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
26013 @item r(x), g(x), b(x)
26014 red, green, and blue value of intensity x.
26016 Default value is @code{st(0, (midi(f)-59.5)/12);
26017 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
26018 r(1-ld(1)) + b(ld(1))}.
26021 Specify image file to draw the axis. This option override @var{fontfile} and
26022 @var{fontcolor} option.
26025 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
26026 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
26027 Default value is @code{1}.
26030 Set colorspace. The accepted values are:
26033 Unspecified (default)
26042 BT.470BG or BT.601-6 625
26045 SMPTE-170M or BT.601-6 525
26051 BT.2020 with non-constant luminance
26056 Set spectrogram color scheme. This is list of floating point values with format
26057 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
26058 The default is @code{1|0.5|0|0|0.5|1}.
26062 @subsection Examples
26066 Playing audio while showing the spectrum:
26068 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
26072 Same as above, but with frame rate 30 fps:
26074 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
26078 Playing at 1280x720:
26080 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
26084 Disable sonogram display:
26090 A1 and its harmonics: A1, A2, (near)E3, A3:
26092 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),
26093 asplit[a][out1]; [a] showcqt [out0]'
26097 Same as above, but with more accuracy in frequency domain:
26099 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),
26100 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
26106 bar_v=10:sono_v=bar_v*a_weighting(f)
26110 Custom gamma, now spectrum is linear to the amplitude.
26116 Custom tlength equation:
26118 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)))'
26122 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
26124 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
26128 Custom font using fontconfig:
26130 font='Courier New,Monospace,mono|bold'
26134 Custom frequency range with custom axis using image file:
26136 axisfile=myaxis.png:basefreq=40:endfreq=10000
26142 Convert input audio to video output representing the audio power spectrum.
26143 Audio amplitude is on Y-axis while frequency is on X-axis.
26145 The filter accepts the following options:
26149 Specify size of video. For the syntax of this option, check the
26150 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26151 Default is @code{1024x512}.
26155 This set how each frequency bin will be represented.
26157 It accepts the following values:
26163 Default is @code{bar}.
26166 Set amplitude scale.
26168 It accepts the following values:
26182 Default is @code{log}.
26185 Set frequency scale.
26187 It accepts the following values:
26196 Reverse logarithmic scale.
26198 Default is @code{lin}.
26201 Set window size. Allowed range is from 16 to 65536.
26203 Default is @code{2048}
26206 Set windowing function.
26208 It accepts the following values:
26231 Default is @code{hanning}.
26234 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26235 which means optimal overlap for selected window function will be picked.
26238 Set time averaging. Setting this to 0 will display current maximal peaks.
26239 Default is @code{1}, which means time averaging is disabled.
26242 Specify list of colors separated by space or by '|' which will be used to
26243 draw channel frequencies. Unrecognized or missing colors will be replaced
26247 Set channel display mode.
26249 It accepts the following values:
26254 Default is @code{combined}.
26257 Set minimum amplitude used in @code{log} amplitude scaler.
26260 Set data display mode.
26262 It accepts the following values:
26268 Default is @code{magnitude}.
26271 @section showspatial
26273 Convert stereo input audio to a video output, representing the spatial relationship
26274 between two channels.
26276 The filter accepts the following options:
26280 Specify the video size for the output. For the syntax of this option, check the
26281 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26282 Default value is @code{512x512}.
26285 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
26288 Set window function.
26290 It accepts the following values:
26315 Default value is @code{hann}.
26318 Set ratio of overlap window. Default value is @code{0.5}.
26319 When value is @code{1} overlap is set to recommended size for specific
26320 window function currently used.
26323 @anchor{showspectrum}
26324 @section showspectrum
26326 Convert input audio to a video output, representing the audio frequency
26329 The filter accepts the following options:
26333 Specify the video size for the output. For the syntax of this option, check the
26334 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26335 Default value is @code{640x512}.
26338 Specify how the spectrum should slide along the window.
26340 It accepts the following values:
26343 the samples start again on the left when they reach the right
26345 the samples scroll from right to left
26347 frames are only produced when the samples reach the right
26349 the samples scroll from left to right
26352 Default value is @code{replace}.
26355 Specify display mode.
26357 It accepts the following values:
26360 all channels are displayed in the same row
26362 all channels are displayed in separate rows
26365 Default value is @samp{combined}.
26368 Specify display color mode.
26370 It accepts the following values:
26373 each channel is displayed in a separate color
26375 each channel is displayed using the same color scheme
26377 each channel is displayed using the rainbow color scheme
26379 each channel is displayed using the moreland color scheme
26381 each channel is displayed using the nebulae color scheme
26383 each channel is displayed using the fire color scheme
26385 each channel is displayed using the fiery color scheme
26387 each channel is displayed using the fruit color scheme
26389 each channel is displayed using the cool color scheme
26391 each channel is displayed using the magma color scheme
26393 each channel is displayed using the green color scheme
26395 each channel is displayed using the viridis color scheme
26397 each channel is displayed using the plasma color scheme
26399 each channel is displayed using the cividis color scheme
26401 each channel is displayed using the terrain color scheme
26404 Default value is @samp{channel}.
26407 Specify scale used for calculating intensity color values.
26409 It accepts the following values:
26414 square root, default
26425 Default value is @samp{sqrt}.
26428 Specify frequency scale.
26430 It accepts the following values:
26438 Default value is @samp{lin}.
26441 Set saturation modifier for displayed colors. Negative values provide
26442 alternative color scheme. @code{0} is no saturation at all.
26443 Saturation must be in [-10.0, 10.0] range.
26444 Default value is @code{1}.
26447 Set window function.
26449 It accepts the following values:
26474 Default value is @code{hann}.
26477 Set orientation of time vs frequency axis. Can be @code{vertical} or
26478 @code{horizontal}. Default is @code{vertical}.
26481 Set ratio of overlap window. Default value is @code{0}.
26482 When value is @code{1} overlap is set to recommended size for specific
26483 window function currently used.
26486 Set scale gain for calculating intensity color values.
26487 Default value is @code{1}.
26490 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
26493 Set color rotation, must be in [-1.0, 1.0] range.
26494 Default value is @code{0}.
26497 Set start frequency from which to display spectrogram. Default is @code{0}.
26500 Set stop frequency to which to display spectrogram. Default is @code{0}.
26503 Set upper frame rate limit. Default is @code{auto}, unlimited.
26506 Draw time and frequency axes and legends. Default is disabled.
26509 The usage is very similar to the showwaves filter; see the examples in that
26512 @subsection Examples
26516 Large window with logarithmic color scaling:
26518 showspectrum=s=1280x480:scale=log
26522 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
26524 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
26525 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
26529 @section showspectrumpic
26531 Convert input audio to a single video frame, representing the audio frequency
26534 The filter accepts the following options:
26538 Specify the video size for the output. For the syntax of this option, check the
26539 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26540 Default value is @code{4096x2048}.
26543 Specify display mode.
26545 It accepts the following values:
26548 all channels are displayed in the same row
26550 all channels are displayed in separate rows
26552 Default value is @samp{combined}.
26555 Specify display color mode.
26557 It accepts the following values:
26560 each channel is displayed in a separate color
26562 each channel is displayed using the same color scheme
26564 each channel is displayed using the rainbow color scheme
26566 each channel is displayed using the moreland color scheme
26568 each channel is displayed using the nebulae color scheme
26570 each channel is displayed using the fire color scheme
26572 each channel is displayed using the fiery color scheme
26574 each channel is displayed using the fruit color scheme
26576 each channel is displayed using the cool color scheme
26578 each channel is displayed using the magma color scheme
26580 each channel is displayed using the green color scheme
26582 each channel is displayed using the viridis color scheme
26584 each channel is displayed using the plasma color scheme
26586 each channel is displayed using the cividis color scheme
26588 each channel is displayed using the terrain color scheme
26590 Default value is @samp{intensity}.
26593 Specify scale used for calculating intensity color values.
26595 It accepts the following values:
26600 square root, default
26610 Default value is @samp{log}.
26613 Specify frequency scale.
26615 It accepts the following values:
26623 Default value is @samp{lin}.
26626 Set saturation modifier for displayed colors. Negative values provide
26627 alternative color scheme. @code{0} is no saturation at all.
26628 Saturation must be in [-10.0, 10.0] range.
26629 Default value is @code{1}.
26632 Set window function.
26634 It accepts the following values:
26658 Default value is @code{hann}.
26661 Set orientation of time vs frequency axis. Can be @code{vertical} or
26662 @code{horizontal}. Default is @code{vertical}.
26665 Set scale gain for calculating intensity color values.
26666 Default value is @code{1}.
26669 Draw time and frequency axes and legends. Default is enabled.
26672 Set color rotation, must be in [-1.0, 1.0] range.
26673 Default value is @code{0}.
26676 Set start frequency from which to display spectrogram. Default is @code{0}.
26679 Set stop frequency to which to display spectrogram. Default is @code{0}.
26682 @subsection Examples
26686 Extract an audio spectrogram of a whole audio track
26687 in a 1024x1024 picture using @command{ffmpeg}:
26689 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26693 @section showvolume
26695 Convert input audio volume to a video output.
26697 The filter accepts the following options:
26704 Set border width, allowed range is [0, 5]. Default is 1.
26707 Set channel width, allowed range is [80, 8192]. Default is 400.
26710 Set channel height, allowed range is [1, 900]. Default is 20.
26713 Set fade, allowed range is [0, 1]. Default is 0.95.
26716 Set volume color expression.
26718 The expression can use the following variables:
26722 Current max volume of channel in dB.
26728 Current channel number, starting from 0.
26732 If set, displays channel names. Default is enabled.
26735 If set, displays volume values. Default is enabled.
26738 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26739 default is @code{h}.
26742 Set step size, allowed range is [0, 5]. Default is 0, which means
26746 Set background opacity, allowed range is [0, 1]. Default is 0.
26749 Set metering mode, can be peak: @code{p} or rms: @code{r},
26750 default is @code{p}.
26753 Set display scale, can be linear: @code{lin} or log: @code{log},
26754 default is @code{lin}.
26758 If set to > 0., display a line for the max level
26759 in the previous seconds.
26760 default is disabled: @code{0.}
26763 The color of the max line. Use when @code{dm} option is set to > 0.
26764 default is: @code{orange}
26769 Convert input audio to a video output, representing the samples waves.
26771 The filter accepts the following options:
26775 Specify the video size for the output. For the syntax of this option, check the
26776 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26777 Default value is @code{600x240}.
26782 Available values are:
26785 Draw a point for each sample.
26788 Draw a vertical line for each sample.
26791 Draw a point for each sample and a line between them.
26794 Draw a centered vertical line for each sample.
26797 Default value is @code{point}.
26800 Set the number of samples which are printed on the same column. A
26801 larger value will decrease the frame rate. Must be a positive
26802 integer. This option can be set only if the value for @var{rate}
26803 is not explicitly specified.
26806 Set the (approximate) output frame rate. This is done by setting the
26807 option @var{n}. Default value is "25".
26809 @item split_channels
26810 Set if channels should be drawn separately or overlap. Default value is 0.
26813 Set colors separated by '|' which are going to be used for drawing of each channel.
26816 Set amplitude scale.
26818 Available values are:
26836 Set the draw mode. This is mostly useful to set for high @var{n}.
26838 Available values are:
26841 Scale pixel values for each drawn sample.
26844 Draw every sample directly.
26847 Default value is @code{scale}.
26850 @subsection Examples
26854 Output the input file audio and the corresponding video representation
26857 amovie=a.mp3,asplit[out0],showwaves[out1]
26861 Create a synthetic signal and show it with showwaves, forcing a
26862 frame rate of 30 frames per second:
26864 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26868 @section showwavespic
26870 Convert input audio to a single video frame, representing the samples waves.
26872 The filter accepts the following options:
26876 Specify the video size for the output. For the syntax of this option, check the
26877 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26878 Default value is @code{600x240}.
26880 @item split_channels
26881 Set if channels should be drawn separately or overlap. Default value is 0.
26884 Set colors separated by '|' which are going to be used for drawing of each channel.
26887 Set amplitude scale.
26889 Available values are:
26909 Available values are:
26912 Scale pixel values for each drawn sample.
26915 Draw every sample directly.
26918 Default value is @code{scale}.
26921 Set the filter mode.
26923 Available values are:
26926 Use average samples values for each drawn sample.
26929 Use peak samples values for each drawn sample.
26932 Default value is @code{average}.
26935 @subsection Examples
26939 Extract a channel split representation of the wave form of a whole audio track
26940 in a 1024x800 picture using @command{ffmpeg}:
26942 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26946 @section sidedata, asidedata
26948 Delete frame side data, or select frames based on it.
26950 This filter accepts the following options:
26954 Set mode of operation of the filter.
26956 Can be one of the following:
26960 Select every frame with side data of @code{type}.
26963 Delete side data of @code{type}. If @code{type} is not set, delete all side
26969 Set side data type used with all modes. Must be set for @code{select} mode. For
26970 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26971 in @file{libavutil/frame.h}. For example, to choose
26972 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26976 @section spectrumsynth
26978 Synthesize audio from 2 input video spectrums, first input stream represents
26979 magnitude across time and second represents phase across time.
26980 The filter will transform from frequency domain as displayed in videos back
26981 to time domain as presented in audio output.
26983 This filter is primarily created for reversing processed @ref{showspectrum}
26984 filter outputs, but can synthesize sound from other spectrograms too.
26985 But in such case results are going to be poor if the phase data is not
26986 available, because in such cases phase data need to be recreated, usually
26987 it's just recreated from random noise.
26988 For best results use gray only output (@code{channel} color mode in
26989 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26990 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26991 @code{data} option. Inputs videos should generally use @code{fullframe}
26992 slide mode as that saves resources needed for decoding video.
26994 The filter accepts the following options:
26998 Specify sample rate of output audio, the sample rate of audio from which
26999 spectrum was generated may differ.
27002 Set number of channels represented in input video spectrums.
27005 Set scale which was used when generating magnitude input spectrum.
27006 Can be @code{lin} or @code{log}. Default is @code{log}.
27009 Set slide which was used when generating inputs spectrums.
27010 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
27011 Default is @code{fullframe}.
27014 Set window function used for resynthesis.
27017 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
27018 which means optimal overlap for selected window function will be picked.
27021 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
27022 Default is @code{vertical}.
27025 @subsection Examples
27029 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
27030 then resynthesize videos back to audio with spectrumsynth:
27032 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
27033 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
27034 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
27038 @section split, asplit
27040 Split input into several identical outputs.
27042 @code{asplit} works with audio input, @code{split} with video.
27044 The filter accepts a single parameter which specifies the number of outputs. If
27045 unspecified, it defaults to 2.
27047 @subsection Examples
27051 Create two separate outputs from the same input:
27053 [in] split [out0][out1]
27057 To create 3 or more outputs, you need to specify the number of
27060 [in] asplit=3 [out0][out1][out2]
27064 Create two separate outputs from the same input, one cropped and
27067 [in] split [splitout1][splitout2];
27068 [splitout1] crop=100:100:0:0 [cropout];
27069 [splitout2] pad=200:200:100:100 [padout];
27073 Create 5 copies of the input audio with @command{ffmpeg}:
27075 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
27081 Receive commands sent through a libzmq client, and forward them to
27082 filters in the filtergraph.
27084 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
27085 must be inserted between two video filters, @code{azmq} between two
27086 audio filters. Both are capable to send messages to any filter type.
27088 To enable these filters you need to install the libzmq library and
27089 headers and configure FFmpeg with @code{--enable-libzmq}.
27091 For more information about libzmq see:
27092 @url{http://www.zeromq.org/}
27094 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
27095 receives messages sent through a network interface defined by the
27096 @option{bind_address} (or the abbreviation "@option{b}") option.
27097 Default value of this option is @file{tcp://localhost:5555}. You may
27098 want to alter this value to your needs, but do not forget to escape any
27099 ':' signs (see @ref{filtergraph escaping}).
27101 The received message must be in the form:
27103 @var{TARGET} @var{COMMAND} [@var{ARG}]
27106 @var{TARGET} specifies the target of the command, usually the name of
27107 the filter class or a specific filter instance name. The default
27108 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
27109 but you can override this by using the @samp{filter_name@@id} syntax
27110 (see @ref{Filtergraph syntax}).
27112 @var{COMMAND} specifies the name of the command for the target filter.
27114 @var{ARG} is optional and specifies the optional argument list for the
27115 given @var{COMMAND}.
27117 Upon reception, the message is processed and the corresponding command
27118 is injected into the filtergraph. Depending on the result, the filter
27119 will send a reply to the client, adopting the format:
27121 @var{ERROR_CODE} @var{ERROR_REASON}
27125 @var{MESSAGE} is optional.
27127 @subsection Examples
27129 Look at @file{tools/zmqsend} for an example of a zmq client which can
27130 be used to send commands processed by these filters.
27132 Consider the following filtergraph generated by @command{ffplay}.
27133 In this example the last overlay filter has an instance name. All other
27134 filters will have default instance names.
27137 ffplay -dumpgraph 1 -f lavfi "
27138 color=s=100x100:c=red [l];
27139 color=s=100x100:c=blue [r];
27140 nullsrc=s=200x100, zmq [bg];
27141 [bg][l] overlay [bg+l];
27142 [bg+l][r] overlay@@my=x=100 "
27145 To change the color of the left side of the video, the following
27146 command can be used:
27148 echo Parsed_color_0 c yellow | tools/zmqsend
27151 To change the right side:
27153 echo Parsed_color_1 c pink | tools/zmqsend
27156 To change the position of the right side:
27158 echo overlay@@my x 150 | tools/zmqsend
27162 @c man end MULTIMEDIA FILTERS
27164 @chapter Multimedia Sources
27165 @c man begin MULTIMEDIA SOURCES
27167 Below is a description of the currently available multimedia sources.
27171 This is the same as @ref{movie} source, except it selects an audio
27177 Read audio and/or video stream(s) from a movie container.
27179 It accepts the following parameters:
27183 The name of the resource to read (not necessarily a file; it can also be a
27184 device or a stream accessed through some protocol).
27186 @item format_name, f
27187 Specifies the format assumed for the movie to read, and can be either
27188 the name of a container or an input device. If not specified, the
27189 format is guessed from @var{movie_name} or by probing.
27191 @item seek_point, sp
27192 Specifies the seek point in seconds. The frames will be output
27193 starting from this seek point. The parameter is evaluated with
27194 @code{av_strtod}, so the numerical value may be suffixed by an IS
27195 postfix. The default value is "0".
27198 Specifies the streams to read. Several streams can be specified,
27199 separated by "+". The source will then have as many outputs, in the
27200 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
27201 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
27202 respectively the default (best suited) video and audio stream. Default
27203 is "dv", or "da" if the filter is called as "amovie".
27205 @item stream_index, si
27206 Specifies the index of the video stream to read. If the value is -1,
27207 the most suitable video stream will be automatically selected. The default
27208 value is "-1". Deprecated. If the filter is called "amovie", it will select
27209 audio instead of video.
27212 Specifies how many times to read the stream in sequence.
27213 If the value is 0, the stream will be looped infinitely.
27214 Default value is "1".
27216 Note that when the movie is looped the source timestamps are not
27217 changed, so it will generate non monotonically increasing timestamps.
27219 @item discontinuity
27220 Specifies the time difference between frames above which the point is
27221 considered a timestamp discontinuity which is removed by adjusting the later
27225 It allows overlaying a second video on top of the main input of
27226 a filtergraph, as shown in this graph:
27228 input -----------> deltapts0 --> overlay --> output
27231 movie --> scale--> deltapts1 -------+
27233 @subsection Examples
27237 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
27238 on top of the input labelled "in":
27240 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
27241 [in] setpts=PTS-STARTPTS [main];
27242 [main][over] overlay=16:16 [out]
27246 Read from a video4linux2 device, and overlay it on top of the input
27249 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
27250 [in] setpts=PTS-STARTPTS [main];
27251 [main][over] overlay=16:16 [out]
27255 Read the first video stream and the audio stream with id 0x81 from
27256 dvd.vob; the video is connected to the pad named "video" and the audio is
27257 connected to the pad named "audio":
27259 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
27263 @subsection Commands
27265 Both movie and amovie support the following commands:
27268 Perform seek using "av_seek_frame".
27269 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
27272 @var{stream_index}: If stream_index is -1, a default
27273 stream is selected, and @var{timestamp} is automatically converted
27274 from AV_TIME_BASE units to the stream specific time_base.
27276 @var{timestamp}: Timestamp in AVStream.time_base units
27277 or, if no stream is specified, in AV_TIME_BASE units.
27279 @var{flags}: Flags which select direction and seeking mode.
27283 Get movie duration in AV_TIME_BASE units.
27287 @c man end MULTIMEDIA SOURCES