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 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Commands
1091 This filter supports the all above options as @ref{commands}.
1093 @subsection Examples
1097 Fade in first 15 seconds of audio:
1099 afade=t=in:ss=0:d=15
1103 Fade out last 25 seconds of a 900 seconds audio:
1105 afade=t=out:st=875:d=25
1110 Denoise audio samples with FFT.
1112 A description of the accepted parameters follows.
1116 Set the noise reduction in dB, allowed range is 0.01 to 97.
1117 Default value is 12 dB.
1120 Set the noise floor in dB, allowed range is -80 to -20.
1121 Default value is -50 dB.
1126 It accepts the following values:
1135 Select shellac noise.
1138 Select custom noise, defined in @code{bn} option.
1140 Default value is white noise.
1144 Set custom band noise for every one of 15 bands.
1145 Bands are separated by ' ' or '|'.
1148 Set the residual floor in dB, allowed range is -80 to -20.
1149 Default value is -38 dB.
1152 Enable noise tracking. By default is disabled.
1153 With this enabled, noise floor is automatically adjusted.
1156 Enable residual tracking. By default is disabled.
1159 Set the output mode.
1161 It accepts the following values:
1164 Pass input unchanged.
1167 Pass noise filtered out.
1172 Default value is @var{o}.
1176 @subsection Commands
1178 This filter supports the following commands:
1180 @item sample_noise, sn
1181 Start or stop measuring noise profile.
1182 Syntax for the command is : "start" or "stop" string.
1183 After measuring noise profile is stopped it will be
1184 automatically applied in filtering.
1186 @item noise_reduction, nr
1187 Change noise reduction. Argument is single float number.
1188 Syntax for the command is : "@var{noise_reduction}"
1190 @item noise_floor, nf
1191 Change noise floor. Argument is single float number.
1192 Syntax for the command is : "@var{noise_floor}"
1194 @item output_mode, om
1195 Change output mode operation.
1196 Syntax for the command is : "i", "o" or "n" string.
1200 Apply arbitrary expressions to samples in frequency domain.
1204 Set frequency domain real expression for each separate channel separated
1205 by '|'. Default is "re".
1206 If the number of input channels is greater than the number of
1207 expressions, the last specified expression is used for the remaining
1211 Set frequency domain imaginary expression for each separate channel
1212 separated by '|'. Default is "im".
1214 Each expression in @var{real} and @var{imag} can contain the following
1215 constants and functions:
1222 current frequency bin number
1225 number of available bins
1228 channel number of the current expression
1237 current real part of frequency bin of current channel
1240 current imaginary part of frequency bin of current channel
1243 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1246 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1250 Set window size. Allowed range is from 16 to 131072.
1251 Default is @code{4096}
1254 Set window function. Default is @code{hann}.
1257 Set window overlap. If set to 1, the recommended overlap for selected
1258 window function will be picked. Default is @code{0.75}.
1261 @subsection Examples
1265 Leave almost only low frequencies in audio:
1267 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1271 Apply robotize effect:
1273 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1277 Apply whisper effect:
1279 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"
1286 Apply an arbitrary Finite Impulse Response filter.
1288 This filter is designed for applying long FIR filters,
1289 up to 60 seconds long.
1291 It can be used as component for digital crossover filters,
1292 room equalization, cross talk cancellation, wavefield synthesis,
1293 auralization, ambiophonics, ambisonics and spatialization.
1295 This filter uses the streams higher than first one as FIR coefficients.
1296 If the non-first stream holds a single channel, it will be used
1297 for all input channels in the first stream, otherwise
1298 the number of channels in the non-first stream must be same as
1299 the number of channels in the first stream.
1301 It accepts the following parameters:
1305 Set dry gain. This sets input gain.
1308 Set wet gain. This sets final output gain.
1311 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1314 Enable applying gain measured from power of IR.
1316 Set which approach to use for auto gain measurement.
1320 Do not apply any gain.
1323 select peak gain, very conservative approach. This is default value.
1326 select DC gain, limited application.
1329 select gain to noise approach, this is most popular one.
1333 Set gain to be applied to IR coefficients before filtering.
1334 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1337 Set format of IR stream. Can be @code{mono} or @code{input}.
1338 Default is @code{input}.
1341 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1342 Allowed range is 0.1 to 60 seconds.
1345 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1346 By default it is disabled.
1349 Set for which IR channel to display frequency response. By default is first channel
1350 displayed. This option is used only when @var{response} is enabled.
1353 Set video stream size. This option is used only when @var{response} is enabled.
1356 Set video stream frame rate. This option is used only when @var{response} is enabled.
1359 Set minimal partition size used for convolution. Default is @var{8192}.
1360 Allowed range is from @var{1} to @var{32768}.
1361 Lower values decreases latency at cost of higher CPU usage.
1364 Set maximal partition size used for convolution. Default is @var{8192}.
1365 Allowed range is from @var{8} to @var{32768}.
1366 Lower values may increase CPU usage.
1369 Set number of input impulse responses streams which will be switchable at runtime.
1370 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1373 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1374 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1375 This option can be changed at runtime via @ref{commands}.
1378 @subsection Examples
1382 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1384 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1391 Set output format constraints for the input audio. The framework will
1392 negotiate the most appropriate format to minimize conversions.
1394 It accepts the following parameters:
1397 @item sample_fmts, f
1398 A '|'-separated list of requested sample formats.
1400 @item sample_rates, r
1401 A '|'-separated list of requested sample rates.
1403 @item channel_layouts, cl
1404 A '|'-separated list of requested channel layouts.
1406 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1407 for the required syntax.
1410 If a parameter is omitted, all values are allowed.
1412 Force the output to either unsigned 8-bit or signed 16-bit stereo
1414 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1418 Apply frequency shift to input audio samples.
1420 The filter accepts the following options:
1424 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1425 Default value is 0.0.
1428 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1429 Default value is 1.0.
1432 @subsection Commands
1434 This filter supports the all above options as @ref{commands}.
1438 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1439 processing reduces disturbing noise between useful signals.
1441 Gating is done by detecting the volume below a chosen level @var{threshold}
1442 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1443 floor is set via @var{range}. Because an exact manipulation of the signal
1444 would cause distortion of the waveform the reduction can be levelled over
1445 time. This is done by setting @var{attack} and @var{release}.
1447 @var{attack} determines how long the signal has to fall below the threshold
1448 before any reduction will occur and @var{release} sets the time the signal
1449 has to rise above the threshold to reduce the reduction again.
1450 Shorter signals than the chosen attack time will be left untouched.
1454 Set input level before filtering.
1455 Default is 1. Allowed range is from 0.015625 to 64.
1458 Set the mode of operation. Can be @code{upward} or @code{downward}.
1459 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1460 will be amplified, expanding dynamic range in upward direction.
1461 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1464 Set the level of gain reduction when the signal is below the threshold.
1465 Default is 0.06125. Allowed range is from 0 to 1.
1466 Setting this to 0 disables reduction and then filter behaves like expander.
1469 If a signal rises above this level the gain reduction is released.
1470 Default is 0.125. Allowed range is from 0 to 1.
1473 Set a ratio by which the signal is reduced.
1474 Default is 2. Allowed range is from 1 to 9000.
1477 Amount of milliseconds the signal has to rise above the threshold before gain
1479 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1482 Amount of milliseconds the signal has to fall below the threshold before the
1483 reduction is increased again. Default is 250 milliseconds.
1484 Allowed range is from 0.01 to 9000.
1487 Set amount of amplification of signal after processing.
1488 Default is 1. Allowed range is from 1 to 64.
1491 Curve the sharp knee around the threshold to enter gain reduction more softly.
1492 Default is 2.828427125. Allowed range is from 1 to 8.
1495 Choose if exact signal should be taken for detection or an RMS like one.
1496 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1499 Choose if the average level between all channels or the louder channel affects
1501 Default is @code{average}. Can be @code{average} or @code{maximum}.
1504 @subsection Commands
1506 This filter supports the all above options as @ref{commands}.
1510 Apply an arbitrary Infinite Impulse Response filter.
1512 It accepts the following parameters:
1516 Set B/numerator/zeros/reflection coefficients.
1519 Set A/denominator/poles/ladder coefficients.
1531 Set coefficients format.
1535 lattice-ladder function
1537 analog transfer function
1539 digital transfer function
1541 Z-plane zeros/poles, cartesian (default)
1543 Z-plane zeros/poles, polar radians
1545 Z-plane zeros/poles, polar degrees
1551 Set type of processing.
1563 Set filtering precision.
1567 double-precision floating-point (default)
1569 single-precision floating-point
1577 Normalize filter coefficients, by default is enabled.
1578 Enabling it will normalize magnitude response at DC to 0dB.
1581 How much to use filtered signal in output. Default is 1.
1582 Range is between 0 and 1.
1585 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1586 By default it is disabled.
1589 Set for which IR channel to display frequency response. By default is first channel
1590 displayed. This option is used only when @var{response} is enabled.
1593 Set video stream size. This option is used only when @var{response} is enabled.
1596 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1599 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1600 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1603 Different coefficients and gains can be provided for every channel, in such case
1604 use '|' to separate coefficients or gains. Last provided coefficients will be
1605 used for all remaining channels.
1607 @subsection Examples
1611 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1613 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
1617 Same as above but in @code{zp} format:
1619 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
1623 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1625 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1631 The limiter prevents an input signal from rising over a desired threshold.
1632 This limiter uses lookahead technology to prevent your signal from distorting.
1633 It means that there is a small delay after the signal is processed. Keep in mind
1634 that the delay it produces is the attack time you set.
1636 The filter accepts the following options:
1640 Set input gain. Default is 1.
1643 Set output gain. Default is 1.
1646 Don't let signals above this level pass the limiter. Default is 1.
1649 The limiter will reach its attenuation level in this amount of time in
1650 milliseconds. Default is 5 milliseconds.
1653 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1654 Default is 50 milliseconds.
1657 When gain reduction is always needed ASC takes care of releasing to an
1658 average reduction level rather than reaching a reduction of 0 in the release
1662 Select how much the release time is affected by ASC, 0 means nearly no changes
1663 in release time while 1 produces higher release times.
1666 Auto level output signal. Default is enabled.
1667 This normalizes audio back to 0dB if enabled.
1670 Depending on picked setting it is recommended to upsample input 2x or 4x times
1671 with @ref{aresample} before applying this filter.
1675 Apply a two-pole all-pass filter with central frequency (in Hz)
1676 @var{frequency}, and filter-width @var{width}.
1677 An all-pass filter changes the audio's frequency to phase relationship
1678 without changing its frequency to amplitude relationship.
1680 The filter accepts the following options:
1684 Set frequency in Hz.
1687 Set method to specify band-width of filter.
1702 Specify the band-width of a filter in width_type units.
1705 How much to use filtered signal in output. Default is 1.
1706 Range is between 0 and 1.
1709 Specify which channels to filter, by default all available are filtered.
1712 Normalize biquad coefficients, by default is disabled.
1713 Enabling it will normalize magnitude response at DC to 0dB.
1716 Set the filter order, can be 1 or 2. Default is 2.
1719 Set transform type of IIR filter.
1728 Set precison of filtering.
1731 Pick automatic sample format depending on surround filters.
1733 Always use signed 16-bit.
1735 Always use signed 32-bit.
1737 Always use float 32-bit.
1739 Always use float 64-bit.
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Change allpass frequency.
1749 Syntax for the command is : "@var{frequency}"
1752 Change allpass width_type.
1753 Syntax for the command is : "@var{width_type}"
1756 Change allpass width.
1757 Syntax for the command is : "@var{width}"
1761 Syntax for the command is : "@var{mix}"
1768 The filter accepts the following options:
1772 Set the number of loops. Setting this value to -1 will result in infinite loops.
1776 Set maximal number of samples. Default is 0.
1779 Set first sample of loop. Default is 0.
1785 Merge two or more audio streams into a single multi-channel stream.
1787 The filter accepts the following options:
1792 Set the number of inputs. Default is 2.
1796 If the channel layouts of the inputs are disjoint, and therefore compatible,
1797 the channel layout of the output will be set accordingly and the channels
1798 will be reordered as necessary. If the channel layouts of the inputs are not
1799 disjoint, the output will have all the channels of the first input then all
1800 the channels of the second input, in that order, and the channel layout of
1801 the output will be the default value corresponding to the total number of
1804 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1805 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1806 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1807 first input, b1 is the first channel of the second input).
1809 On the other hand, if both input are in stereo, the output channels will be
1810 in the default order: a1, a2, b1, b2, and the channel layout will be
1811 arbitrarily set to 4.0, which may or may not be the expected value.
1813 All inputs must have the same sample rate, and format.
1815 If inputs do not have the same duration, the output will stop with the
1818 @subsection Examples
1822 Merge two mono files into a stereo stream:
1824 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1828 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1830 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
1836 Mixes multiple audio inputs into a single output.
1838 Note that this filter only supports float samples (the @var{amerge}
1839 and @var{pan} audio filters support many formats). If the @var{amix}
1840 input has integer samples then @ref{aresample} will be automatically
1841 inserted to perform the conversion to float samples.
1845 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1847 will mix 3 input audio streams to a single output with the same duration as the
1848 first input and a dropout transition time of 3 seconds.
1850 It accepts the following parameters:
1854 The number of inputs. If unspecified, it defaults to 2.
1857 How to determine the end-of-stream.
1861 The duration of the longest input. (default)
1864 The duration of the shortest input.
1867 The duration of the first input.
1871 @item dropout_transition
1872 The transition time, in seconds, for volume renormalization when an input
1873 stream ends. The default value is 2 seconds.
1876 Specify weight of each input audio stream as sequence.
1877 Each weight is separated by space. By default all inputs have same weight.
1880 @subsection Commands
1882 This filter supports the following commands:
1885 Syntax is same as option with same name.
1890 Multiply first audio stream with second audio stream and store result
1891 in output audio stream. Multiplication is done by multiplying each
1892 sample from first stream with sample at same position from second stream.
1894 With this element-wise multiplication one can create amplitude fades and
1895 amplitude modulations.
1897 @section anequalizer
1899 High-order parametric multiband equalizer for each channel.
1901 It accepts the following parameters:
1905 This option string is in format:
1906 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1907 Each equalizer band is separated by '|'.
1911 Set channel number to which equalization will be applied.
1912 If input doesn't have that channel the entry is ignored.
1915 Set central frequency for band.
1916 If input doesn't have that frequency the entry is ignored.
1919 Set band width in Hertz.
1922 Set band gain in dB.
1925 Set filter type for band, optional, can be:
1929 Butterworth, this is default.
1940 With this option activated frequency response of anequalizer is displayed
1944 Set video stream size. Only useful if curves option is activated.
1947 Set max gain that will be displayed. Only useful if curves option is activated.
1948 Setting this to a reasonable value makes it possible to display gain which is derived from
1949 neighbour bands which are too close to each other and thus produce higher gain
1950 when both are activated.
1953 Set frequency scale used to draw frequency response in video output.
1954 Can be linear or logarithmic. Default is logarithmic.
1957 Set color for each channel curve which is going to be displayed in video stream.
1958 This is list of color names separated by space or by '|'.
1959 Unrecognised or missing colors will be replaced by white color.
1962 @subsection Examples
1966 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1967 for first 2 channels using Chebyshev type 1 filter:
1969 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1973 @subsection Commands
1975 This filter supports the following commands:
1978 Alter existing filter parameters.
1979 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1981 @var{fN} is existing filter number, starting from 0, if no such filter is available
1983 @var{freq} set new frequency parameter.
1984 @var{width} set new width parameter in Hertz.
1985 @var{gain} set new gain parameter in dB.
1987 Full filter invocation with asendcmd may look like this:
1988 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1993 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1995 Each sample is adjusted by looking for other samples with similar contexts. This
1996 context similarity is defined by comparing their surrounding patches of size
1997 @option{p}. Patches are searched in an area of @option{r} around the sample.
1999 The filter accepts the following options:
2003 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2006 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2007 Default value is 2 milliseconds.
2010 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2011 Default value is 6 milliseconds.
2014 Set the output mode.
2016 It accepts the following values:
2019 Pass input unchanged.
2022 Pass noise filtered out.
2027 Default value is @var{o}.
2031 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2034 @subsection Commands
2036 This filter supports the all above options as @ref{commands}.
2039 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2041 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2042 relate to producing the least mean square of the error signal (difference between the desired,
2043 2nd input audio stream and the actual signal, the 1st input audio stream).
2045 A description of the accepted options follows.
2058 Set the filter leakage.
2061 It accepts the following values:
2070 Pass filtered samples.
2073 Pass difference between desired and filtered samples.
2075 Default value is @var{o}.
2079 @subsection Examples
2083 One of many usages of this filter is noise reduction, input audio is filtered
2084 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2086 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2090 @subsection Commands
2092 This filter supports the same commands as options, excluding option @code{order}.
2096 Pass the audio source unchanged to the output.
2100 Pad the end of an audio stream with silence.
2102 This can be used together with @command{ffmpeg} @option{-shortest} to
2103 extend audio streams to the same length as the video stream.
2105 A description of the accepted options follows.
2109 Set silence packet size. Default value is 4096.
2112 Set the number of samples of silence to add to the end. After the
2113 value is reached, the stream is terminated. This option is mutually
2114 exclusive with @option{whole_len}.
2117 Set the minimum total number of samples in the output audio stream. If
2118 the value is longer than the input audio length, silence is added to
2119 the end, until the value is reached. This option is mutually exclusive
2120 with @option{pad_len}.
2123 Specify the duration of samples of silence to add. See
2124 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2125 for the accepted syntax. Used only if set to non-zero value.
2128 Specify the minimum total duration in the output audio stream. See
2129 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2130 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2131 the input audio length, silence is added to the end, until the value is reached.
2132 This option is mutually exclusive with @option{pad_dur}
2135 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2136 nor @option{whole_dur} option is set, the filter will add silence to the end of
2137 the input stream indefinitely.
2139 @subsection Examples
2143 Add 1024 samples of silence to the end of the input:
2149 Make sure the audio output will contain at least 10000 samples, pad
2150 the input with silence if required:
2152 apad=whole_len=10000
2156 Use @command{ffmpeg} to pad the audio input with silence, so that the
2157 video stream will always result the shortest and will be converted
2158 until the end in the output file when using the @option{shortest}
2161 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2166 Add a phasing effect to the input audio.
2168 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2169 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2171 A description of the accepted parameters follows.
2175 Set input gain. Default is 0.4.
2178 Set output gain. Default is 0.74
2181 Set delay in milliseconds. Default is 3.0.
2184 Set decay. Default is 0.4.
2187 Set modulation speed in Hz. Default is 0.5.
2190 Set modulation type. Default is triangular.
2192 It accepts the following values:
2199 @section aphaseshift
2200 Apply phase shift to input audio samples.
2202 The filter accepts the following options:
2206 Specify phase shift. Allowed range is from -1.0 to 1.0.
2207 Default value is 0.0.
2210 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2211 Default value is 1.0.
2214 @subsection Commands
2216 This filter supports the all above options as @ref{commands}.
2220 Audio pulsator is something between an autopanner and a tremolo.
2221 But it can produce funny stereo effects as well. Pulsator changes the volume
2222 of the left and right channel based on a LFO (low frequency oscillator) with
2223 different waveforms and shifted phases.
2224 This filter have the ability to define an offset between left and right
2225 channel. An offset of 0 means that both LFO shapes match each other.
2226 The left and right channel are altered equally - a conventional tremolo.
2227 An offset of 50% means that the shape of the right channel is exactly shifted
2228 in phase (or moved backwards about half of the frequency) - pulsator acts as
2229 an autopanner. At 1 both curves match again. Every setting in between moves the
2230 phase shift gapless between all stages and produces some "bypassing" sounds with
2231 sine and triangle waveforms. The more you set the offset near 1 (starting from
2232 the 0.5) the faster the signal passes from the left to the right speaker.
2234 The filter accepts the following options:
2238 Set input gain. By default it is 1. Range is [0.015625 - 64].
2241 Set output gain. By default it is 1. Range is [0.015625 - 64].
2244 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2245 sawup or sawdown. Default is sine.
2248 Set modulation. Define how much of original signal is affected by the LFO.
2251 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2254 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2257 Set pulse width. Default is 1. Allowed range is [0 - 2].
2260 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2263 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2267 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2271 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2272 if timing is set to hz.
2278 Resample the input audio to the specified parameters, using the
2279 libswresample library. If none are specified then the filter will
2280 automatically convert between its input and output.
2282 This filter is also able to stretch/squeeze the audio data to make it match
2283 the timestamps or to inject silence / cut out audio to make it match the
2284 timestamps, do a combination of both or do neither.
2286 The filter accepts the syntax
2287 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2288 expresses a sample rate and @var{resampler_options} is a list of
2289 @var{key}=@var{value} pairs, separated by ":". See the
2290 @ref{Resampler Options,,"Resampler Options" section in the
2291 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2292 for the complete list of supported options.
2294 @subsection Examples
2298 Resample the input audio to 44100Hz:
2304 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2305 samples per second compensation:
2307 aresample=async=1000
2313 Reverse an audio clip.
2315 Warning: This filter requires memory to buffer the entire clip, so trimming
2318 @subsection Examples
2322 Take the first 5 seconds of a clip, and reverse it.
2324 atrim=end=5,areverse
2330 Reduce noise from speech using Recurrent Neural Networks.
2332 This filter accepts the following options:
2336 Set train model file to load. This option is always required.
2339 Set how much to mix filtered samples into final output.
2340 Allowed range is from -1 to 1. Default value is 1.
2341 Negative values are special, they set how much to keep filtered noise
2342 in the final filter output. Set this option to -1 to hear actual
2343 noise removed from input signal.
2346 @section asetnsamples
2348 Set the number of samples per each output audio frame.
2350 The last output packet may contain a different number of samples, as
2351 the filter will flush all the remaining samples when the input audio
2354 The filter accepts the following options:
2358 @item nb_out_samples, n
2359 Set the number of frames per each output audio frame. The number is
2360 intended as the number of samples @emph{per each channel}.
2361 Default value is 1024.
2364 If set to 1, the filter will pad the last audio frame with zeroes, so
2365 that the last frame will contain the same number of samples as the
2366 previous ones. Default value is 1.
2369 For example, to set the number of per-frame samples to 1234 and
2370 disable padding for the last frame, use:
2372 asetnsamples=n=1234:p=0
2377 Set the sample rate without altering the PCM data.
2378 This will result in a change of speed and pitch.
2380 The filter accepts the following options:
2383 @item sample_rate, r
2384 Set the output sample rate. Default is 44100 Hz.
2389 Show a line containing various information for each input audio frame.
2390 The input audio is not modified.
2392 The shown line contains a sequence of key/value pairs of the form
2393 @var{key}:@var{value}.
2395 The following values are shown in the output:
2399 The (sequential) number of the input frame, starting from 0.
2402 The presentation timestamp of the input frame, in time base units; the time base
2403 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2406 The presentation timestamp of the input frame in seconds.
2409 position of the frame in the input stream, -1 if this information in
2410 unavailable and/or meaningless (for example in case of synthetic audio)
2419 The sample rate for the audio frame.
2422 The number of samples (per channel) in the frame.
2425 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2426 audio, the data is treated as if all the planes were concatenated.
2428 @item plane_checksums
2429 A list of Adler-32 checksums for each data plane.
2433 Apply audio soft clipping.
2435 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2436 along a smooth curve, rather than the abrupt shape of hard-clipping.
2438 This filter accepts the following options:
2442 Set type of soft-clipping.
2444 It accepts the following values:
2458 Set additional parameter which controls sigmoid function.
2461 Set oversampling factor.
2464 @subsection Commands
2466 This filter supports the all above options as @ref{commands}.
2469 Automatic Speech Recognition
2471 This filter uses PocketSphinx for speech recognition. To enable
2472 compilation of this filter, you need to configure FFmpeg with
2473 @code{--enable-pocketsphinx}.
2475 It accepts the following options:
2479 Set sampling rate of input audio. Defaults is @code{16000}.
2480 This need to match speech models, otherwise one will get poor results.
2483 Set dictionary containing acoustic model files.
2486 Set pronunciation dictionary.
2489 Set language model file.
2492 Set language model set.
2495 Set which language model to use.
2498 Set output for log messages.
2501 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2506 Display time domain statistical information about the audio channels.
2507 Statistics are calculated and displayed for each audio channel and,
2508 where applicable, an overall figure is also given.
2510 It accepts the following option:
2513 Short window length in seconds, used for peak and trough RMS measurement.
2514 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2518 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2519 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2522 Available keys for each channel are:
2568 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2569 this @code{lavfi.astats.Overall.Peak_count}.
2571 For description what each key means read below.
2574 Set number of frame after which stats are going to be recalculated.
2575 Default is disabled.
2577 @item measure_perchannel
2578 Select the entries which need to be measured per channel. The metadata keys can
2579 be used as flags, default is @option{all} which measures everything.
2580 @option{none} disables all per channel measurement.
2582 @item measure_overall
2583 Select the entries which need to be measured overall. The metadata keys can
2584 be used as flags, default is @option{all} which measures everything.
2585 @option{none} disables all overall measurement.
2589 A description of each shown parameter follows:
2593 Mean amplitude displacement from zero.
2596 Minimal sample level.
2599 Maximal sample level.
2601 @item Min difference
2602 Minimal difference between two consecutive samples.
2604 @item Max difference
2605 Maximal difference between two consecutive samples.
2607 @item Mean difference
2608 Mean difference between two consecutive samples.
2609 The average of each difference between two consecutive samples.
2611 @item RMS difference
2612 Root Mean Square difference between two consecutive samples.
2616 Standard peak and RMS level measured in dBFS.
2620 Peak and trough values for RMS level measured over a short window.
2623 Standard ratio of peak to RMS level (note: not in dB).
2626 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2627 (i.e. either @var{Min level} or @var{Max level}).
2630 Number of occasions (not the number of samples) that the signal attained either
2631 @var{Min level} or @var{Max level}.
2633 @item Noise floor dB
2634 Minimum local peak measured in dBFS over a short window.
2636 @item Noise floor count
2637 Number of occasions (not the number of samples) that the signal attained
2641 Overall bit depth of audio. Number of bits used for each sample.
2644 Measured dynamic range of audio in dB.
2646 @item Zero crossings
2647 Number of points where the waveform crosses the zero level axis.
2649 @item Zero crossings rate
2650 Rate of Zero crossings and number of audio samples.
2654 Boost subwoofer frequencies.
2656 The filter accepts the following options:
2660 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2661 Default value is 0.7.
2664 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2665 Default value is 0.7.
2668 Set delay line decay gain value. Allowed range is from 0 to 1.
2669 Default value is 0.7.
2672 Set delay line feedback gain value. Allowed range is from 0 to 1.
2673 Default value is 0.9.
2676 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2677 Default value is 100.
2680 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2681 Default value is 0.5.
2684 Set delay. Allowed range is from 1 to 100.
2685 Default value is 20.
2688 @subsection Commands
2690 This filter supports the all above options as @ref{commands}.
2693 Cut subwoofer frequencies.
2695 This filter allows to set custom, steeper
2696 roll off than highpass filter, and thus is able to more attenuate
2697 frequency content in stop-band.
2699 The filter accepts the following options:
2703 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2704 Default value is 20.
2707 Set filter order. Available values are from 3 to 20.
2708 Default value is 10.
2711 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2714 @subsection Commands
2716 This filter supports the all above options as @ref{commands}.
2719 Cut super frequencies.
2721 The filter accepts the following options:
2725 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2726 Default value is 20000.
2729 Set filter order. Available values are from 3 to 20.
2730 Default value is 10.
2733 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2736 @subsection Commands
2738 This filter supports the all above options as @ref{commands}.
2741 Apply high order Butterworth band-pass filter.
2743 The filter accepts the following options:
2747 Set center frequency in Hertz. Allowed range is 2 to 999999.
2748 Default value is 1000.
2751 Set filter order. Available values are from 4 to 20.
2755 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2758 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2761 @subsection Commands
2763 This filter supports the all above options as @ref{commands}.
2766 Apply high order Butterworth band-stop filter.
2768 The filter accepts the following options:
2772 Set center frequency in Hertz. Allowed range is 2 to 999999.
2773 Default value is 1000.
2776 Set filter order. Available values are from 4 to 20.
2780 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2783 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2786 @subsection Commands
2788 This filter supports the all above options as @ref{commands}.
2794 The filter accepts exactly one parameter, the audio tempo. If not
2795 specified then the filter will assume nominal 1.0 tempo. Tempo must
2796 be in the [0.5, 100.0] range.
2798 Note that tempo greater than 2 will skip some samples rather than
2799 blend them in. If for any reason this is a concern it is always
2800 possible to daisy-chain several instances of atempo to achieve the
2801 desired product tempo.
2803 @subsection Examples
2807 Slow down audio to 80% tempo:
2813 To speed up audio to 300% tempo:
2819 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2821 atempo=sqrt(3),atempo=sqrt(3)
2825 @subsection Commands
2827 This filter supports the following commands:
2830 Change filter tempo scale factor.
2831 Syntax for the command is : "@var{tempo}"
2836 Trim the input so that the output contains one continuous subpart of the input.
2838 It accepts the following parameters:
2841 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2842 sample with the timestamp @var{start} will be the first sample in the output.
2845 Specify time of the first audio sample that will be dropped, i.e. the
2846 audio sample immediately preceding the one with the timestamp @var{end} will be
2847 the last sample in the output.
2850 Same as @var{start}, except this option sets the start timestamp in samples
2854 Same as @var{end}, except this option sets the end timestamp in samples instead
2858 The maximum duration of the output in seconds.
2861 The number of the first sample that should be output.
2864 The number of the first sample that should be dropped.
2867 @option{start}, @option{end}, and @option{duration} are expressed as time
2868 duration specifications; see
2869 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2871 Note that the first two sets of the start/end options and the @option{duration}
2872 option look at the frame timestamp, while the _sample options simply count the
2873 samples that pass through the filter. So start/end_pts and start/end_sample will
2874 give different results when the timestamps are wrong, inexact or do not start at
2875 zero. Also note that this filter does not modify the timestamps. If you wish
2876 to have the output timestamps start at zero, insert the asetpts filter after the
2879 If multiple start or end options are set, this filter tries to be greedy and
2880 keep all samples that match at least one of the specified constraints. To keep
2881 only the part that matches all the constraints at once, chain multiple atrim
2884 The defaults are such that all the input is kept. So it is possible to set e.g.
2885 just the end values to keep everything before the specified time.
2890 Drop everything except the second minute of input:
2892 ffmpeg -i INPUT -af atrim=60:120
2896 Keep only the first 1000 samples:
2898 ffmpeg -i INPUT -af atrim=end_sample=1000
2903 @section axcorrelate
2904 Calculate normalized cross-correlation between two input audio streams.
2906 Resulted samples are always between -1 and 1 inclusive.
2907 If result is 1 it means two input samples are highly correlated in that selected segment.
2908 Result 0 means they are not correlated at all.
2909 If result is -1 it means two input samples are out of phase, which means they cancel each
2912 The filter accepts the following options:
2916 Set size of segment over which cross-correlation is calculated.
2917 Default is 256. Allowed range is from 2 to 131072.
2920 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2921 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2922 are always zero and thus need much less calculations to make.
2923 This is generally not true, but is valid for typical audio streams.
2926 @subsection Examples
2930 Calculate correlation between channels in stereo audio stream:
2932 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2938 Apply a two-pole Butterworth band-pass filter with central
2939 frequency @var{frequency}, and (3dB-point) band-width width.
2940 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2941 instead of the default: constant 0dB peak gain.
2942 The filter roll off at 6dB per octave (20dB per decade).
2944 The filter accepts the following options:
2948 Set the filter's central frequency. Default is @code{3000}.
2951 Constant skirt gain if set to 1. Defaults to 0.
2954 Set method to specify band-width of filter.
2969 Specify the band-width of a filter in width_type units.
2972 How much to use filtered signal in output. Default is 1.
2973 Range is between 0 and 1.
2976 Specify which channels to filter, by default all available are filtered.
2979 Normalize biquad coefficients, by default is disabled.
2980 Enabling it will normalize magnitude response at DC to 0dB.
2983 Set transform type of IIR filter.
2992 Set precison of filtering.
2995 Pick automatic sample format depending on surround filters.
2997 Always use signed 16-bit.
2999 Always use signed 32-bit.
3001 Always use float 32-bit.
3003 Always use float 64-bit.
3007 @subsection Commands
3009 This filter supports the following commands:
3012 Change bandpass frequency.
3013 Syntax for the command is : "@var{frequency}"
3016 Change bandpass width_type.
3017 Syntax for the command is : "@var{width_type}"
3020 Change bandpass width.
3021 Syntax for the command is : "@var{width}"
3024 Change bandpass mix.
3025 Syntax for the command is : "@var{mix}"
3030 Apply a two-pole Butterworth band-reject filter with central
3031 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3032 The filter roll off at 6dB per octave (20dB per decade).
3034 The filter accepts the following options:
3038 Set the filter's central frequency. Default is @code{3000}.
3041 Set method to specify band-width of filter.
3056 Specify the band-width of a filter in width_type units.
3059 How much to use filtered signal in output. Default is 1.
3060 Range is between 0 and 1.
3063 Specify which channels to filter, by default all available are filtered.
3066 Normalize biquad coefficients, by default is disabled.
3067 Enabling it will normalize magnitude response at DC to 0dB.
3070 Set transform type of IIR filter.
3079 Set precison of filtering.
3082 Pick automatic sample format depending on surround filters.
3084 Always use signed 16-bit.
3086 Always use signed 32-bit.
3088 Always use float 32-bit.
3090 Always use float 64-bit.
3094 @subsection Commands
3096 This filter supports the following commands:
3099 Change bandreject frequency.
3100 Syntax for the command is : "@var{frequency}"
3103 Change bandreject width_type.
3104 Syntax for the command is : "@var{width_type}"
3107 Change bandreject width.
3108 Syntax for the command is : "@var{width}"
3111 Change bandreject mix.
3112 Syntax for the command is : "@var{mix}"
3115 @section bass, lowshelf
3117 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3118 shelving filter with a response similar to that of a standard
3119 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3121 The filter accepts the following options:
3125 Give the gain at 0 Hz. Its useful range is about -20
3126 (for a large cut) to +20 (for a large boost).
3127 Beware of clipping when using a positive gain.
3130 Set the filter's central frequency and so can be used
3131 to extend or reduce the frequency range to be boosted or cut.
3132 The default value is @code{100} Hz.
3135 Set method to specify band-width of filter.
3150 Determine how steep is the filter's shelf transition.
3153 How much to use filtered signal in output. Default is 1.
3154 Range is between 0 and 1.
3157 Specify which channels to filter, by default all available are filtered.
3160 Normalize biquad coefficients, by default is disabled.
3161 Enabling it will normalize magnitude response at DC to 0dB.
3164 Set transform type of IIR filter.
3173 Set precison of filtering.
3176 Pick automatic sample format depending on surround filters.
3178 Always use signed 16-bit.
3180 Always use signed 32-bit.
3182 Always use float 32-bit.
3184 Always use float 64-bit.
3188 @subsection Commands
3190 This filter supports the following commands:
3193 Change bass frequency.
3194 Syntax for the command is : "@var{frequency}"
3197 Change bass width_type.
3198 Syntax for the command is : "@var{width_type}"
3202 Syntax for the command is : "@var{width}"
3206 Syntax for the command is : "@var{gain}"
3210 Syntax for the command is : "@var{mix}"
3215 Apply a biquad IIR filter with the given coefficients.
3216 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3217 are the numerator and denominator coefficients respectively.
3218 and @var{channels}, @var{c} specify which channels to filter, by default all
3219 available are filtered.
3221 @subsection Commands
3223 This filter supports the following commands:
3231 Change biquad parameter.
3232 Syntax for the command is : "@var{value}"
3235 How much to use filtered signal in output. Default is 1.
3236 Range is between 0 and 1.
3239 Specify which channels to filter, by default all available are filtered.
3242 Normalize biquad coefficients, by default is disabled.
3243 Enabling it will normalize magnitude response at DC to 0dB.
3246 Set transform type of IIR filter.
3255 Set precison of filtering.
3258 Pick automatic sample format depending on surround filters.
3260 Always use signed 16-bit.
3262 Always use signed 32-bit.
3264 Always use float 32-bit.
3266 Always use float 64-bit.
3271 Bauer stereo to binaural transformation, which improves headphone listening of
3272 stereo audio records.
3274 To enable compilation of this filter you need to configure FFmpeg with
3275 @code{--enable-libbs2b}.
3277 It accepts the following parameters:
3281 Pre-defined crossfeed level.
3285 Default level (fcut=700, feed=50).
3288 Chu Moy circuit (fcut=700, feed=60).
3291 Jan Meier circuit (fcut=650, feed=95).
3296 Cut frequency (in Hz).
3305 Remap input channels to new locations.
3307 It accepts the following parameters:
3310 Map channels from input to output. The argument is a '|'-separated list of
3311 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3312 @var{in_channel} form. @var{in_channel} can be either the name of the input
3313 channel (e.g. FL for front left) or its index in the input channel layout.
3314 @var{out_channel} is the name of the output channel or its index in the output
3315 channel layout. If @var{out_channel} is not given then it is implicitly an
3316 index, starting with zero and increasing by one for each mapping.
3318 @item channel_layout
3319 The channel layout of the output stream.
3322 If no mapping is present, the filter will implicitly map input channels to
3323 output channels, preserving indices.
3325 @subsection Examples
3329 For example, assuming a 5.1+downmix input MOV file,
3331 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3333 will create an output WAV file tagged as stereo from the downmix channels of
3337 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3339 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3343 @section channelsplit
3345 Split each channel from an input audio stream into a separate output stream.
3347 It accepts the following parameters:
3349 @item channel_layout
3350 The channel layout of the input stream. The default is "stereo".
3352 A channel layout describing the channels to be extracted as separate output streams
3353 or "all" to extract each input channel as a separate stream. The default is "all".
3355 Choosing channels not present in channel layout in the input will result in an error.
3358 @subsection Examples
3362 For example, assuming a stereo input MP3 file,
3364 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3366 will create an output Matroska file with two audio streams, one containing only
3367 the left channel and the other the right channel.
3370 Split a 5.1 WAV file into per-channel files:
3372 ffmpeg -i in.wav -filter_complex
3373 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3374 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3375 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3380 Extract only LFE from a 5.1 WAV file:
3382 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3383 -map '[LFE]' lfe.wav
3388 Add a chorus effect to the audio.
3390 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3392 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3393 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3394 The modulation depth defines the range the modulated delay is played before or after
3395 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3396 sound tuned around the original one, like in a chorus where some vocals are slightly
3399 It accepts the following parameters:
3402 Set input gain. Default is 0.4.
3405 Set output gain. Default is 0.4.
3408 Set delays. A typical delay is around 40ms to 60ms.
3420 @subsection Examples
3426 chorus=0.7:0.9:55:0.4:0.25:2
3432 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3436 Fuller sounding chorus with three delays:
3438 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
3443 Compress or expand the audio's dynamic range.
3445 It accepts the following parameters:
3451 A list of times in seconds for each channel over which the instantaneous level
3452 of the input signal is averaged to determine its volume. @var{attacks} refers to
3453 increase of volume and @var{decays} refers to decrease of volume. For most
3454 situations, the attack time (response to the audio getting louder) should be
3455 shorter than the decay time, because the human ear is more sensitive to sudden
3456 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3457 a typical value for decay is 0.8 seconds.
3458 If specified number of attacks & decays is lower than number of channels, the last
3459 set attack/decay will be used for all remaining channels.
3462 A list of points for the transfer function, specified in dB relative to the
3463 maximum possible signal amplitude. Each key points list must be defined using
3464 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3465 @code{x0/y0 x1/y1 x2/y2 ....}
3467 The input values must be in strictly increasing order but the transfer function
3468 does not have to be monotonically rising. The point @code{0/0} is assumed but
3469 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3470 function are @code{-70/-70|-60/-20|1/0}.
3473 Set the curve radius in dB for all joints. It defaults to 0.01.
3476 Set the additional gain in dB to be applied at all points on the transfer
3477 function. This allows for easy adjustment of the overall gain.
3481 Set an initial volume, in dB, to be assumed for each channel when filtering
3482 starts. This permits the user to supply a nominal level initially, so that, for
3483 example, a very large gain is not applied to initial signal levels before the
3484 companding has begun to operate. A typical value for audio which is initially
3485 quiet is -90 dB. It defaults to 0.
3488 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3489 delayed before being fed to the volume adjuster. Specifying a delay
3490 approximately equal to the attack/decay times allows the filter to effectively
3491 operate in predictive rather than reactive mode. It defaults to 0.
3495 @subsection Examples
3499 Make music with both quiet and loud passages suitable for listening to in a
3502 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3505 Another example for audio with whisper and explosion parts:
3507 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3511 A noise gate for when the noise is at a lower level than the signal:
3513 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3517 Here is another noise gate, this time for when the noise is at a higher level
3518 than the signal (making it, in some ways, similar to squelch):
3520 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3524 2:1 compression starting at -6dB:
3526 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3530 2:1 compression starting at -9dB:
3532 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3536 2:1 compression starting at -12dB:
3538 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3542 2:1 compression starting at -18dB:
3544 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3548 3:1 compression starting at -15dB:
3550 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3556 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3562 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
3566 Hard limiter at -6dB:
3568 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3572 Hard limiter at -12dB:
3574 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3578 Hard noise gate at -35 dB:
3580 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3586 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3590 @section compensationdelay
3592 Compensation Delay Line is a metric based delay to compensate differing
3593 positions of microphones or speakers.
3595 For example, you have recorded guitar with two microphones placed in
3596 different locations. Because the front of sound wave has fixed speed in
3597 normal conditions, the phasing of microphones can vary and depends on
3598 their location and interposition. The best sound mix can be achieved when
3599 these microphones are in phase (synchronized). Note that a distance of
3600 ~30 cm between microphones makes one microphone capture the signal in
3601 antiphase to the other microphone. That makes the final mix sound moody.
3602 This filter helps to solve phasing problems by adding different delays
3603 to each microphone track and make them synchronized.
3605 The best result can be reached when you take one track as base and
3606 synchronize other tracks one by one with it.
3607 Remember that synchronization/delay tolerance depends on sample rate, too.
3608 Higher sample rates will give more tolerance.
3610 The filter accepts the following parameters:
3614 Set millimeters distance. This is compensation distance for fine tuning.
3618 Set cm distance. This is compensation distance for tightening distance setup.
3622 Set meters distance. This is compensation distance for hard distance setup.
3626 Set dry amount. Amount of unprocessed (dry) signal.
3630 Set wet amount. Amount of processed (wet) signal.
3634 Set temperature in degrees Celsius. This is the temperature of the environment.
3639 Apply headphone crossfeed filter.
3641 Crossfeed is the process of blending the left and right channels of stereo
3643 It is mainly used to reduce extreme stereo separation of low frequencies.
3645 The intent is to produce more speaker like sound to the listener.
3647 The filter accepts the following options:
3651 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3652 This sets gain of low shelf filter for side part of stereo image.
3653 Default is -6dB. Max allowed is -30db when strength is set to 1.
3656 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3657 This sets cut off frequency of low shelf filter. Default is cut off near
3658 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3661 Set curve slope of low shelf filter. Default is 0.5.
3662 Allowed range is from 0.01 to 1.
3665 Set input gain. Default is 0.9.
3668 Set output gain. Default is 1.
3671 @subsection Commands
3673 This filter supports the all above options as @ref{commands}.
3675 @section crystalizer
3676 Simple algorithm to expand audio dynamic range.
3678 The filter accepts the following options:
3682 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3683 (unchanged sound) to 10.0 (maximum effect).
3686 Enable clipping. By default is enabled.
3689 @subsection Commands
3691 This filter supports the all above options as @ref{commands}.
3694 Apply a DC shift to the audio.
3696 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3697 in the recording chain) from the audio. The effect of a DC offset is reduced
3698 headroom and hence volume. The @ref{astats} filter can be used to determine if
3699 a signal has a DC offset.
3703 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3707 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3708 used to prevent clipping.
3713 Apply de-essing to the audio samples.
3717 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3721 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3725 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3729 Set the output mode.
3731 It accepts the following values:
3734 Pass input unchanged.
3737 Pass ess filtered out.
3742 Default value is @var{o}.
3748 Measure audio dynamic range.
3750 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3751 is found in transition material. And anything less that 8 have very poor dynamics
3752 and is very compressed.
3754 The filter accepts the following options:
3758 Set window length in seconds used to split audio into segments of equal length.
3759 Default is 3 seconds.
3763 Dynamic Audio Normalizer.
3765 This filter applies a certain amount of gain to the input audio in order
3766 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3767 contrast to more "simple" normalization algorithms, the Dynamic Audio
3768 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3769 This allows for applying extra gain to the "quiet" sections of the audio
3770 while avoiding distortions or clipping the "loud" sections. In other words:
3771 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3772 sections, in the sense that the volume of each section is brought to the
3773 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3774 this goal *without* applying "dynamic range compressing". It will retain 100%
3775 of the dynamic range *within* each section of the audio file.
3779 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3780 Default is 500 milliseconds.
3781 The Dynamic Audio Normalizer processes the input audio in small chunks,
3782 referred to as frames. This is required, because a peak magnitude has no
3783 meaning for just a single sample value. Instead, we need to determine the
3784 peak magnitude for a contiguous sequence of sample values. While a "standard"
3785 normalizer would simply use the peak magnitude of the complete file, the
3786 Dynamic Audio Normalizer determines the peak magnitude individually for each
3787 frame. The length of a frame is specified in milliseconds. By default, the
3788 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3789 been found to give good results with most files.
3790 Note that the exact frame length, in number of samples, will be determined
3791 automatically, based on the sampling rate of the individual input audio file.
3794 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3795 number. Default is 31.
3796 Probably the most important parameter of the Dynamic Audio Normalizer is the
3797 @code{window size} of the Gaussian smoothing filter. The filter's window size
3798 is specified in frames, centered around the current frame. For the sake of
3799 simplicity, this must be an odd number. Consequently, the default value of 31
3800 takes into account the current frame, as well as the 15 preceding frames and
3801 the 15 subsequent frames. Using a larger window results in a stronger
3802 smoothing effect and thus in less gain variation, i.e. slower gain
3803 adaptation. Conversely, using a smaller window results in a weaker smoothing
3804 effect and thus in more gain variation, i.e. faster gain adaptation.
3805 In other words, the more you increase this value, the more the Dynamic Audio
3806 Normalizer will behave like a "traditional" normalization filter. On the
3807 contrary, the more you decrease this value, the more the Dynamic Audio
3808 Normalizer will behave like a dynamic range compressor.
3811 Set the target peak value. This specifies the highest permissible magnitude
3812 level for the normalized audio input. This filter will try to approach the
3813 target peak magnitude as closely as possible, but at the same time it also
3814 makes sure that the normalized signal will never exceed the peak magnitude.
3815 A frame's maximum local gain factor is imposed directly by the target peak
3816 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3817 It is not recommended to go above this value.
3820 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3821 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3822 factor for each input frame, i.e. the maximum gain factor that does not
3823 result in clipping or distortion. The maximum gain factor is determined by
3824 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3825 additionally bounds the frame's maximum gain factor by a predetermined
3826 (global) maximum gain factor. This is done in order to avoid excessive gain
3827 factors in "silent" or almost silent frames. By default, the maximum gain
3828 factor is 10.0, For most inputs the default value should be sufficient and
3829 it usually is not recommended to increase this value. Though, for input
3830 with an extremely low overall volume level, it may be necessary to allow even
3831 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3832 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3833 Instead, a "sigmoid" threshold function will be applied. This way, the
3834 gain factors will smoothly approach the threshold value, but never exceed that
3838 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3839 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3840 This means that the maximum local gain factor for each frame is defined
3841 (only) by the frame's highest magnitude sample. This way, the samples can
3842 be amplified as much as possible without exceeding the maximum signal
3843 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3844 Normalizer can also take into account the frame's root mean square,
3845 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3846 determine the power of a time-varying signal. It is therefore considered
3847 that the RMS is a better approximation of the "perceived loudness" than
3848 just looking at the signal's peak magnitude. Consequently, by adjusting all
3849 frames to a constant RMS value, a uniform "perceived loudness" can be
3850 established. If a target RMS value has been specified, a frame's local gain
3851 factor is defined as the factor that would result in exactly that RMS value.
3852 Note, however, that the maximum local gain factor is still restricted by the
3853 frame's highest magnitude sample, in order to prevent clipping.
3856 Enable channels coupling. By default is enabled.
3857 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3858 amount. This means the same gain factor will be applied to all channels, i.e.
3859 the maximum possible gain factor is determined by the "loudest" channel.
3860 However, in some recordings, it may happen that the volume of the different
3861 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3862 In this case, this option can be used to disable the channel coupling. This way,
3863 the gain factor will be determined independently for each channel, depending
3864 only on the individual channel's highest magnitude sample. This allows for
3865 harmonizing the volume of the different channels.
3868 Enable DC bias correction. By default is disabled.
3869 An audio signal (in the time domain) is a sequence of sample values.
3870 In the Dynamic Audio Normalizer these sample values are represented in the
3871 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3872 audio signal, or "waveform", should be centered around the zero point.
3873 That means if we calculate the mean value of all samples in a file, or in a
3874 single frame, then the result should be 0.0 or at least very close to that
3875 value. If, however, there is a significant deviation of the mean value from
3876 0.0, in either positive or negative direction, this is referred to as a
3877 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3878 Audio Normalizer provides optional DC bias correction.
3879 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3880 the mean value, or "DC correction" offset, of each input frame and subtract
3881 that value from all of the frame's sample values which ensures those samples
3882 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3883 boundaries, the DC correction offset values will be interpolated smoothly
3884 between neighbouring frames.
3886 @item altboundary, b
3887 Enable alternative boundary mode. By default is disabled.
3888 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3889 around each frame. This includes the preceding frames as well as the
3890 subsequent frames. However, for the "boundary" frames, located at the very
3891 beginning and at the very end of the audio file, not all neighbouring
3892 frames are available. In particular, for the first few frames in the audio
3893 file, the preceding frames are not known. And, similarly, for the last few
3894 frames in the audio file, the subsequent frames are not known. Thus, the
3895 question arises which gain factors should be assumed for the missing frames
3896 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3897 to deal with this situation. The default boundary mode assumes a gain factor
3898 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3899 "fade out" at the beginning and at the end of the input, respectively.
3902 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3903 By default, the Dynamic Audio Normalizer does not apply "traditional"
3904 compression. This means that signal peaks will not be pruned and thus the
3905 full dynamic range will be retained within each local neighbourhood. However,
3906 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3907 normalization algorithm with a more "traditional" compression.
3908 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3909 (thresholding) function. If (and only if) the compression feature is enabled,
3910 all input frames will be processed by a soft knee thresholding function prior
3911 to the actual normalization process. Put simply, the thresholding function is
3912 going to prune all samples whose magnitude exceeds a certain threshold value.
3913 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3914 value. Instead, the threshold value will be adjusted for each individual
3916 In general, smaller parameters result in stronger compression, and vice versa.
3917 Values below 3.0 are not recommended, because audible distortion may appear.
3920 Set the target threshold value. This specifies the lowest permissible
3921 magnitude level for the audio input which will be normalized.
3922 If input frame volume is above this value frame will be normalized.
3923 Otherwise frame may not be normalized at all. The default value is set
3924 to 0, which means all input frames will be normalized.
3925 This option is mostly useful if digital noise is not wanted to be amplified.
3928 @subsection Commands
3930 This filter supports the all above options as @ref{commands}.
3934 Make audio easier to listen to on headphones.
3936 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3937 so that when listened to on headphones the stereo image is moved from
3938 inside your head (standard for headphones) to outside and in front of
3939 the listener (standard for speakers).
3945 Apply a two-pole peaking equalisation (EQ) filter. With this
3946 filter, the signal-level at and around a selected frequency can
3947 be increased or decreased, whilst (unlike bandpass and bandreject
3948 filters) that at all other frequencies is unchanged.
3950 In order to produce complex equalisation curves, this filter can
3951 be given several times, each with a different central frequency.
3953 The filter accepts the following options:
3957 Set the filter's central frequency in Hz.
3960 Set method to specify band-width of filter.
3975 Specify the band-width of a filter in width_type units.
3978 Set the required gain or attenuation in dB.
3979 Beware of clipping when using a positive gain.
3982 How much to use filtered signal in output. Default is 1.
3983 Range is between 0 and 1.
3986 Specify which channels to filter, by default all available are filtered.
3989 Normalize biquad coefficients, by default is disabled.
3990 Enabling it will normalize magnitude response at DC to 0dB.
3993 Set transform type of IIR filter.
4002 Set precison of filtering.
4005 Pick automatic sample format depending on surround filters.
4007 Always use signed 16-bit.
4009 Always use signed 32-bit.
4011 Always use float 32-bit.
4013 Always use float 64-bit.
4017 @subsection Examples
4020 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4022 equalizer=f=1000:t=h:width=200:g=-10
4026 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4028 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4032 @subsection Commands
4034 This filter supports the following commands:
4037 Change equalizer frequency.
4038 Syntax for the command is : "@var{frequency}"
4041 Change equalizer width_type.
4042 Syntax for the command is : "@var{width_type}"
4045 Change equalizer width.
4046 Syntax for the command is : "@var{width}"
4049 Change equalizer gain.
4050 Syntax for the command is : "@var{gain}"
4053 Change equalizer mix.
4054 Syntax for the command is : "@var{mix}"
4057 @section extrastereo
4059 Linearly increases the difference between left and right channels which
4060 adds some sort of "live" effect to playback.
4062 The filter accepts the following options:
4066 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4067 (average of both channels), with 1.0 sound will be unchanged, with
4068 -1.0 left and right channels will be swapped.
4071 Enable clipping. By default is enabled.
4074 @subsection Commands
4076 This filter supports the all above options as @ref{commands}.
4078 @section firequalizer
4079 Apply FIR Equalization using arbitrary frequency response.
4081 The filter accepts the following option:
4085 Set gain curve equation (in dB). The expression can contain variables:
4088 the evaluated frequency
4092 channel number, set to 0 when multichannels evaluation is disabled
4094 channel id, see libavutil/channel_layout.h, set to the first channel id when
4095 multichannels evaluation is disabled
4099 channel_layout, see libavutil/channel_layout.h
4104 @item gain_interpolate(f)
4105 interpolate gain on frequency f based on gain_entry
4106 @item cubic_interpolate(f)
4107 same as gain_interpolate, but smoother
4109 This option is also available as command. Default is @code{gain_interpolate(f)}.
4112 Set gain entry for gain_interpolate function. The expression can
4116 store gain entry at frequency f with value g
4118 This option is also available as command.
4121 Set filter delay in seconds. Higher value means more accurate.
4122 Default is @code{0.01}.
4125 Set filter accuracy in Hz. Lower value means more accurate.
4126 Default is @code{5}.
4129 Set window function. Acceptable values are:
4132 rectangular window, useful when gain curve is already smooth
4134 hann window (default)
4140 3-terms continuous 1st derivative nuttall window
4142 minimum 3-terms discontinuous nuttall window
4144 4-terms continuous 1st derivative nuttall window
4146 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4148 blackman-harris window
4154 If enabled, use fixed number of audio samples. This improves speed when
4155 filtering with large delay. Default is disabled.
4158 Enable multichannels evaluation on gain. Default is disabled.
4161 Enable zero phase mode by subtracting timestamp to compensate delay.
4162 Default is disabled.
4165 Set scale used by gain. Acceptable values are:
4168 linear frequency, linear gain
4170 linear frequency, logarithmic (in dB) gain (default)
4172 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4174 logarithmic frequency, logarithmic gain
4178 Set file for dumping, suitable for gnuplot.
4181 Set scale for dumpfile. Acceptable values are same with scale option.
4185 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4186 Default is disabled.
4189 Enable minimum phase impulse response. Default is disabled.
4192 @subsection Examples
4197 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4200 lowpass at 1000 Hz with gain_entry:
4202 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4205 custom equalization:
4207 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4210 higher delay with zero phase to compensate delay:
4212 firequalizer=delay=0.1:fixed=on:zero_phase=on
4215 lowpass on left channel, highpass on right channel:
4217 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4218 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4223 Apply a flanging effect to the audio.
4225 The filter accepts the following options:
4229 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4232 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4235 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4239 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4240 Default value is 71.
4243 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4246 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4247 Default value is @var{sinusoidal}.
4250 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4251 Default value is 25.
4254 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4255 Default is @var{linear}.
4259 Apply Haas effect to audio.
4261 Note that this makes most sense to apply on mono signals.
4262 With this filter applied to mono signals it give some directionality and
4263 stretches its stereo image.
4265 The filter accepts the following options:
4269 Set input level. By default is @var{1}, or 0dB
4272 Set output level. By default is @var{1}, or 0dB.
4275 Set gain applied to side part of signal. By default is @var{1}.
4278 Set kind of middle source. Can be one of the following:
4288 Pick middle part signal of stereo image.
4291 Pick side part signal of stereo image.
4295 Change middle phase. By default is disabled.
4298 Set left channel delay. By default is @var{2.05} milliseconds.
4301 Set left channel balance. By default is @var{-1}.
4304 Set left channel gain. By default is @var{1}.
4307 Change left phase. By default is disabled.
4310 Set right channel delay. By defaults is @var{2.12} milliseconds.
4313 Set right channel balance. By default is @var{1}.
4316 Set right channel gain. By default is @var{1}.
4319 Change right phase. By default is enabled.
4324 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4325 embedded HDCD codes is expanded into a 20-bit PCM stream.
4327 The filter supports the Peak Extend and Low-level Gain Adjustment features
4328 of HDCD, and detects the Transient Filter flag.
4331 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4334 When using the filter with wav, note the default encoding for wav is 16-bit,
4335 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4336 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4338 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4339 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4342 The filter accepts the following options:
4345 @item disable_autoconvert
4346 Disable any automatic format conversion or resampling in the filter graph.
4348 @item process_stereo
4349 Process the stereo channels together. If target_gain does not match between
4350 channels, consider it invalid and use the last valid target_gain.
4353 Set the code detect timer period in ms.
4356 Always extend peaks above -3dBFS even if PE isn't signaled.
4359 Replace audio with a solid tone and adjust the amplitude to signal some
4360 specific aspect of the decoding process. The output file can be loaded in
4361 an audio editor alongside the original to aid analysis.
4363 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4370 Gain adjustment level at each sample
4372 Samples where peak extend occurs
4374 Samples where the code detect timer is active
4376 Samples where the target gain does not match between channels
4382 Apply head-related transfer functions (HRTFs) to create virtual
4383 loudspeakers around the user for binaural listening via headphones.
4384 The HRIRs are provided via additional streams, for each channel
4385 one stereo input stream is needed.
4387 The filter accepts the following options:
4391 Set mapping of input streams for convolution.
4392 The argument is a '|'-separated list of channel names in order as they
4393 are given as additional stream inputs for filter.
4394 This also specify number of input streams. Number of input streams
4395 must be not less than number of channels in first stream plus one.
4398 Set gain applied to audio. Value is in dB. Default is 0.
4401 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4402 processing audio in time domain which is slow.
4403 @var{freq} is processing audio in frequency domain which is fast.
4404 Default is @var{freq}.
4407 Set custom gain for LFE channels. Value is in dB. Default is 0.
4410 Set size of frame in number of samples which will be processed at once.
4411 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4414 Set format of hrir stream.
4415 Default value is @var{stereo}. Alternative value is @var{multich}.
4416 If value is set to @var{stereo}, number of additional streams should
4417 be greater or equal to number of input channels in first input stream.
4418 Also each additional stream should have stereo number of channels.
4419 If value is set to @var{multich}, number of additional streams should
4420 be exactly one. Also number of input channels of additional stream
4421 should be equal or greater than twice number of channels of first input
4425 @subsection Examples
4429 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4430 each amovie filter use stereo file with IR coefficients as input.
4431 The files give coefficients for each position of virtual loudspeaker:
4434 -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"
4439 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4440 but now in @var{multich} @var{hrir} format.
4442 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"
4449 Apply a high-pass filter with 3dB point frequency.
4450 The filter can be either single-pole, or double-pole (the default).
4451 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4453 The filter accepts the following options:
4457 Set frequency in Hz. Default is 3000.
4460 Set number of poles. Default is 2.
4463 Set method to specify band-width of filter.
4478 Specify the band-width of a filter in width_type units.
4479 Applies only to double-pole filter.
4480 The default is 0.707q and gives a Butterworth response.
4483 How much to use filtered signal in output. Default is 1.
4484 Range is between 0 and 1.
4487 Specify which channels to filter, by default all available are filtered.
4490 Normalize biquad coefficients, by default is disabled.
4491 Enabling it will normalize magnitude response at DC to 0dB.
4494 Set transform type of IIR filter.
4503 Set precison of filtering.
4506 Pick automatic sample format depending on surround filters.
4508 Always use signed 16-bit.
4510 Always use signed 32-bit.
4512 Always use float 32-bit.
4514 Always use float 64-bit.
4518 @subsection Commands
4520 This filter supports the following commands:
4523 Change highpass frequency.
4524 Syntax for the command is : "@var{frequency}"
4527 Change highpass width_type.
4528 Syntax for the command is : "@var{width_type}"
4531 Change highpass width.
4532 Syntax for the command is : "@var{width}"
4535 Change highpass mix.
4536 Syntax for the command is : "@var{mix}"
4541 Join multiple input streams into one multi-channel stream.
4543 It accepts the following parameters:
4547 The number of input streams. It defaults to 2.
4549 @item channel_layout
4550 The desired output channel layout. It defaults to stereo.
4553 Map channels from inputs to output. The argument is a '|'-separated list of
4554 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4555 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4556 can be either the name of the input channel (e.g. FL for front left) or its
4557 index in the specified input stream. @var{out_channel} is the name of the output
4561 The filter will attempt to guess the mappings when they are not specified
4562 explicitly. It does so by first trying to find an unused matching input channel
4563 and if that fails it picks the first unused input channel.
4565 Join 3 inputs (with properly set channel layouts):
4567 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4570 Build a 5.1 output from 6 single-channel streams:
4572 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4573 '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'
4579 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4581 To enable compilation of this filter you need to configure FFmpeg with
4582 @code{--enable-ladspa}.
4586 Specifies the name of LADSPA plugin library to load. If the environment
4587 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4588 each one of the directories specified by the colon separated list in
4589 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4590 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4591 @file{/usr/lib/ladspa/}.
4594 Specifies the plugin within the library. Some libraries contain only
4595 one plugin, but others contain many of them. If this is not set filter
4596 will list all available plugins within the specified library.
4599 Set the '|' separated list of controls which are zero or more floating point
4600 values that determine the behavior of the loaded plugin (for example delay,
4602 Controls need to be defined using the following syntax:
4603 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4604 @var{valuei} is the value set on the @var{i}-th control.
4605 Alternatively they can be also defined using the following syntax:
4606 @var{value0}|@var{value1}|@var{value2}|..., where
4607 @var{valuei} is the value set on the @var{i}-th control.
4608 If @option{controls} is set to @code{help}, all available controls and
4609 their valid ranges are printed.
4611 @item sample_rate, s
4612 Specify the sample rate, default to 44100. Only used if plugin have
4616 Set the number of samples per channel per each output frame, default
4617 is 1024. Only used if plugin have zero inputs.
4620 Set the minimum duration of the sourced audio. See
4621 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4622 for the accepted syntax.
4623 Note that the resulting duration may be greater than the specified duration,
4624 as the generated audio is always cut at the end of a complete frame.
4625 If not specified, or the expressed duration is negative, the audio is
4626 supposed to be generated forever.
4627 Only used if plugin have zero inputs.
4630 Enable latency compensation, by default is disabled.
4631 Only used if plugin have inputs.
4634 @subsection Examples
4638 List all available plugins within amp (LADSPA example plugin) library:
4644 List all available controls and their valid ranges for @code{vcf_notch}
4645 plugin from @code{VCF} library:
4647 ladspa=f=vcf:p=vcf_notch:c=help
4651 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4654 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4658 Add reverberation to the audio using TAP-plugins
4659 (Tom's Audio Processing plugins):
4661 ladspa=file=tap_reverb:tap_reverb
4665 Generate white noise, with 0.2 amplitude:
4667 ladspa=file=cmt:noise_source_white:c=c0=.2
4671 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4672 @code{C* Audio Plugin Suite} (CAPS) library:
4674 ladspa=file=caps:Click:c=c1=20'
4678 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4680 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4684 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4685 @code{SWH Plugins} collection:
4687 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4691 Attenuate low frequencies using Multiband EQ from Steve Harris
4692 @code{SWH Plugins} collection:
4694 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4698 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4701 ladspa=caps:Narrower
4705 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4707 ladspa=caps:White:.2
4711 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4713 ladspa=caps:Fractal:c=c1=1
4717 Dynamic volume normalization using @code{VLevel} plugin:
4719 ladspa=vlevel-ladspa:vlevel_mono
4723 @subsection Commands
4725 This filter supports the following commands:
4728 Modify the @var{N}-th control value.
4730 If the specified value is not valid, it is ignored and prior one is kept.
4735 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4736 Support for both single pass (livestreams, files) and double pass (files) modes.
4737 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4738 detect true peaks, the audio stream will be upsampled to 192 kHz.
4739 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4741 The filter accepts the following options:
4745 Set integrated loudness target.
4746 Range is -70.0 - -5.0. Default value is -24.0.
4749 Set loudness range target.
4750 Range is 1.0 - 20.0. Default value is 7.0.
4753 Set maximum true peak.
4754 Range is -9.0 - +0.0. Default value is -2.0.
4756 @item measured_I, measured_i
4757 Measured IL of input file.
4758 Range is -99.0 - +0.0.
4760 @item measured_LRA, measured_lra
4761 Measured LRA of input file.
4762 Range is 0.0 - 99.0.
4764 @item measured_TP, measured_tp
4765 Measured true peak of input file.
4766 Range is -99.0 - +99.0.
4768 @item measured_thresh
4769 Measured threshold of input file.
4770 Range is -99.0 - +0.0.
4773 Set offset gain. Gain is applied before the true-peak limiter.
4774 Range is -99.0 - +99.0. Default is +0.0.
4777 Normalize by linearly scaling the source audio.
4778 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4779 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4780 be lower than source LRA and the change in integrated loudness shouldn't
4781 result in a true peak which exceeds the target TP. If any of these
4782 conditions aren't met, normalization mode will revert to @var{dynamic}.
4783 Options are @code{true} or @code{false}. Default is @code{true}.
4786 Treat mono input files as "dual-mono". If a mono file is intended for playback
4787 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4788 If set to @code{true}, this option will compensate for this effect.
4789 Multi-channel input files are not affected by this option.
4790 Options are true or false. Default is false.
4793 Set print format for stats. Options are summary, json, or none.
4794 Default value is none.
4799 Apply a low-pass filter with 3dB point frequency.
4800 The filter can be either single-pole or double-pole (the default).
4801 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4803 The filter accepts the following options:
4807 Set frequency in Hz. Default is 500.
4810 Set number of poles. Default is 2.
4813 Set method to specify band-width of filter.
4828 Specify the band-width of a filter in width_type units.
4829 Applies only to double-pole filter.
4830 The default is 0.707q and gives a Butterworth response.
4833 How much to use filtered signal in output. Default is 1.
4834 Range is between 0 and 1.
4837 Specify which channels to filter, by default all available are filtered.
4840 Normalize biquad coefficients, by default is disabled.
4841 Enabling it will normalize magnitude response at DC to 0dB.
4844 Set transform type of IIR filter.
4853 Set precison of filtering.
4856 Pick automatic sample format depending on surround filters.
4858 Always use signed 16-bit.
4860 Always use signed 32-bit.
4862 Always use float 32-bit.
4864 Always use float 64-bit.
4868 @subsection Examples
4871 Lowpass only LFE channel, it LFE is not present it does nothing:
4877 @subsection Commands
4879 This filter supports the following commands:
4882 Change lowpass frequency.
4883 Syntax for the command is : "@var{frequency}"
4886 Change lowpass width_type.
4887 Syntax for the command is : "@var{width_type}"
4890 Change lowpass width.
4891 Syntax for the command is : "@var{width}"
4895 Syntax for the command is : "@var{mix}"
4900 Load a LV2 (LADSPA Version 2) plugin.
4902 To enable compilation of this filter you need to configure FFmpeg with
4903 @code{--enable-lv2}.
4907 Specifies the plugin URI. You may need to escape ':'.
4910 Set the '|' separated list of controls which are zero or more floating point
4911 values that determine the behavior of the loaded plugin (for example delay,
4913 If @option{controls} is set to @code{help}, all available controls and
4914 their valid ranges are printed.
4916 @item sample_rate, s
4917 Specify the sample rate, default to 44100. Only used if plugin have
4921 Set the number of samples per channel per each output frame, default
4922 is 1024. Only used if plugin have zero inputs.
4925 Set the minimum duration of the sourced audio. See
4926 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4927 for the accepted syntax.
4928 Note that the resulting duration may be greater than the specified duration,
4929 as the generated audio is always cut at the end of a complete frame.
4930 If not specified, or the expressed duration is negative, the audio is
4931 supposed to be generated forever.
4932 Only used if plugin have zero inputs.
4935 @subsection Examples
4939 Apply bass enhancer plugin from Calf:
4941 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4945 Apply vinyl plugin from Calf:
4947 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4951 Apply bit crusher plugin from ArtyFX:
4953 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4958 Multiband Compress or expand the audio's dynamic range.
4960 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4961 This is akin to the crossover of a loudspeaker, and results in flat frequency
4962 response when absent compander action.
4964 It accepts the following parameters:
4968 This option syntax is:
4969 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4970 For explanation of each item refer to compand filter documentation.
4976 Mix channels with specific gain levels. The filter accepts the output
4977 channel layout followed by a set of channels definitions.
4979 This filter is also designed to efficiently remap the channels of an audio
4982 The filter accepts parameters of the form:
4983 "@var{l}|@var{outdef}|@var{outdef}|..."
4987 output channel layout or number of channels
4990 output channel specification, of the form:
4991 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4994 output channel to define, either a channel name (FL, FR, etc.) or a channel
4995 number (c0, c1, etc.)
4998 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5001 input channel to use, see out_name for details; it is not possible to mix
5002 named and numbered input channels
5005 If the `=' in a channel specification is replaced by `<', then the gains for
5006 that specification will be renormalized so that the total is 1, thus
5007 avoiding clipping noise.
5009 @subsection Mixing examples
5011 For example, if you want to down-mix from stereo to mono, but with a bigger
5012 factor for the left channel:
5014 pan=1c|c0=0.9*c0+0.1*c1
5017 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5018 7-channels surround:
5020 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5023 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5024 that should be preferred (see "-ac" option) unless you have very specific
5027 @subsection Remapping examples
5029 The channel remapping will be effective if, and only if:
5032 @item gain coefficients are zeroes or ones,
5033 @item only one input per channel output,
5036 If all these conditions are satisfied, the filter will notify the user ("Pure
5037 channel mapping detected"), and use an optimized and lossless method to do the
5040 For example, if you have a 5.1 source and want a stereo audio stream by
5041 dropping the extra channels:
5043 pan="stereo| c0=FL | c1=FR"
5046 Given the same source, you can also switch front left and front right channels
5047 and keep the input channel layout:
5049 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5052 If the input is a stereo audio stream, you can mute the front left channel (and
5053 still keep the stereo channel layout) with:
5058 Still with a stereo audio stream input, you can copy the right channel in both
5059 front left and right:
5061 pan="stereo| c0=FR | c1=FR"
5066 ReplayGain scanner filter. This filter takes an audio stream as an input and
5067 outputs it unchanged.
5068 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5072 Convert the audio sample format, sample rate and channel layout. It is
5073 not meant to be used directly.
5076 Apply time-stretching and pitch-shifting with librubberband.
5078 To enable compilation of this filter, you need to configure FFmpeg with
5079 @code{--enable-librubberband}.
5081 The filter accepts the following options:
5085 Set tempo scale factor.
5088 Set pitch scale factor.
5091 Set transients detector.
5092 Possible values are:
5101 Possible values are:
5110 Possible values are:
5117 Set processing window size.
5118 Possible values are:
5127 Possible values are:
5134 Enable formant preservation when shift pitching.
5135 Possible values are:
5143 Possible values are:
5152 Possible values are:
5159 @subsection Commands
5161 This filter supports the following commands:
5164 Change filter tempo scale factor.
5165 Syntax for the command is : "@var{tempo}"
5168 Change filter pitch scale factor.
5169 Syntax for the command is : "@var{pitch}"
5172 @section sidechaincompress
5174 This filter acts like normal compressor but has the ability to compress
5175 detected signal using second input signal.
5176 It needs two input streams and returns one output stream.
5177 First input stream will be processed depending on second stream signal.
5178 The filtered signal then can be filtered with other filters in later stages of
5179 processing. See @ref{pan} and @ref{amerge} filter.
5181 The filter accepts the following options:
5185 Set input gain. Default is 1. Range is between 0.015625 and 64.
5188 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5189 Default is @code{downward}.
5192 If a signal of second stream raises above this level it will affect the gain
5193 reduction of first stream.
5194 By default is 0.125. Range is between 0.00097563 and 1.
5197 Set a ratio about which the signal is reduced. 1:2 means that if the level
5198 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5199 Default is 2. Range is between 1 and 20.
5202 Amount of milliseconds the signal has to rise above the threshold before gain
5203 reduction starts. Default is 20. Range is between 0.01 and 2000.
5206 Amount of milliseconds the signal has to fall below the threshold before
5207 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5210 Set the amount by how much signal will be amplified after processing.
5211 Default is 1. Range is from 1 to 64.
5214 Curve the sharp knee around the threshold to enter gain reduction more softly.
5215 Default is 2.82843. Range is between 1 and 8.
5218 Choose if the @code{average} level between all channels of side-chain stream
5219 or the louder(@code{maximum}) channel of side-chain stream affects the
5220 reduction. Default is @code{average}.
5223 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5224 of @code{rms}. Default is @code{rms} which is mainly smoother.
5227 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5230 How much to use compressed signal in output. Default is 1.
5231 Range is between 0 and 1.
5234 @subsection Commands
5236 This filter supports the all above options as @ref{commands}.
5238 @subsection Examples
5242 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5243 depending on the signal of 2nd input and later compressed signal to be
5244 merged with 2nd input:
5246 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5250 @section sidechaingate
5252 A sidechain gate acts like a normal (wideband) gate but has the ability to
5253 filter the detected signal before sending it to the gain reduction stage.
5254 Normally a gate uses the full range signal to detect a level above the
5256 For example: If you cut all lower frequencies from your sidechain signal
5257 the gate will decrease the volume of your track only if not enough highs
5258 appear. With this technique you are able to reduce the resonation of a
5259 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5261 It needs two input streams and returns one output stream.
5262 First input stream will be processed depending on second stream signal.
5264 The filter accepts the following options:
5268 Set input level before filtering.
5269 Default is 1. Allowed range is from 0.015625 to 64.
5272 Set the mode of operation. Can be @code{upward} or @code{downward}.
5273 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5274 will be amplified, expanding dynamic range in upward direction.
5275 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5278 Set the level of gain reduction when the signal is below the threshold.
5279 Default is 0.06125. Allowed range is from 0 to 1.
5280 Setting this to 0 disables reduction and then filter behaves like expander.
5283 If a signal rises above this level the gain reduction is released.
5284 Default is 0.125. Allowed range is from 0 to 1.
5287 Set a ratio about which the signal is reduced.
5288 Default is 2. Allowed range is from 1 to 9000.
5291 Amount of milliseconds the signal has to rise above the threshold before gain
5293 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5296 Amount of milliseconds the signal has to fall below the threshold before the
5297 reduction is increased again. Default is 250 milliseconds.
5298 Allowed range is from 0.01 to 9000.
5301 Set amount of amplification of signal after processing.
5302 Default is 1. Allowed range is from 1 to 64.
5305 Curve the sharp knee around the threshold to enter gain reduction more softly.
5306 Default is 2.828427125. Allowed range is from 1 to 8.
5309 Choose if exact signal should be taken for detection or an RMS like one.
5310 Default is rms. Can be peak or rms.
5313 Choose if the average level between all channels or the louder channel affects
5315 Default is average. Can be average or maximum.
5318 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5321 @subsection Commands
5323 This filter supports the all above options as @ref{commands}.
5325 @section silencedetect
5327 Detect silence in an audio stream.
5329 This filter logs a message when it detects that the input audio volume is less
5330 or equal to a noise tolerance value for a duration greater or equal to the
5331 minimum detected noise duration.
5333 The printed times and duration are expressed in seconds. The
5334 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5335 is set on the first frame whose timestamp equals or exceeds the detection
5336 duration and it contains the timestamp of the first frame of the silence.
5338 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5339 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5340 keys are set on the first frame after the silence. If @option{mono} is
5341 enabled, and each channel is evaluated separately, the @code{.X}
5342 suffixed keys are used, and @code{X} corresponds to the channel number.
5344 The filter accepts the following options:
5348 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5349 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5352 Set silence duration until notification (default is 2 seconds). See
5353 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5354 for the accepted syntax.
5357 Process each channel separately, instead of combined. By default is disabled.
5360 @subsection Examples
5364 Detect 5 seconds of silence with -50dB noise tolerance:
5366 silencedetect=n=-50dB:d=5
5370 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5371 tolerance in @file{silence.mp3}:
5373 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5377 @section silenceremove
5379 Remove silence from the beginning, middle or end of the audio.
5381 The filter accepts the following options:
5385 This value is used to indicate if audio should be trimmed at beginning of
5386 the audio. A value of zero indicates no silence should be trimmed from the
5387 beginning. When specifying a non-zero value, it trims audio up until it
5388 finds non-silence. Normally, when trimming silence from beginning of audio
5389 the @var{start_periods} will be @code{1} but it can be increased to higher
5390 values to trim all audio up to specific count of non-silence periods.
5391 Default value is @code{0}.
5393 @item start_duration
5394 Specify the amount of time that non-silence must be detected before it stops
5395 trimming audio. By increasing the duration, bursts of noises can be treated
5396 as silence and trimmed off. Default value is @code{0}.
5398 @item start_threshold
5399 This indicates what sample value should be treated as silence. For digital
5400 audio, a value of @code{0} may be fine but for audio recorded from analog,
5401 you may wish to increase the value to account for background noise.
5402 Can be specified in dB (in case "dB" is appended to the specified value)
5403 or amplitude ratio. Default value is @code{0}.
5406 Specify max duration of silence at beginning that will be kept after
5407 trimming. Default is 0, which is equal to trimming all samples detected
5411 Specify mode of detection of silence end in start of multi-channel audio.
5412 Can be @var{any} or @var{all}. Default is @var{any}.
5413 With @var{any}, any sample that is detected as non-silence will cause
5414 stopped trimming of silence.
5415 With @var{all}, only if all channels are detected as non-silence will cause
5416 stopped trimming of silence.
5419 Set the count for trimming silence from the end of audio.
5420 To remove silence from the middle of a file, specify a @var{stop_periods}
5421 that is negative. This value is then treated as a positive value and is
5422 used to indicate the effect should restart processing as specified by
5423 @var{start_periods}, making it suitable for removing periods of silence
5424 in the middle of the audio.
5425 Default value is @code{0}.
5428 Specify a duration of silence that must exist before audio is not copied any
5429 more. By specifying a higher duration, silence that is wanted can be left in
5431 Default value is @code{0}.
5433 @item stop_threshold
5434 This is the same as @option{start_threshold} but for trimming silence from
5436 Can be specified in dB (in case "dB" is appended to the specified value)
5437 or amplitude ratio. Default value is @code{0}.
5440 Specify max duration of silence at end that will be kept after
5441 trimming. Default is 0, which is equal to trimming all samples detected
5445 Specify mode of detection of silence start in end of multi-channel audio.
5446 Can be @var{any} or @var{all}. Default is @var{any}.
5447 With @var{any}, any sample that is detected as non-silence will cause
5448 stopped trimming of silence.
5449 With @var{all}, only if all channels are detected as non-silence will cause
5450 stopped trimming of silence.
5453 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5454 and works better with digital silence which is exactly 0.
5455 Default value is @code{rms}.
5458 Set duration in number of seconds used to calculate size of window in number
5459 of samples for detecting silence.
5460 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5463 @subsection Examples
5467 The following example shows how this filter can be used to start a recording
5468 that does not contain the delay at the start which usually occurs between
5469 pressing the record button and the start of the performance:
5471 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5475 Trim all silence encountered from beginning to end where there is more than 1
5476 second of silence in audio:
5478 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5482 Trim all digital silence samples, using peak detection, from beginning to end
5483 where there is more than 0 samples of digital silence in audio and digital
5484 silence is detected in all channels at same positions in stream:
5486 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5492 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5493 loudspeakers around the user for binaural listening via headphones (audio
5494 formats up to 9 channels supported).
5495 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5496 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5497 Austrian Academy of Sciences.
5499 To enable compilation of this filter you need to configure FFmpeg with
5500 @code{--enable-libmysofa}.
5502 The filter accepts the following options:
5506 Set the SOFA file used for rendering.
5509 Set gain applied to audio. Value is in dB. Default is 0.
5512 Set rotation of virtual loudspeakers in deg. Default is 0.
5515 Set elevation of virtual speakers in deg. Default is 0.
5518 Set distance in meters between loudspeakers and the listener with near-field
5519 HRTFs. Default is 1.
5522 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5523 processing audio in time domain which is slow.
5524 @var{freq} is processing audio in frequency domain which is fast.
5525 Default is @var{freq}.
5528 Set custom positions of virtual loudspeakers. Syntax for this option is:
5529 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5530 Each virtual loudspeaker is described with short channel name following with
5531 azimuth and elevation in degrees.
5532 Each virtual loudspeaker description is separated by '|'.
5533 For example to override front left and front right channel positions use:
5534 'speakers=FL 45 15|FR 345 15'.
5535 Descriptions with unrecognised channel names are ignored.
5538 Set custom gain for LFE channels. Value is in dB. Default is 0.
5541 Set custom frame size in number of samples. Default is 1024.
5542 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5543 is set to @var{freq}.
5546 Should all IRs be normalized upon importing SOFA file.
5547 By default is enabled.
5550 Should nearest IRs be interpolated with neighbor IRs if exact position
5551 does not match. By default is disabled.
5554 Minphase all IRs upon loading of SOFA file. By default is disabled.
5557 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5560 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5563 @subsection Examples
5567 Using ClubFritz6 sofa file:
5569 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5573 Using ClubFritz12 sofa file and bigger radius with small rotation:
5575 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5579 Similar as above but with custom speaker positions for front left, front right, back left and back right
5580 and also with custom gain:
5582 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5589 This filter expands or compresses each half-cycle of audio samples
5590 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5591 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5593 The filter accepts the following options:
5597 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5598 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5601 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5602 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5603 would be such that local peak value reaches target peak value but never to surpass it and that
5604 ratio between new and previous peak value does not surpass this option value.
5606 @item compression, c
5607 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5608 This option controls maximum local half-cycle of samples compression. This option is used
5609 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5610 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5611 that peak's half-cycle will be compressed by current compression factor.
5614 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5615 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5616 Any half-cycle samples with their local peak value below or same as this option value will be
5617 compressed by current compression factor, otherwise, if greater than threshold value they will be
5618 expanded with expansion factor so that it could reach peak target value but never surpass it.
5621 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5622 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5623 each new half-cycle until it reaches @option{expansion} value.
5624 Setting this options too high may lead to distortions.
5627 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5628 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5629 each new half-cycle until it reaches @option{compression} value.
5632 Specify which channels to filter, by default all available channels are filtered.
5635 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5636 option. When enabled any half-cycle of samples with their local peak value below or same as
5637 @option{threshold} option will be expanded otherwise it will be compressed.
5640 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5641 When disabled each filtered channel gain calculation is independent, otherwise when this option
5642 is enabled the minimum of all possible gains for each filtered channel is used.
5645 @subsection Commands
5647 This filter supports the all above options as @ref{commands}.
5649 @section stereotools
5651 This filter has some handy utilities to manage stereo signals, for converting
5652 M/S stereo recordings to L/R signal while having control over the parameters
5653 or spreading the stereo image of master track.
5655 The filter accepts the following options:
5659 Set input level before filtering for both channels. Defaults is 1.
5660 Allowed range is from 0.015625 to 64.
5663 Set output level after filtering for both channels. Defaults is 1.
5664 Allowed range is from 0.015625 to 64.
5667 Set input balance between both channels. Default is 0.
5668 Allowed range is from -1 to 1.
5671 Set output balance between both channels. Default is 0.
5672 Allowed range is from -1 to 1.
5675 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5676 clipping. Disabled by default.
5679 Mute the left channel. Disabled by default.
5682 Mute the right channel. Disabled by default.
5685 Change the phase of the left channel. Disabled by default.
5688 Change the phase of the right channel. Disabled by default.
5691 Set stereo mode. Available values are:
5695 Left/Right to Left/Right, this is default.
5698 Left/Right to Mid/Side.
5701 Mid/Side to Left/Right.
5704 Left/Right to Left/Left.
5707 Left/Right to Right/Right.
5710 Left/Right to Left + Right.
5713 Left/Right to Right/Left.
5716 Mid/Side to Left/Left.
5719 Mid/Side to Right/Right.
5722 Mid/Side to Right/Left.
5725 Left/Right to Left - Right.
5729 Set level of side signal. Default is 1.
5730 Allowed range is from 0.015625 to 64.
5733 Set balance of side signal. Default is 0.
5734 Allowed range is from -1 to 1.
5737 Set level of the middle signal. Default is 1.
5738 Allowed range is from 0.015625 to 64.
5741 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5744 Set stereo base between mono and inversed channels. Default is 0.
5745 Allowed range is from -1 to 1.
5748 Set delay in milliseconds how much to delay left from right channel and
5749 vice versa. Default is 0. Allowed range is from -20 to 20.
5752 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5755 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5757 @item bmode_in, bmode_out
5758 Set balance mode for balance_in/balance_out option.
5760 Can be one of the following:
5764 Classic balance mode. Attenuate one channel at time.
5765 Gain is raised up to 1.
5768 Similar as classic mode above but gain is raised up to 2.
5771 Equal power distribution, from -6dB to +6dB range.
5775 @subsection Commands
5777 This filter supports the all above options as @ref{commands}.
5779 @subsection Examples
5783 Apply karaoke like effect:
5785 stereotools=mlev=0.015625
5789 Convert M/S signal to L/R:
5791 "stereotools=mode=ms>lr"
5795 @section stereowiden
5797 This filter enhance the stereo effect by suppressing signal common to both
5798 channels and by delaying the signal of left into right and vice versa,
5799 thereby widening the stereo effect.
5801 The filter accepts the following options:
5805 Time in milliseconds of the delay of left signal into right and vice versa.
5806 Default is 20 milliseconds.
5809 Amount of gain in delayed signal into right and vice versa. Gives a delay
5810 effect of left signal in right output and vice versa which gives widening
5811 effect. Default is 0.3.
5814 Cross feed of left into right with inverted phase. This helps in suppressing
5815 the mono. If the value is 1 it will cancel all the signal common to both
5816 channels. Default is 0.3.
5819 Set level of input signal of original channel. Default is 0.8.
5822 @subsection Commands
5824 This filter supports the all above options except @code{delay} as @ref{commands}.
5826 @section superequalizer
5827 Apply 18 band equalizer.
5829 The filter accepts the following options:
5836 Set 131Hz band gain.
5838 Set 185Hz band gain.
5840 Set 262Hz band gain.
5842 Set 370Hz band gain.
5844 Set 523Hz band gain.
5846 Set 740Hz band gain.
5848 Set 1047Hz band gain.
5850 Set 1480Hz band gain.
5852 Set 2093Hz band gain.
5854 Set 2960Hz band gain.
5856 Set 4186Hz band gain.
5858 Set 5920Hz band gain.
5860 Set 8372Hz band gain.
5862 Set 11840Hz band gain.
5864 Set 16744Hz band gain.
5866 Set 20000Hz band gain.
5870 Apply audio surround upmix filter.
5872 This filter allows to produce multichannel output from audio stream.
5874 The filter accepts the following options:
5878 Set output channel layout. By default, this is @var{5.1}.
5880 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5881 for the required syntax.
5884 Set input channel layout. By default, this is @var{stereo}.
5886 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5887 for the required syntax.
5890 Set input volume level. By default, this is @var{1}.
5893 Set output volume level. By default, this is @var{1}.
5896 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5899 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5902 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5905 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5906 In @var{add} mode, LFE channel is created from input audio and added to output.
5907 In @var{sub} mode, LFE channel is created from input audio and added to output but
5908 also all non-LFE output channels are subtracted with output LFE channel.
5911 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5912 Default is @var{90}.
5915 Set front center input volume. By default, this is @var{1}.
5918 Set front center output volume. By default, this is @var{1}.
5921 Set front left input volume. By default, this is @var{1}.
5924 Set front left output volume. By default, this is @var{1}.
5927 Set front right input volume. By default, this is @var{1}.
5930 Set front right output volume. By default, this is @var{1}.
5933 Set side left input volume. By default, this is @var{1}.
5936 Set side left output volume. By default, this is @var{1}.
5939 Set side right input volume. By default, this is @var{1}.
5942 Set side right output volume. By default, this is @var{1}.
5945 Set back left input volume. By default, this is @var{1}.
5948 Set back left output volume. By default, this is @var{1}.
5951 Set back right input volume. By default, this is @var{1}.
5954 Set back right output volume. By default, this is @var{1}.
5957 Set back center input volume. By default, this is @var{1}.
5960 Set back center output volume. By default, this is @var{1}.
5963 Set LFE input volume. By default, this is @var{1}.
5966 Set LFE output volume. By default, this is @var{1}.
5969 Set spread usage of stereo image across X axis for all channels.
5972 Set spread usage of stereo image across Y axis for all channels.
5974 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5975 Set spread usage of stereo image across X axis for each channel.
5977 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5978 Set spread usage of stereo image across Y axis for each channel.
5981 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5984 Set window function.
5986 It accepts the following values:
6009 Default is @code{hann}.
6012 Set window overlap. If set to 1, the recommended overlap for selected
6013 window function will be picked. Default is @code{0.5}.
6016 @section treble, highshelf
6018 Boost or cut treble (upper) frequencies of the audio using a two-pole
6019 shelving filter with a response similar to that of a standard
6020 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6022 The filter accepts the following options:
6026 Give the gain at whichever is the lower of ~22 kHz and the
6027 Nyquist frequency. Its useful range is about -20 (for a large cut)
6028 to +20 (for a large boost). Beware of clipping when using a positive gain.
6031 Set the filter's central frequency and so can be used
6032 to extend or reduce the frequency range to be boosted or cut.
6033 The default value is @code{3000} Hz.
6036 Set method to specify band-width of filter.
6051 Determine how steep is the filter's shelf transition.
6054 How much to use filtered signal in output. Default is 1.
6055 Range is between 0 and 1.
6058 Specify which channels to filter, by default all available are filtered.
6061 Normalize biquad coefficients, by default is disabled.
6062 Enabling it will normalize magnitude response at DC to 0dB.
6065 Set transform type of IIR filter.
6074 Set precison of filtering.
6077 Pick automatic sample format depending on surround filters.
6079 Always use signed 16-bit.
6081 Always use signed 32-bit.
6083 Always use float 32-bit.
6085 Always use float 64-bit.
6089 @subsection Commands
6091 This filter supports the following commands:
6094 Change treble frequency.
6095 Syntax for the command is : "@var{frequency}"
6098 Change treble width_type.
6099 Syntax for the command is : "@var{width_type}"
6102 Change treble width.
6103 Syntax for the command is : "@var{width}"
6107 Syntax for the command is : "@var{gain}"
6111 Syntax for the command is : "@var{mix}"
6116 Sinusoidal amplitude modulation.
6118 The filter accepts the following options:
6122 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6123 (20 Hz or lower) will result in a tremolo effect.
6124 This filter may also be used as a ring modulator by specifying
6125 a modulation frequency higher than 20 Hz.
6126 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6129 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6130 Default value is 0.5.
6135 Sinusoidal phase modulation.
6137 The filter accepts the following options:
6141 Modulation frequency in Hertz.
6142 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6145 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6146 Default value is 0.5.
6151 Adjust the input audio volume.
6153 It accepts the following parameters:
6157 Set audio volume expression.
6159 Output values are clipped to the maximum value.
6161 The output audio volume is given by the relation:
6163 @var{output_volume} = @var{volume} * @var{input_volume}
6166 The default value for @var{volume} is "1.0".
6169 This parameter represents the mathematical precision.
6171 It determines which input sample formats will be allowed, which affects the
6172 precision of the volume scaling.
6176 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6178 32-bit floating-point; this limits input sample format to FLT. (default)
6180 64-bit floating-point; this limits input sample format to DBL.
6184 Choose the behaviour on encountering ReplayGain side data in input frames.
6188 Remove ReplayGain side data, ignoring its contents (the default).
6191 Ignore ReplayGain side data, but leave it in the frame.
6194 Prefer the track gain, if present.
6197 Prefer the album gain, if present.
6200 @item replaygain_preamp
6201 Pre-amplification gain in dB to apply to the selected replaygain gain.
6203 Default value for @var{replaygain_preamp} is 0.0.
6205 @item replaygain_noclip
6206 Prevent clipping by limiting the gain applied.
6208 Default value for @var{replaygain_noclip} is 1.
6211 Set when the volume expression is evaluated.
6213 It accepts the following values:
6216 only evaluate expression once during the filter initialization, or
6217 when the @samp{volume} command is sent
6220 evaluate expression for each incoming frame
6223 Default value is @samp{once}.
6226 The volume expression can contain the following parameters.
6230 frame number (starting at zero)
6233 @item nb_consumed_samples
6234 number of samples consumed by the filter
6236 number of samples in the current frame
6238 original frame position in the file
6244 PTS at start of stream
6246 time at start of stream
6252 last set volume value
6255 Note that when @option{eval} is set to @samp{once} only the
6256 @var{sample_rate} and @var{tb} variables are available, all other
6257 variables will evaluate to NAN.
6259 @subsection Commands
6261 This filter supports the following commands:
6264 Modify the volume expression.
6265 The command accepts the same syntax of the corresponding option.
6267 If the specified expression is not valid, it is kept at its current
6271 @subsection Examples
6275 Halve the input audio volume:
6279 volume=volume=-6.0206dB
6282 In all the above example the named key for @option{volume} can be
6283 omitted, for example like in:
6289 Increase input audio power by 6 decibels using fixed-point precision:
6291 volume=volume=6dB:precision=fixed
6295 Fade volume after time 10 with an annihilation period of 5 seconds:
6297 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6301 @section volumedetect
6303 Detect the volume of the input video.
6305 The filter has no parameters. The input is not modified. Statistics about
6306 the volume will be printed in the log when the input stream end is reached.
6308 In particular it will show the mean volume (root mean square), maximum
6309 volume (on a per-sample basis), and the beginning of a histogram of the
6310 registered volume values (from the maximum value to a cumulated 1/1000 of
6313 All volumes are in decibels relative to the maximum PCM value.
6315 @subsection Examples
6317 Here is an excerpt of the output:
6319 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6320 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6321 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6322 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6323 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6324 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6325 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6326 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6327 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6333 The mean square energy is approximately -27 dB, or 10^-2.7.
6335 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6337 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6340 In other words, raising the volume by +4 dB does not cause any clipping,
6341 raising it by +5 dB causes clipping for 6 samples, etc.
6343 @c man end AUDIO FILTERS
6345 @chapter Audio Sources
6346 @c man begin AUDIO SOURCES
6348 Below is a description of the currently available audio sources.
6352 Buffer audio frames, and make them available to the filter chain.
6354 This source is mainly intended for a programmatic use, in particular
6355 through the interface defined in @file{libavfilter/buffersrc.h}.
6357 It accepts the following parameters:
6361 The timebase which will be used for timestamps of submitted frames. It must be
6362 either a floating-point number or in @var{numerator}/@var{denominator} form.
6365 The sample rate of the incoming audio buffers.
6368 The sample format of the incoming audio buffers.
6369 Either a sample format name or its corresponding integer representation from
6370 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6372 @item channel_layout
6373 The channel layout of the incoming audio buffers.
6374 Either a channel layout name from channel_layout_map in
6375 @file{libavutil/channel_layout.c} or its corresponding integer representation
6376 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6379 The number of channels of the incoming audio buffers.
6380 If both @var{channels} and @var{channel_layout} are specified, then they
6385 @subsection Examples
6388 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6391 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6392 Since the sample format with name "s16p" corresponds to the number
6393 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6396 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6401 Generate an audio signal specified by an expression.
6403 This source accepts in input one or more expressions (one for each
6404 channel), which are evaluated and used to generate a corresponding
6407 This source accepts the following options:
6411 Set the '|'-separated expressions list for each separate channel. In case the
6412 @option{channel_layout} option is not specified, the selected channel layout
6413 depends on the number of provided expressions. Otherwise the last
6414 specified expression is applied to the remaining output channels.
6416 @item channel_layout, c
6417 Set the channel layout. The number of channels in the specified layout
6418 must be equal to the number of specified expressions.
6421 Set the minimum duration of the sourced audio. See
6422 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6423 for the accepted syntax.
6424 Note that the resulting duration may be greater than the specified
6425 duration, as the generated audio is always cut at the end of a
6428 If not specified, or the expressed duration is negative, the audio is
6429 supposed to be generated forever.
6432 Set the number of samples per channel per each output frame,
6435 @item sample_rate, s
6436 Specify the sample rate, default to 44100.
6439 Each expression in @var{exprs} can contain the following constants:
6443 number of the evaluated sample, starting from 0
6446 time of the evaluated sample expressed in seconds, starting from 0
6453 @subsection Examples
6463 Generate a sin signal with frequency of 440 Hz, set sample rate to
6466 aevalsrc="sin(440*2*PI*t):s=8000"
6470 Generate a two channels signal, specify the channel layout (Front
6471 Center + Back Center) explicitly:
6473 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6477 Generate white noise:
6479 aevalsrc="-2+random(0)"
6483 Generate an amplitude modulated signal:
6485 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6489 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6491 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6498 Generate a FIR coefficients using frequency sampling method.
6500 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6502 The filter accepts the following options:
6506 Set number of filter coefficents in output audio stream.
6507 Default value is 1025.
6510 Set frequency points from where magnitude and phase are set.
6511 This must be in non decreasing order, and first element must be 0, while last element
6512 must be 1. Elements are separated by white spaces.
6515 Set magnitude value for every frequency point set by @option{frequency}.
6516 Number of values must be same as number of frequency points.
6517 Values are separated by white spaces.
6520 Set phase value for every frequency point set by @option{frequency}.
6521 Number of values must be same as number of frequency points.
6522 Values are separated by white spaces.
6524 @item sample_rate, r
6525 Set sample rate, default is 44100.
6528 Set number of samples per each frame. Default is 1024.
6531 Set window function. Default is blackman.
6536 The null audio source, return unprocessed audio frames. It is mainly useful
6537 as a template and to be employed in analysis / debugging tools, or as
6538 the source for filters which ignore the input data (for example the sox
6541 This source accepts the following options:
6545 @item channel_layout, cl
6547 Specifies the channel layout, and can be either an integer or a string
6548 representing a channel layout. The default value of @var{channel_layout}
6551 Check the channel_layout_map definition in
6552 @file{libavutil/channel_layout.c} for the mapping between strings and
6553 channel layout values.
6555 @item sample_rate, r
6556 Specifies the sample rate, and defaults to 44100.
6559 Set the number of samples per requested frames.
6562 Set the duration of the sourced audio. See
6563 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6564 for the accepted syntax.
6566 If not specified, or the expressed duration is negative, the audio is
6567 supposed to be generated forever.
6570 @subsection Examples
6574 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6576 anullsrc=r=48000:cl=4
6580 Do the same operation with a more obvious syntax:
6582 anullsrc=r=48000:cl=mono
6586 All the parameters need to be explicitly defined.
6590 Synthesize a voice utterance using the libflite library.
6592 To enable compilation of this filter you need to configure FFmpeg with
6593 @code{--enable-libflite}.
6595 Note that versions of the flite library prior to 2.0 are not thread-safe.
6597 The filter accepts the following options:
6602 If set to 1, list the names of the available voices and exit
6603 immediately. Default value is 0.
6606 Set the maximum number of samples per frame. Default value is 512.
6609 Set the filename containing the text to speak.
6612 Set the text to speak.
6615 Set the voice to use for the speech synthesis. Default value is
6616 @code{kal}. See also the @var{list_voices} option.
6619 @subsection Examples
6623 Read from file @file{speech.txt}, and synthesize the text using the
6624 standard flite voice:
6626 flite=textfile=speech.txt
6630 Read the specified text selecting the @code{slt} voice:
6632 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6636 Input text to ffmpeg:
6638 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6642 Make @file{ffplay} speak the specified text, using @code{flite} and
6643 the @code{lavfi} device:
6645 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6649 For more information about libflite, check:
6650 @url{http://www.festvox.org/flite/}
6654 Generate a noise audio signal.
6656 The filter accepts the following options:
6659 @item sample_rate, r
6660 Specify the sample rate. Default value is 48000 Hz.
6663 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6667 Specify the duration of the generated audio stream. Not specifying this option
6668 results in noise with an infinite length.
6670 @item color, colour, c
6671 Specify the color of noise. Available noise colors are white, pink, brown,
6672 blue, violet and velvet. Default color is white.
6675 Specify a value used to seed the PRNG.
6678 Set the number of samples per each output frame, default is 1024.
6681 @subsection Examples
6686 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6688 anoisesrc=d=60:c=pink:r=44100:a=0.5
6694 Generate odd-tap Hilbert transform FIR coefficients.
6696 The resulting stream can be used with @ref{afir} filter for phase-shifting
6697 the signal by 90 degrees.
6699 This is used in many matrix coding schemes and for analytic signal generation.
6700 The process is often written as a multiplication by i (or j), the imaginary unit.
6702 The filter accepts the following options:
6706 @item sample_rate, s
6707 Set sample rate, default is 44100.
6710 Set length of FIR filter, default is 22051.
6713 Set number of samples per each frame.
6716 Set window function to be used when generating FIR coefficients.
6721 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6723 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6725 The filter accepts the following options:
6728 @item sample_rate, r
6729 Set sample rate, default is 44100.
6732 Set number of samples per each frame. Default is 1024.
6735 Set high-pass frequency. Default is 0.
6738 Set low-pass frequency. Default is 0.
6739 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6740 is higher than 0 then filter will create band-pass filter coefficients,
6741 otherwise band-reject filter coefficients.
6744 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6747 Set Kaiser window beta.
6750 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6753 Enable rounding, by default is disabled.
6756 Set number of taps for high-pass filter.
6759 Set number of taps for low-pass filter.
6764 Generate an audio signal made of a sine wave with amplitude 1/8.
6766 The audio signal is bit-exact.
6768 The filter accepts the following options:
6773 Set the carrier frequency. Default is 440 Hz.
6775 @item beep_factor, b
6776 Enable a periodic beep every second with frequency @var{beep_factor} times
6777 the carrier frequency. Default is 0, meaning the beep is disabled.
6779 @item sample_rate, r
6780 Specify the sample rate, default is 44100.
6783 Specify the duration of the generated audio stream.
6785 @item samples_per_frame
6786 Set the number of samples per output frame.
6788 The expression can contain the following constants:
6792 The (sequential) number of the output audio frame, starting from 0.
6795 The PTS (Presentation TimeStamp) of the output audio frame,
6796 expressed in @var{TB} units.
6799 The PTS of the output audio frame, expressed in seconds.
6802 The timebase of the output audio frames.
6805 Default is @code{1024}.
6808 @subsection Examples
6813 Generate a simple 440 Hz sine wave:
6819 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6823 sine=frequency=220:beep_factor=4:duration=5
6827 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6830 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6834 @c man end AUDIO SOURCES
6836 @chapter Audio Sinks
6837 @c man begin AUDIO SINKS
6839 Below is a description of the currently available audio sinks.
6841 @section abuffersink
6843 Buffer audio frames, and make them available to the end of filter chain.
6845 This sink is mainly intended for programmatic use, in particular
6846 through the interface defined in @file{libavfilter/buffersink.h}
6847 or the options system.
6849 It accepts a pointer to an AVABufferSinkContext structure, which
6850 defines the incoming buffers' formats, to be passed as the opaque
6851 parameter to @code{avfilter_init_filter} for initialization.
6854 Null audio sink; do absolutely nothing with the input audio. It is
6855 mainly useful as a template and for use in analysis / debugging
6858 @c man end AUDIO SINKS
6860 @chapter Video Filters
6861 @c man begin VIDEO FILTERS
6863 When you configure your FFmpeg build, you can disable any of the
6864 existing filters using @code{--disable-filters}.
6865 The configure output will show the video filters included in your
6868 Below is a description of the currently available video filters.
6872 Mark a region of interest in a video frame.
6874 The frame data is passed through unchanged, but metadata is attached
6875 to the frame indicating regions of interest which can affect the
6876 behaviour of later encoding. Multiple regions can be marked by
6877 applying the filter multiple times.
6881 Region distance in pixels from the left edge of the frame.
6883 Region distance in pixels from the top edge of the frame.
6885 Region width in pixels.
6887 Region height in pixels.
6889 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6890 and may contain the following variables:
6893 Width of the input frame.
6895 Height of the input frame.
6899 Quantisation offset to apply within the region.
6901 This must be a real value in the range -1 to +1. A value of zero
6902 indicates no quality change. A negative value asks for better quality
6903 (less quantisation), while a positive value asks for worse quality
6904 (greater quantisation).
6906 The range is calibrated so that the extreme values indicate the
6907 largest possible offset - if the rest of the frame is encoded with the
6908 worst possible quality, an offset of -1 indicates that this region
6909 should be encoded with the best possible quality anyway. Intermediate
6910 values are then interpolated in some codec-dependent way.
6912 For example, in 10-bit H.264 the quantisation parameter varies between
6913 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6914 this region should be encoded with a QP around one-tenth of the full
6915 range better than the rest of the frame. So, if most of the frame
6916 were to be encoded with a QP of around 30, this region would get a QP
6917 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6918 An extreme value of -1 would indicate that this region should be
6919 encoded with the best possible quality regardless of the treatment of
6920 the rest of the frame - that is, should be encoded at a QP of -12.
6922 If set to true, remove any existing regions of interest marked on the
6923 frame before adding the new one.
6926 @subsection Examples
6930 Mark the centre quarter of the frame as interesting.
6932 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6935 Mark the 100-pixel-wide region on the left edge of the frame as very
6936 uninteresting (to be encoded at much lower quality than the rest of
6939 addroi=0:0:100:ih:+1/5
6943 @section alphaextract
6945 Extract the alpha component from the input as a grayscale video. This
6946 is especially useful with the @var{alphamerge} filter.
6950 Add or replace the alpha component of the primary input with the
6951 grayscale value of a second input. This is intended for use with
6952 @var{alphaextract} to allow the transmission or storage of frame
6953 sequences that have alpha in a format that doesn't support an alpha
6956 For example, to reconstruct full frames from a normal YUV-encoded video
6957 and a separate video created with @var{alphaextract}, you might use:
6959 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6964 Amplify differences between current pixel and pixels of adjacent frames in
6965 same pixel location.
6967 This filter accepts the following options:
6971 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6972 For example radius of 3 will instruct filter to calculate average of 7 frames.
6975 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6978 Set threshold for difference amplification. Any difference greater or equal to
6979 this value will not alter source pixel. Default is 10.
6980 Allowed range is from 0 to 65535.
6983 Set tolerance for difference amplification. Any difference lower to
6984 this value will not alter source pixel. Default is 0.
6985 Allowed range is from 0 to 65535.
6988 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6989 This option controls maximum possible value that will decrease source pixel value.
6992 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6993 This option controls maximum possible value that will increase source pixel value.
6996 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6999 @subsection Commands
7001 This filter supports the following @ref{commands} that corresponds to option of same name:
7013 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7014 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7015 Substation Alpha) subtitles files.
7017 This filter accepts the following option in addition to the common options from
7018 the @ref{subtitles} filter:
7022 Set the shaping engine
7024 Available values are:
7027 The default libass shaping engine, which is the best available.
7029 Fast, font-agnostic shaper that can do only substitutions
7031 Slower shaper using OpenType for substitutions and positioning
7034 The default is @code{auto}.
7038 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7040 The filter accepts the following options:
7044 Set threshold A for 1st plane. Default is 0.02.
7045 Valid range is 0 to 0.3.
7048 Set threshold B for 1st plane. Default is 0.04.
7049 Valid range is 0 to 5.
7052 Set threshold A for 2nd plane. Default is 0.02.
7053 Valid range is 0 to 0.3.
7056 Set threshold B for 2nd plane. Default is 0.04.
7057 Valid range is 0 to 5.
7060 Set threshold A for 3rd plane. Default is 0.02.
7061 Valid range is 0 to 0.3.
7064 Set threshold B for 3rd plane. Default is 0.04.
7065 Valid range is 0 to 5.
7067 Threshold A is designed to react on abrupt changes in the input signal and
7068 threshold B is designed to react on continuous changes in the input signal.
7071 Set number of frames filter will use for averaging. Default is 9. Must be odd
7072 number in range [5, 129].
7075 Set what planes of frame filter will use for averaging. Default is all.
7078 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7079 Alternatively can be set to @code{s} serial.
7081 Parallel can be faster then serial, while other way around is never true.
7082 Parallel will abort early on first change being greater then thresholds, while serial
7083 will continue processing other side of frames if they are equal or below thresholds.
7086 @subsection Commands
7087 This filter supports same @ref{commands} as options except option @code{s}.
7088 The command accepts the same syntax of the corresponding option.
7092 Apply average blur filter.
7094 The filter accepts the following options:
7098 Set horizontal radius size.
7101 Set which planes to filter. By default all planes are filtered.
7104 Set vertical radius size, if zero it will be same as @code{sizeX}.
7105 Default is @code{0}.
7108 @subsection Commands
7109 This filter supports same commands as options.
7110 The command accepts the same syntax of the corresponding option.
7112 If the specified expression is not valid, it is kept at its current
7117 Compute the bounding box for the non-black pixels in the input frame
7120 This filter computes the bounding box containing all the pixels with a
7121 luminance value greater than the minimum allowed value.
7122 The parameters describing the bounding box are printed on the filter
7125 The filter accepts the following option:
7129 Set the minimal luminance value. Default is @code{16}.
7133 Apply bilateral filter, spatial smoothing while preserving edges.
7135 The filter accepts the following options:
7138 Set sigma of gaussian function to calculate spatial weight.
7139 Allowed range is 0 to 512. Default is 0.1.
7142 Set sigma of gaussian function to calculate range weight.
7143 Allowed range is 0 to 1. Default is 0.1.
7146 Set planes to filter. Default is first only.
7149 @section bitplanenoise
7151 Show and measure bit plane noise.
7153 The filter accepts the following options:
7157 Set which plane to analyze. Default is @code{1}.
7160 Filter out noisy pixels from @code{bitplane} set above.
7161 Default is disabled.
7164 @section blackdetect
7166 Detect video intervals that are (almost) completely black. Can be
7167 useful to detect chapter transitions, commercials, or invalid
7170 The filter outputs its detection analysis to both the log as well as
7171 frame metadata. If a black segment of at least the specified minimum
7172 duration is found, a line with the start and end timestamps as well
7173 as duration is printed to the log with level @code{info}. In addition,
7174 a log line with level @code{debug} is printed per frame showing the
7175 black amount detected for that frame.
7177 The filter also attaches metadata to the first frame of a black
7178 segment with key @code{lavfi.black_start} and to the first frame
7179 after the black segment ends with key @code{lavfi.black_end}. The
7180 value is the frame's timestamp. This metadata is added regardless
7181 of the minimum duration specified.
7183 The filter accepts the following options:
7186 @item black_min_duration, d
7187 Set the minimum detected black duration expressed in seconds. It must
7188 be a non-negative floating point number.
7190 Default value is 2.0.
7192 @item picture_black_ratio_th, pic_th
7193 Set the threshold for considering a picture "black".
7194 Express the minimum value for the ratio:
7196 @var{nb_black_pixels} / @var{nb_pixels}
7199 for which a picture is considered black.
7200 Default value is 0.98.
7202 @item pixel_black_th, pix_th
7203 Set the threshold for considering a pixel "black".
7205 The threshold expresses the maximum pixel luminance value for which a
7206 pixel is considered "black". The provided value is scaled according to
7207 the following equation:
7209 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7212 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7213 the input video format, the range is [0-255] for YUV full-range
7214 formats and [16-235] for YUV non full-range formats.
7216 Default value is 0.10.
7219 The following example sets the maximum pixel threshold to the minimum
7220 value, and detects only black intervals of 2 or more seconds:
7222 blackdetect=d=2:pix_th=0.00
7227 Detect frames that are (almost) completely black. Can be useful to
7228 detect chapter transitions or commercials. Output lines consist of
7229 the frame number of the detected frame, the percentage of blackness,
7230 the position in the file if known or -1 and the timestamp in seconds.
7232 In order to display the output lines, you need to set the loglevel at
7233 least to the AV_LOG_INFO value.
7235 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7236 The value represents the percentage of pixels in the picture that
7237 are below the threshold value.
7239 It accepts the following parameters:
7244 The percentage of the pixels that have to be below the threshold; it defaults to
7247 @item threshold, thresh
7248 The threshold below which a pixel value is considered black; it defaults to
7256 Blend two video frames into each other.
7258 The @code{blend} filter takes two input streams and outputs one
7259 stream, the first input is the "top" layer and second input is
7260 "bottom" layer. By default, the output terminates when the longest input terminates.
7262 The @code{tblend} (time blend) filter takes two consecutive frames
7263 from one single stream, and outputs the result obtained by blending
7264 the new frame on top of the old frame.
7266 A description of the accepted options follows.
7274 Set blend mode for specific pixel component or all pixel components in case
7275 of @var{all_mode}. Default value is @code{normal}.
7277 Available values for component modes are:
7319 Set blend opacity for specific pixel component or all pixel components in case
7320 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7327 Set blend expression for specific pixel component or all pixel components in case
7328 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7330 The expressions can use the following variables:
7334 The sequential number of the filtered frame, starting from @code{0}.
7338 the coordinates of the current sample
7342 the width and height of currently filtered plane
7346 Width and height scale for the plane being filtered. It is the
7347 ratio between the dimensions of the current plane to the luma plane,
7348 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7349 the luma plane and @code{0.5,0.5} for the chroma planes.
7352 Time of the current frame, expressed in seconds.
7355 Value of pixel component at current location for first video frame (top layer).
7358 Value of pixel component at current location for second video frame (bottom layer).
7362 The @code{blend} filter also supports the @ref{framesync} options.
7364 @subsection Examples
7368 Apply transition from bottom layer to top layer in first 10 seconds:
7370 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7374 Apply linear horizontal transition from top layer to bottom layer:
7376 blend=all_expr='A*(X/W)+B*(1-X/W)'
7380 Apply 1x1 checkerboard effect:
7382 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7386 Apply uncover left effect:
7388 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7392 Apply uncover down effect:
7394 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7398 Apply uncover up-left effect:
7400 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7404 Split diagonally video and shows top and bottom layer on each side:
7406 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7410 Display differences between the current and the previous frame:
7412 tblend=all_mode=grainextract
7418 Denoise frames using Block-Matching 3D algorithm.
7420 The filter accepts the following options.
7424 Set denoising strength. Default value is 1.
7425 Allowed range is from 0 to 999.9.
7426 The denoising algorithm is very sensitive to sigma, so adjust it
7427 according to the source.
7430 Set local patch size. This sets dimensions in 2D.
7433 Set sliding step for processing blocks. Default value is 4.
7434 Allowed range is from 1 to 64.
7435 Smaller values allows processing more reference blocks and is slower.
7438 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7439 When set to 1, no block matching is done. Larger values allows more blocks
7441 Allowed range is from 1 to 256.
7444 Set radius for search block matching. Default is 9.
7445 Allowed range is from 1 to INT32_MAX.
7448 Set step between two search locations for block matching. Default is 1.
7449 Allowed range is from 1 to 64. Smaller is slower.
7452 Set threshold of mean square error for block matching. Valid range is 0 to
7456 Set thresholding parameter for hard thresholding in 3D transformed domain.
7457 Larger values results in stronger hard-thresholding filtering in frequency
7461 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7462 Default is @code{basic}.
7465 If enabled, filter will use 2nd stream for block matching.
7466 Default is disabled for @code{basic} value of @var{estim} option,
7467 and always enabled if value of @var{estim} is @code{final}.
7470 Set planes to filter. Default is all available except alpha.
7473 @subsection Examples
7477 Basic filtering with bm3d:
7479 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7483 Same as above, but filtering only luma:
7485 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7489 Same as above, but with both estimation modes:
7491 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
7495 Same as above, but prefilter with @ref{nlmeans} filter instead:
7497 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
7503 Apply a boxblur algorithm to the input video.
7505 It accepts the following parameters:
7509 @item luma_radius, lr
7510 @item luma_power, lp
7511 @item chroma_radius, cr
7512 @item chroma_power, cp
7513 @item alpha_radius, ar
7514 @item alpha_power, ap
7518 A description of the accepted options follows.
7521 @item luma_radius, lr
7522 @item chroma_radius, cr
7523 @item alpha_radius, ar
7524 Set an expression for the box radius in pixels used for blurring the
7525 corresponding input plane.
7527 The radius value must be a non-negative number, and must not be
7528 greater than the value of the expression @code{min(w,h)/2} for the
7529 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7532 Default value for @option{luma_radius} is "2". If not specified,
7533 @option{chroma_radius} and @option{alpha_radius} default to the
7534 corresponding value set for @option{luma_radius}.
7536 The expressions can contain the following constants:
7540 The input width and height in pixels.
7544 The input chroma image width and height in pixels.
7548 The horizontal and vertical chroma subsample values. For example, for the
7549 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7552 @item luma_power, lp
7553 @item chroma_power, cp
7554 @item alpha_power, ap
7555 Specify how many times the boxblur filter is applied to the
7556 corresponding plane.
7558 Default value for @option{luma_power} is 2. If not specified,
7559 @option{chroma_power} and @option{alpha_power} default to the
7560 corresponding value set for @option{luma_power}.
7562 A value of 0 will disable the effect.
7565 @subsection Examples
7569 Apply a boxblur filter with the luma, chroma, and alpha radii
7572 boxblur=luma_radius=2:luma_power=1
7577 Set the luma radius to 2, and alpha and chroma radius to 0:
7579 boxblur=2:1:cr=0:ar=0
7583 Set the luma and chroma radii to a fraction of the video dimension:
7585 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7591 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7592 Deinterlacing Filter").
7594 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7595 interpolation algorithms.
7596 It accepts the following parameters:
7600 The interlacing mode to adopt. It accepts one of the following values:
7604 Output one frame for each frame.
7606 Output one frame for each field.
7609 The default value is @code{send_field}.
7612 The picture field parity assumed for the input interlaced video. It accepts one
7613 of the following values:
7617 Assume the top field is first.
7619 Assume the bottom field is first.
7621 Enable automatic detection of field parity.
7624 The default value is @code{auto}.
7625 If the interlacing is unknown or the decoder does not export this information,
7626 top field first will be assumed.
7629 Specify which frames to deinterlace. Accepts one of the following
7634 Deinterlace all frames.
7636 Only deinterlace frames marked as interlaced.
7639 The default value is @code{all}.
7644 Apply Contrast Adaptive Sharpen filter to video stream.
7646 The filter accepts the following options:
7650 Set the sharpening strength. Default value is 0.
7653 Set planes to filter. Default value is to filter all
7654 planes except alpha plane.
7658 Remove all color information for all colors except for certain one.
7660 The filter accepts the following options:
7664 The color which will not be replaced with neutral chroma.
7667 Similarity percentage with the above color.
7668 0.01 matches only the exact key color, while 1.0 matches everything.
7672 0.0 makes pixels either fully gray, or not gray at all.
7673 Higher values result in more preserved color.
7676 Signals that the color passed is already in YUV instead of RGB.
7678 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7679 This can be used to pass exact YUV values as hexadecimal numbers.
7682 @subsection Commands
7683 This filter supports same @ref{commands} as options.
7684 The command accepts the same syntax of the corresponding option.
7686 If the specified expression is not valid, it is kept at its current
7690 YUV colorspace color/chroma keying.
7692 The filter accepts the following options:
7696 The color which will be replaced with transparency.
7699 Similarity percentage with the key color.
7701 0.01 matches only the exact key color, while 1.0 matches everything.
7706 0.0 makes pixels either fully transparent, or not transparent at all.
7708 Higher values result in semi-transparent pixels, with a higher transparency
7709 the more similar the pixels color is to the key color.
7712 Signals that the color passed is already in YUV instead of RGB.
7714 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7715 This can be used to pass exact YUV values as hexadecimal numbers.
7718 @subsection Commands
7719 This filter supports same @ref{commands} as options.
7720 The command accepts the same syntax of the corresponding option.
7722 If the specified expression is not valid, it is kept at its current
7725 @subsection Examples
7729 Make every green pixel in the input image transparent:
7731 ffmpeg -i input.png -vf chromakey=green out.png
7735 Overlay a greenscreen-video on top of a static black background.
7737 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
7742 Reduce chrominance noise.
7744 The filter accepts the following options:
7748 Set threshold for averaging chrominance values.
7749 Sum of absolute difference of Y, U and V pixel components of current
7750 pixel and neighbour pixels lower than this threshold will be used in
7751 averaging. Luma component is left unchanged and is copied to output.
7752 Default value is 30. Allowed range is from 1 to 200.
7755 Set horizontal radius of rectangle used for averaging.
7756 Allowed range is from 1 to 100. Default value is 5.
7759 Set vertical radius of rectangle used for averaging.
7760 Allowed range is from 1 to 100. Default value is 5.
7763 Set horizontal step when averaging. Default value is 1.
7764 Allowed range is from 1 to 50.
7765 Mostly useful to speed-up filtering.
7768 Set vertical step when averaging. Default value is 1.
7769 Allowed range is from 1 to 50.
7770 Mostly useful to speed-up filtering.
7773 Set Y threshold for averaging chrominance values.
7774 Set finer control for max allowed difference between Y components
7775 of current pixel and neigbour pixels.
7776 Default value is 200. Allowed range is from 1 to 200.
7779 Set U threshold for averaging chrominance values.
7780 Set finer control for max allowed difference between U components
7781 of current pixel and neigbour pixels.
7782 Default value is 200. Allowed range is from 1 to 200.
7785 Set V threshold for averaging chrominance values.
7786 Set finer control for max allowed difference between V components
7787 of current pixel and neigbour pixels.
7788 Default value is 200. Allowed range is from 1 to 200.
7791 @subsection Commands
7792 This filter supports same @ref{commands} as options.
7793 The command accepts the same syntax of the corresponding option.
7795 @section chromashift
7796 Shift chroma pixels horizontally and/or vertically.
7798 The filter accepts the following options:
7801 Set amount to shift chroma-blue horizontally.
7803 Set amount to shift chroma-blue vertically.
7805 Set amount to shift chroma-red horizontally.
7807 Set amount to shift chroma-red vertically.
7809 Set edge mode, can be @var{smear}, default, or @var{warp}.
7812 @subsection Commands
7814 This filter supports the all above options as @ref{commands}.
7818 Display CIE color diagram with pixels overlaid onto it.
7820 The filter accepts the following options:
7835 @item uhdtv, rec2020
7849 Set what gamuts to draw.
7851 See @code{system} option for available values.
7854 Set ciescope size, by default set to 512.
7857 Set intensity used to map input pixel values to CIE diagram.
7860 Set contrast used to draw tongue colors that are out of active color system gamut.
7863 Correct gamma displayed on scope, by default enabled.
7866 Show white point on CIE diagram, by default disabled.
7869 Set input gamma. Used only with XYZ input color space.
7874 Visualize information exported by some codecs.
7876 Some codecs can export information through frames using side-data or other
7877 means. For example, some MPEG based codecs export motion vectors through the
7878 @var{export_mvs} flag in the codec @option{flags2} option.
7880 The filter accepts the following option:
7884 Set motion vectors to visualize.
7886 Available flags for @var{mv} are:
7890 forward predicted MVs of P-frames
7892 forward predicted MVs of B-frames
7894 backward predicted MVs of B-frames
7898 Display quantization parameters using the chroma planes.
7901 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7903 Available flags for @var{mv_type} are:
7907 forward predicted MVs
7909 backward predicted MVs
7912 @item frame_type, ft
7913 Set frame type to visualize motion vectors of.
7915 Available flags for @var{frame_type} are:
7919 intra-coded frames (I-frames)
7921 predicted frames (P-frames)
7923 bi-directionally predicted frames (B-frames)
7927 @subsection Examples
7931 Visualize forward predicted MVs of all frames using @command{ffplay}:
7933 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7937 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7939 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7943 @section colorbalance
7944 Modify intensity of primary colors (red, green and blue) of input frames.
7946 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7947 regions for the red-cyan, green-magenta or blue-yellow balance.
7949 A positive adjustment value shifts the balance towards the primary color, a negative
7950 value towards the complementary color.
7952 The filter accepts the following options:
7958 Adjust red, green and blue shadows (darkest pixels).
7963 Adjust red, green and blue midtones (medium pixels).
7968 Adjust red, green and blue highlights (brightest pixels).
7970 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7973 Preserve lightness when changing color balance. Default is disabled.
7976 @subsection Examples
7980 Add red color cast to shadows:
7986 @subsection Commands
7988 This filter supports the all above options as @ref{commands}.
7990 @section colorchannelmixer
7992 Adjust video input frames by re-mixing color channels.
7994 This filter modifies a color channel by adding the values associated to
7995 the other channels of the same pixels. For example if the value to
7996 modify is red, the output value will be:
7998 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8001 The filter accepts the following options:
8008 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8009 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8015 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8016 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8022 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8023 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8029 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8030 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8032 Allowed ranges for options are @code{[-2.0, 2.0]}.
8035 @subsection Examples
8039 Convert source to grayscale:
8041 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8044 Simulate sepia tones:
8046 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8050 @subsection Commands
8052 This filter supports the all above options as @ref{commands}.
8055 RGB colorspace color keying.
8057 The filter accepts the following options:
8061 The color which will be replaced with transparency.
8064 Similarity percentage with the key color.
8066 0.01 matches only the exact key color, while 1.0 matches everything.
8071 0.0 makes pixels either fully transparent, or not transparent at all.
8073 Higher values result in semi-transparent pixels, with a higher transparency
8074 the more similar the pixels color is to the key color.
8077 @subsection Examples
8081 Make every green pixel in the input image transparent:
8083 ffmpeg -i input.png -vf colorkey=green out.png
8087 Overlay a greenscreen-video on top of a static background image.
8089 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
8093 @subsection Commands
8094 This filter supports same @ref{commands} as options.
8095 The command accepts the same syntax of the corresponding option.
8097 If the specified expression is not valid, it is kept at its current
8101 Remove all color information for all RGB colors except for certain one.
8103 The filter accepts the following options:
8107 The color which will not be replaced with neutral gray.
8110 Similarity percentage with the above color.
8111 0.01 matches only the exact key color, while 1.0 matches everything.
8114 Blend percentage. 0.0 makes pixels fully gray.
8115 Higher values result in more preserved color.
8118 @subsection Commands
8119 This filter supports same @ref{commands} as options.
8120 The command accepts the same syntax of the corresponding option.
8122 If the specified expression is not valid, it is kept at its current
8125 @section colorlevels
8127 Adjust video input frames using levels.
8129 The filter accepts the following options:
8136 Adjust red, green, blue and alpha input black point.
8137 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8143 Adjust red, green, blue and alpha input white point.
8144 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8146 Input levels are used to lighten highlights (bright tones), darken shadows
8147 (dark tones), change the balance of bright and dark tones.
8153 Adjust red, green, blue and alpha output black point.
8154 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8160 Adjust red, green, blue and alpha output white point.
8161 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8163 Output levels allows manual selection of a constrained output level range.
8166 @subsection Examples
8170 Make video output darker:
8172 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8178 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8182 Make video output lighter:
8184 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8188 Increase brightness:
8190 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8194 @subsection Commands
8196 This filter supports the all above options as @ref{commands}.
8198 @section colormatrix
8200 Convert color matrix.
8202 The filter accepts the following options:
8207 Specify the source and destination color matrix. Both values must be
8210 The accepted values are:
8238 For example to convert from BT.601 to SMPTE-240M, use the command:
8240 colormatrix=bt601:smpte240m
8245 Convert colorspace, transfer characteristics or color primaries.
8246 Input video needs to have an even size.
8248 The filter accepts the following options:
8253 Specify all color properties at once.
8255 The accepted values are:
8285 Specify output colorspace.
8287 The accepted values are:
8296 BT.470BG or BT.601-6 625
8299 SMPTE-170M or BT.601-6 525
8308 BT.2020 with non-constant luminance
8314 Specify output transfer characteristics.
8316 The accepted values are:
8328 Constant gamma of 2.2
8331 Constant gamma of 2.8
8334 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8352 BT.2020 for 10-bits content
8355 BT.2020 for 12-bits content
8361 Specify output color primaries.
8363 The accepted values are:
8372 BT.470BG or BT.601-6 625
8375 SMPTE-170M or BT.601-6 525
8399 Specify output color range.
8401 The accepted values are:
8404 TV (restricted) range
8407 MPEG (restricted) range
8418 Specify output color format.
8420 The accepted values are:
8423 YUV 4:2:0 planar 8-bits
8426 YUV 4:2:0 planar 10-bits
8429 YUV 4:2:0 planar 12-bits
8432 YUV 4:2:2 planar 8-bits
8435 YUV 4:2:2 planar 10-bits
8438 YUV 4:2:2 planar 12-bits
8441 YUV 4:4:4 planar 8-bits
8444 YUV 4:4:4 planar 10-bits
8447 YUV 4:4:4 planar 12-bits
8452 Do a fast conversion, which skips gamma/primary correction. This will take
8453 significantly less CPU, but will be mathematically incorrect. To get output
8454 compatible with that produced by the colormatrix filter, use fast=1.
8457 Specify dithering mode.
8459 The accepted values are:
8465 Floyd-Steinberg dithering
8469 Whitepoint adaptation mode.
8471 The accepted values are:
8474 Bradford whitepoint adaptation
8477 von Kries whitepoint adaptation
8480 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8484 Override all input properties at once. Same accepted values as @ref{all}.
8487 Override input colorspace. Same accepted values as @ref{space}.
8490 Override input color primaries. Same accepted values as @ref{primaries}.
8493 Override input transfer characteristics. Same accepted values as @ref{trc}.
8496 Override input color range. Same accepted values as @ref{range}.
8500 The filter converts the transfer characteristics, color space and color
8501 primaries to the specified user values. The output value, if not specified,
8502 is set to a default value based on the "all" property. If that property is
8503 also not specified, the filter will log an error. The output color range and
8504 format default to the same value as the input color range and format. The
8505 input transfer characteristics, color space, color primaries and color range
8506 should be set on the input data. If any of these are missing, the filter will
8507 log an error and no conversion will take place.
8509 For example to convert the input to SMPTE-240M, use the command:
8511 colorspace=smpte240m
8514 @section convolution
8516 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8518 The filter accepts the following options:
8525 Set matrix for each plane.
8526 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8527 and from 1 to 49 odd number of signed integers in @var{row} mode.
8533 Set multiplier for calculated value for each plane.
8534 If unset or 0, it will be sum of all matrix elements.
8540 Set bias for each plane. This value is added to the result of the multiplication.
8541 Useful for making the overall image brighter or darker. Default is 0.0.
8547 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8548 Default is @var{square}.
8551 @subsection Commands
8553 This filter supports the all above options as @ref{commands}.
8555 @subsection Examples
8561 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"
8567 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"
8573 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"
8579 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"
8583 Apply laplacian edge detector which includes diagonals:
8585 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"
8591 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"
8597 Apply 2D convolution of video stream in frequency domain using second stream
8600 The filter accepts the following options:
8604 Set which planes to process.
8607 Set which impulse video frames will be processed, can be @var{first}
8608 or @var{all}. Default is @var{all}.
8611 The @code{convolve} filter also supports the @ref{framesync} options.
8615 Copy the input video source unchanged to the output. This is mainly useful for
8620 Video filtering on GPU using Apple's CoreImage API on OSX.
8622 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8623 processed by video hardware. However, software-based OpenGL implementations
8624 exist which means there is no guarantee for hardware processing. It depends on
8627 There are many filters and image generators provided by Apple that come with a
8628 large variety of options. The filter has to be referenced by its name along
8631 The coreimage filter accepts the following options:
8634 List all available filters and generators along with all their respective
8635 options as well as possible minimum and maximum values along with the default
8642 Specify all filters by their respective name and options.
8643 Use @var{list_filters} to determine all valid filter names and options.
8644 Numerical options are specified by a float value and are automatically clamped
8645 to their respective value range. Vector and color options have to be specified
8646 by a list of space separated float values. Character escaping has to be done.
8647 A special option name @code{default} is available to use default options for a
8650 It is required to specify either @code{default} or at least one of the filter options.
8651 All omitted options are used with their default values.
8652 The syntax of the filter string is as follows:
8654 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8658 Specify a rectangle where the output of the filter chain is copied into the
8659 input image. It is given by a list of space separated float values:
8661 output_rect=x\ y\ width\ height
8663 If not given, the output rectangle equals the dimensions of the input image.
8664 The output rectangle is automatically cropped at the borders of the input
8665 image. Negative values are valid for each component.
8667 output_rect=25\ 25\ 100\ 100
8671 Several filters can be chained for successive processing without GPU-HOST
8672 transfers allowing for fast processing of complex filter chains.
8673 Currently, only filters with zero (generators) or exactly one (filters) input
8674 image and one output image are supported. Also, transition filters are not yet
8677 Some filters generate output images with additional padding depending on the
8678 respective filter kernel. The padding is automatically removed to ensure the
8679 filter output has the same size as the input image.
8681 For image generators, the size of the output image is determined by the
8682 previous output image of the filter chain or the input image of the whole
8683 filterchain, respectively. The generators do not use the pixel information of
8684 this image to generate their output. However, the generated output is
8685 blended onto this image, resulting in partial or complete coverage of the
8688 The @ref{coreimagesrc} video source can be used for generating input images
8689 which are directly fed into the filter chain. By using it, providing input
8690 images by another video source or an input video is not required.
8692 @subsection Examples
8697 List all filters available:
8699 coreimage=list_filters=true
8703 Use the CIBoxBlur filter with default options to blur an image:
8705 coreimage=filter=CIBoxBlur@@default
8709 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8710 its center at 100x100 and a radius of 50 pixels:
8712 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8716 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8717 given as complete and escaped command-line for Apple's standard bash shell:
8719 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8725 Cover a rectangular object
8727 It accepts the following options:
8731 Filepath of the optional cover image, needs to be in yuv420.
8736 It accepts the following values:
8739 cover it by the supplied image
8741 cover it by interpolating the surrounding pixels
8744 Default value is @var{blur}.
8747 @subsection Examples
8751 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8753 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8759 Crop the input video to given dimensions.
8761 It accepts the following parameters:
8765 The width of the output video. It defaults to @code{iw}.
8766 This expression is evaluated only once during the filter
8767 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8770 The height of the output video. It defaults to @code{ih}.
8771 This expression is evaluated only once during the filter
8772 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8775 The horizontal position, in the input video, of the left edge of the output
8776 video. It defaults to @code{(in_w-out_w)/2}.
8777 This expression is evaluated per-frame.
8780 The vertical position, in the input video, of the top edge of the output video.
8781 It defaults to @code{(in_h-out_h)/2}.
8782 This expression is evaluated per-frame.
8785 If set to 1 will force the output display aspect ratio
8786 to be the same of the input, by changing the output sample aspect
8787 ratio. It defaults to 0.
8790 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8791 width/height/x/y as specified and will not be rounded to nearest smaller value.
8795 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8796 expressions containing the following constants:
8801 The computed values for @var{x} and @var{y}. They are evaluated for
8806 The input width and height.
8810 These are the same as @var{in_w} and @var{in_h}.
8814 The output (cropped) width and height.
8818 These are the same as @var{out_w} and @var{out_h}.
8821 same as @var{iw} / @var{ih}
8824 input sample aspect ratio
8827 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8831 horizontal and vertical chroma subsample values. For example for the
8832 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8835 The number of the input frame, starting from 0.
8838 the position in the file of the input frame, NAN if unknown
8841 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8845 The expression for @var{out_w} may depend on the value of @var{out_h},
8846 and the expression for @var{out_h} may depend on @var{out_w}, but they
8847 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8848 evaluated after @var{out_w} and @var{out_h}.
8850 The @var{x} and @var{y} parameters specify the expressions for the
8851 position of the top-left corner of the output (non-cropped) area. They
8852 are evaluated for each frame. If the evaluated value is not valid, it
8853 is approximated to the nearest valid value.
8855 The expression for @var{x} may depend on @var{y}, and the expression
8856 for @var{y} may depend on @var{x}.
8858 @subsection Examples
8862 Crop area with size 100x100 at position (12,34).
8867 Using named options, the example above becomes:
8869 crop=w=100:h=100:x=12:y=34
8873 Crop the central input area with size 100x100:
8879 Crop the central input area with size 2/3 of the input video:
8881 crop=2/3*in_w:2/3*in_h
8885 Crop the input video central square:
8892 Delimit the rectangle with the top-left corner placed at position
8893 100:100 and the right-bottom corner corresponding to the right-bottom
8894 corner of the input image.
8896 crop=in_w-100:in_h-100:100:100
8900 Crop 10 pixels from the left and right borders, and 20 pixels from
8901 the top and bottom borders
8903 crop=in_w-2*10:in_h-2*20
8907 Keep only the bottom right quarter of the input image:
8909 crop=in_w/2:in_h/2:in_w/2:in_h/2
8913 Crop height for getting Greek harmony:
8915 crop=in_w:1/PHI*in_w
8919 Apply trembling effect:
8921 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)
8925 Apply erratic camera effect depending on timestamp:
8927 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)"
8931 Set x depending on the value of y:
8933 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8937 @subsection Commands
8939 This filter supports the following commands:
8945 Set width/height of the output video and the horizontal/vertical position
8947 The command accepts the same syntax of the corresponding option.
8949 If the specified expression is not valid, it is kept at its current
8955 Auto-detect the crop size.
8957 It calculates the necessary cropping parameters and prints the
8958 recommended parameters via the logging system. The detected dimensions
8959 correspond to the non-black area of the input video.
8961 It accepts the following parameters:
8966 Set higher black value threshold, which can be optionally specified
8967 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8968 value greater to the set value is considered non-black. It defaults to 24.
8969 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8970 on the bitdepth of the pixel format.
8973 The value which the width/height should be divisible by. It defaults to
8974 16. The offset is automatically adjusted to center the video. Use 2 to
8975 get only even dimensions (needed for 4:2:2 video). 16 is best when
8976 encoding to most video codecs.
8979 Set the number of initial frames for which evaluation is skipped.
8980 Default is 2. Range is 0 to INT_MAX.
8982 @item reset_count, reset
8983 Set the counter that determines after how many frames cropdetect will
8984 reset the previously detected largest video area and start over to
8985 detect the current optimal crop area. Default value is 0.
8987 This can be useful when channel logos distort the video area. 0
8988 indicates 'never reset', and returns the largest area encountered during
8995 Delay video filtering until a given wallclock timestamp. The filter first
8996 passes on @option{preroll} amount of frames, then it buffers at most
8997 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8998 it forwards the buffered frames and also any subsequent frames coming in its
9001 The filter can be used synchronize the output of multiple ffmpeg processes for
9002 realtime output devices like decklink. By putting the delay in the filtering
9003 chain and pre-buffering frames the process can pass on data to output almost
9004 immediately after the target wallclock timestamp is reached.
9006 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9012 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9015 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9018 The maximum duration of content to buffer before waiting for the cue expressed
9019 in seconds. Default is 0.
9026 Apply color adjustments using curves.
9028 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9029 component (red, green and blue) has its values defined by @var{N} key points
9030 tied from each other using a smooth curve. The x-axis represents the pixel
9031 values from the input frame, and the y-axis the new pixel values to be set for
9034 By default, a component curve is defined by the two points @var{(0;0)} and
9035 @var{(1;1)}. This creates a straight line where each original pixel value is
9036 "adjusted" to its own value, which means no change to the image.
9038 The filter allows you to redefine these two points and add some more. A new
9039 curve (using a natural cubic spline interpolation) will be define to pass
9040 smoothly through all these new coordinates. The new defined points needs to be
9041 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9042 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9043 the vector spaces, the values will be clipped accordingly.
9045 The filter accepts the following options:
9049 Select one of the available color presets. This option can be used in addition
9050 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9051 options takes priority on the preset values.
9052 Available presets are:
9055 @item color_negative
9058 @item increase_contrast
9060 @item linear_contrast
9061 @item medium_contrast
9063 @item strong_contrast
9066 Default is @code{none}.
9068 Set the master key points. These points will define a second pass mapping. It
9069 is sometimes called a "luminance" or "value" mapping. It can be used with
9070 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9071 post-processing LUT.
9073 Set the key points for the red component.
9075 Set the key points for the green component.
9077 Set the key points for the blue component.
9079 Set the key points for all components (not including master).
9080 Can be used in addition to the other key points component
9081 options. In this case, the unset component(s) will fallback on this
9082 @option{all} setting.
9084 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9086 Save Gnuplot script of the curves in specified file.
9089 To avoid some filtergraph syntax conflicts, each key points list need to be
9090 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9092 @subsection Examples
9096 Increase slightly the middle level of blue:
9098 curves=blue='0/0 0.5/0.58 1/1'
9104 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'
9106 Here we obtain the following coordinates for each components:
9109 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9111 @code{(0;0) (0.50;0.48) (1;1)}
9113 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9117 The previous example can also be achieved with the associated built-in preset:
9119 curves=preset=vintage
9129 Use a Photoshop preset and redefine the points of the green component:
9131 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9135 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9136 and @command{gnuplot}:
9138 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9139 gnuplot -p /tmp/curves.plt
9145 Video data analysis filter.
9147 This filter shows hexadecimal pixel values of part of video.
9149 The filter accepts the following options:
9153 Set output video size.
9156 Set x offset from where to pick pixels.
9159 Set y offset from where to pick pixels.
9162 Set scope mode, can be one of the following:
9165 Draw hexadecimal pixel values with white color on black background.
9168 Draw hexadecimal pixel values with input video pixel color on black
9172 Draw hexadecimal pixel values on color background picked from input video,
9173 the text color is picked in such way so its always visible.
9177 Draw rows and columns numbers on left and top of video.
9180 Set background opacity.
9183 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9187 Apply Directional blur filter.
9189 The filter accepts the following options:
9193 Set angle of directional blur. Default is @code{45}.
9196 Set radius of directional blur. Default is @code{5}.
9199 Set which planes to filter. By default all planes are filtered.
9202 @subsection Commands
9203 This filter supports same @ref{commands} as options.
9204 The command accepts the same syntax of the corresponding option.
9206 If the specified expression is not valid, it is kept at its current
9211 Denoise frames using 2D DCT (frequency domain filtering).
9213 This filter is not designed for real time.
9215 The filter accepts the following options:
9219 Set the noise sigma constant.
9221 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9222 coefficient (absolute value) below this threshold with be dropped.
9224 If you need a more advanced filtering, see @option{expr}.
9226 Default is @code{0}.
9229 Set number overlapping pixels for each block. Since the filter can be slow, you
9230 may want to reduce this value, at the cost of a less effective filter and the
9231 risk of various artefacts.
9233 If the overlapping value doesn't permit processing the whole input width or
9234 height, a warning will be displayed and according borders won't be denoised.
9236 Default value is @var{blocksize}-1, which is the best possible setting.
9239 Set the coefficient factor expression.
9241 For each coefficient of a DCT block, this expression will be evaluated as a
9242 multiplier value for the coefficient.
9244 If this is option is set, the @option{sigma} option will be ignored.
9246 The absolute value of the coefficient can be accessed through the @var{c}
9250 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9251 @var{blocksize}, which is the width and height of the processed blocks.
9253 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9254 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9255 on the speed processing. Also, a larger block size does not necessarily means a
9259 @subsection Examples
9261 Apply a denoise with a @option{sigma} of @code{4.5}:
9266 The same operation can be achieved using the expression system:
9268 dctdnoiz=e='gte(c, 4.5*3)'
9271 Violent denoise using a block size of @code{16x16}:
9278 Remove banding artifacts from input video.
9279 It works by replacing banded pixels with average value of referenced pixels.
9281 The filter accepts the following options:
9288 Set banding detection threshold for each plane. Default is 0.02.
9289 Valid range is 0.00003 to 0.5.
9290 If difference between current pixel and reference pixel is less than threshold,
9291 it will be considered as banded.
9294 Banding detection range in pixels. Default is 16. If positive, random number
9295 in range 0 to set value will be used. If negative, exact absolute value
9297 The range defines square of four pixels around current pixel.
9300 Set direction in radians from which four pixel will be compared. If positive,
9301 random direction from 0 to set direction will be picked. If negative, exact of
9302 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9303 will pick only pixels on same row and -PI/2 will pick only pixels on same
9307 If enabled, current pixel is compared with average value of all four
9308 surrounding pixels. The default is enabled. If disabled current pixel is
9309 compared with all four surrounding pixels. The pixel is considered banded
9310 if only all four differences with surrounding pixels are less than threshold.
9313 If enabled, current pixel is changed if and only if all pixel components are banded,
9314 e.g. banding detection threshold is triggered for all color components.
9315 The default is disabled.
9320 Remove blocking artifacts from input video.
9322 The filter accepts the following options:
9326 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9327 This controls what kind of deblocking is applied.
9330 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9336 Set blocking detection thresholds. Allowed range is 0 to 1.
9337 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9338 Using higher threshold gives more deblocking strength.
9339 Setting @var{alpha} controls threshold detection at exact edge of block.
9340 Remaining options controls threshold detection near the edge. Each one for
9341 below/above or left/right. Setting any of those to @var{0} disables
9345 Set planes to filter. Default is to filter all available planes.
9348 @subsection Examples
9352 Deblock using weak filter and block size of 4 pixels.
9354 deblock=filter=weak:block=4
9358 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9359 deblocking more edges.
9361 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9365 Similar as above, but filter only first plane.
9367 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9371 Similar as above, but filter only second and third plane.
9373 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9380 Drop duplicated frames at regular intervals.
9382 The filter accepts the following options:
9386 Set the number of frames from which one will be dropped. Setting this to
9387 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9388 Default is @code{5}.
9391 Set the threshold for duplicate detection. If the difference metric for a frame
9392 is less than or equal to this value, then it is declared as duplicate. Default
9396 Set scene change threshold. Default is @code{15}.
9400 Set the size of the x and y-axis blocks used during metric calculations.
9401 Larger blocks give better noise suppression, but also give worse detection of
9402 small movements. Must be a power of two. Default is @code{32}.
9405 Mark main input as a pre-processed input and activate clean source input
9406 stream. This allows the input to be pre-processed with various filters to help
9407 the metrics calculation while keeping the frame selection lossless. When set to
9408 @code{1}, the first stream is for the pre-processed input, and the second
9409 stream is the clean source from where the kept frames are chosen. Default is
9413 Set whether or not chroma is considered in the metric calculations. Default is
9419 Apply 2D deconvolution of video stream in frequency domain using second stream
9422 The filter accepts the following options:
9426 Set which planes to process.
9429 Set which impulse video frames will be processed, can be @var{first}
9430 or @var{all}. Default is @var{all}.
9433 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9434 and height are not same and not power of 2 or if stream prior to convolving
9438 The @code{deconvolve} filter also supports the @ref{framesync} options.
9442 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9444 It accepts the following options:
9448 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9449 @var{rainbows} for cross-color reduction.
9452 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9455 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9458 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9461 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9466 Apply deflate effect to the video.
9468 This filter replaces the pixel by the local(3x3) average by taking into account
9469 only values lower than the pixel.
9471 It accepts the following options:
9478 Limit the maximum change for each plane, default is 65535.
9479 If 0, plane will remain unchanged.
9482 @subsection Commands
9484 This filter supports the all above options as @ref{commands}.
9488 Remove temporal frame luminance variations.
9490 It accepts the following options:
9494 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9497 Set averaging mode to smooth temporal luminance variations.
9499 Available values are:
9524 Do not actually modify frame. Useful when one only wants metadata.
9529 Remove judder produced by partially interlaced telecined content.
9531 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9532 source was partially telecined content then the output of @code{pullup,dejudder}
9533 will have a variable frame rate. May change the recorded frame rate of the
9534 container. Aside from that change, this filter will not affect constant frame
9537 The option available in this filter is:
9541 Specify the length of the window over which the judder repeats.
9543 Accepts any integer greater than 1. Useful values are:
9547 If the original was telecined from 24 to 30 fps (Film to NTSC).
9550 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9553 If a mixture of the two.
9556 The default is @samp{4}.
9561 Suppress a TV station logo by a simple interpolation of the surrounding
9562 pixels. Just set a rectangle covering the logo and watch it disappear
9563 (and sometimes something even uglier appear - your mileage may vary).
9565 It accepts the following parameters:
9570 Specify the top left corner coordinates of the logo. They must be
9575 Specify the width and height of the logo to clear. They must be
9579 Specify the thickness of the fuzzy edge of the rectangle (added to
9580 @var{w} and @var{h}). The default value is 1. This option is
9581 deprecated, setting higher values should no longer be necessary and
9585 When set to 1, a green rectangle is drawn on the screen to simplify
9586 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9587 The default value is 0.
9589 The rectangle is drawn on the outermost pixels which will be (partly)
9590 replaced with interpolated values. The values of the next pixels
9591 immediately outside this rectangle in each direction will be used to
9592 compute the interpolated pixel values inside the rectangle.
9596 @subsection Examples
9600 Set a rectangle covering the area with top left corner coordinates 0,0
9601 and size 100x77, and a band of size 10:
9603 delogo=x=0:y=0:w=100:h=77:band=10
9611 Remove the rain in the input image/video by applying the derain methods based on
9612 convolutional neural networks. Supported models:
9616 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9617 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9620 Training as well as model generation scripts are provided in
9621 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9623 Native model files (.model) can be generated from TensorFlow model
9624 files (.pb) by using tools/python/convert.py
9626 The filter accepts the following options:
9630 Specify which filter to use. This option accepts the following values:
9634 Derain filter. To conduct derain filter, you need to use a derain model.
9637 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9639 Default value is @samp{derain}.
9642 Specify which DNN backend to use for model loading and execution. This option accepts
9643 the following values:
9647 Native implementation of DNN loading and execution.
9650 TensorFlow backend. To enable this backend you
9651 need to install the TensorFlow for C library (see
9652 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9653 @code{--enable-libtensorflow}
9655 Default value is @samp{native}.
9658 Set path to model file specifying network architecture and its parameters.
9659 Note that different backends use different file formats. TensorFlow and native
9660 backend can load files for only its format.
9663 It can also be finished with @ref{dnn_processing} filter.
9667 Attempt to fix small changes in horizontal and/or vertical shift. This
9668 filter helps remove camera shake from hand-holding a camera, bumping a
9669 tripod, moving on a vehicle, etc.
9671 The filter accepts the following options:
9679 Specify a rectangular area where to limit the search for motion
9681 If desired the search for motion vectors can be limited to a
9682 rectangular area of the frame defined by its top left corner, width
9683 and height. These parameters have the same meaning as the drawbox
9684 filter which can be used to visualise the position of the bounding
9687 This is useful when simultaneous movement of subjects within the frame
9688 might be confused for camera motion by the motion vector search.
9690 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9691 then the full frame is used. This allows later options to be set
9692 without specifying the bounding box for the motion vector search.
9694 Default - search the whole frame.
9698 Specify the maximum extent of movement in x and y directions in the
9699 range 0-64 pixels. Default 16.
9702 Specify how to generate pixels to fill blanks at the edge of the
9703 frame. Available values are:
9706 Fill zeroes at blank locations
9708 Original image at blank locations
9710 Extruded edge value at blank locations
9712 Mirrored edge at blank locations
9714 Default value is @samp{mirror}.
9717 Specify the blocksize to use for motion search. Range 4-128 pixels,
9721 Specify the contrast threshold for blocks. Only blocks with more than
9722 the specified contrast (difference between darkest and lightest
9723 pixels) will be considered. Range 1-255, default 125.
9726 Specify the search strategy. Available values are:
9729 Set exhaustive search
9731 Set less exhaustive search.
9733 Default value is @samp{exhaustive}.
9736 If set then a detailed log of the motion search is written to the
9743 Remove unwanted contamination of foreground colors, caused by reflected color of
9744 greenscreen or bluescreen.
9746 This filter accepts the following options:
9750 Set what type of despill to use.
9753 Set how spillmap will be generated.
9756 Set how much to get rid of still remaining spill.
9759 Controls amount of red in spill area.
9762 Controls amount of green in spill area.
9763 Should be -1 for greenscreen.
9766 Controls amount of blue in spill area.
9767 Should be -1 for bluescreen.
9770 Controls brightness of spill area, preserving colors.
9773 Modify alpha from generated spillmap.
9776 @subsection Commands
9778 This filter supports the all above options as @ref{commands}.
9782 Apply an exact inverse of the telecine operation. It requires a predefined
9783 pattern specified using the pattern option which must be the same as that passed
9784 to the telecine filter.
9786 This filter accepts the following options:
9795 The default value is @code{top}.
9799 A string of numbers representing the pulldown pattern you wish to apply.
9800 The default value is @code{23}.
9803 A number representing position of the first frame with respect to the telecine
9804 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9809 Apply dilation effect to the video.
9811 This filter replaces the pixel by the local(3x3) maximum.
9813 It accepts the following options:
9820 Limit the maximum change for each plane, default is 65535.
9821 If 0, plane will remain unchanged.
9824 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9827 Flags to local 3x3 coordinates maps like this:
9834 @subsection Commands
9836 This filter supports the all above options as @ref{commands}.
9840 Displace pixels as indicated by second and third input stream.
9842 It takes three input streams and outputs one stream, the first input is the
9843 source, and second and third input are displacement maps.
9845 The second input specifies how much to displace pixels along the
9846 x-axis, while the third input specifies how much to displace pixels
9848 If one of displacement map streams terminates, last frame from that
9849 displacement map will be used.
9851 Note that once generated, displacements maps can be reused over and over again.
9853 A description of the accepted options follows.
9857 Set displace behavior for pixels that are out of range.
9859 Available values are:
9862 Missing pixels are replaced by black pixels.
9865 Adjacent pixels will spread out to replace missing pixels.
9868 Out of range pixels are wrapped so they point to pixels of other side.
9871 Out of range pixels will be replaced with mirrored pixels.
9873 Default is @samp{smear}.
9877 @subsection Examples
9881 Add ripple effect to rgb input of video size hd720:
9883 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
9887 Add wave effect to rgb input of video size hd720:
9889 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
9893 @anchor{dnn_processing}
9894 @section dnn_processing
9896 Do image processing with deep neural networks. It works together with another filter
9897 which converts the pixel format of the Frame to what the dnn network requires.
9899 The filter accepts the following options:
9903 Specify which DNN backend to use for model loading and execution. This option accepts
9904 the following values:
9908 Native implementation of DNN loading and execution.
9911 TensorFlow backend. To enable this backend you
9912 need to install the TensorFlow for C library (see
9913 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9914 @code{--enable-libtensorflow}
9917 OpenVINO backend. To enable this backend you
9918 need to build and install the OpenVINO for C library (see
9919 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9920 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9921 be needed if the header files and libraries are not installed into system path)
9925 Default value is @samp{native}.
9928 Set path to model file specifying network architecture and its parameters.
9929 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9930 backend can load files for only its format.
9932 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9935 Set the input name of the dnn network.
9938 Set the output name of the dnn network.
9942 @subsection Examples
9946 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9948 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9952 Halve the pixel value of the frame with format gray32f:
9954 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
9958 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9960 ./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
9964 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9966 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9973 Draw a colored box on the input image.
9975 It accepts the following parameters:
9980 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9984 The expressions which specify the width and height of the box; if 0 they are interpreted as
9985 the input width and height. It defaults to 0.
9988 Specify the color of the box to write. For the general syntax of this option,
9989 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9990 value @code{invert} is used, the box edge color is the same as the
9991 video with inverted luma.
9994 The expression which sets the thickness of the box edge.
9995 A value of @code{fill} will create a filled box. Default value is @code{3}.
9997 See below for the list of accepted constants.
10000 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10001 will overwrite the video's color and alpha pixels.
10002 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10005 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10006 following constants:
10010 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10014 horizontal and vertical chroma subsample values. For example for the
10015 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10019 The input width and height.
10022 The input sample aspect ratio.
10026 The x and y offset coordinates where the box is drawn.
10030 The width and height of the drawn box.
10033 The thickness of the drawn box.
10035 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10036 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10040 @subsection Examples
10044 Draw a black box around the edge of the input image:
10050 Draw a box with color red and an opacity of 50%:
10052 drawbox=10:20:200:60:red@@0.5
10055 The previous example can be specified as:
10057 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10061 Fill the box with pink color:
10063 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10067 Draw a 2-pixel red 2.40:1 mask:
10069 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
10073 @subsection Commands
10074 This filter supports same commands as options.
10075 The command accepts the same syntax of the corresponding option.
10077 If the specified expression is not valid, it is kept at its current
10082 Draw a graph using input video metadata.
10084 It accepts the following parameters:
10088 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10091 Set 1st foreground color expression.
10094 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10097 Set 2nd foreground color expression.
10100 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10103 Set 3rd foreground color expression.
10106 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10109 Set 4th foreground color expression.
10112 Set minimal value of metadata value.
10115 Set maximal value of metadata value.
10118 Set graph background color. Default is white.
10123 Available values for mode is:
10130 Default is @code{line}.
10135 Available values for slide is:
10138 Draw new frame when right border is reached.
10141 Replace old columns with new ones.
10144 Scroll from right to left.
10147 Scroll from left to right.
10150 Draw single picture.
10153 Default is @code{frame}.
10156 Set size of graph video. For the syntax of this option, check the
10157 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10158 The default value is @code{900x256}.
10161 Set the output frame rate. Default value is @code{25}.
10163 The foreground color expressions can use the following variables:
10166 Minimal value of metadata value.
10169 Maximal value of metadata value.
10172 Current metadata key value.
10175 The color is defined as 0xAABBGGRR.
10178 Example using metadata from @ref{signalstats} filter:
10180 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10183 Example using metadata from @ref{ebur128} filter:
10185 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10190 Draw a grid on the input image.
10192 It accepts the following parameters:
10197 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10201 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10202 input width and height, respectively, minus @code{thickness}, so image gets
10203 framed. Default to 0.
10206 Specify the color of the grid. For the general syntax of this option,
10207 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10208 value @code{invert} is used, the grid color is the same as the
10209 video with inverted luma.
10212 The expression which sets the thickness of the grid line. Default value is @code{1}.
10214 See below for the list of accepted constants.
10217 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10218 will overwrite the video's color and alpha pixels.
10219 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10222 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10223 following constants:
10227 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10231 horizontal and vertical chroma subsample values. For example for the
10232 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10236 The input grid cell width and height.
10239 The input sample aspect ratio.
10243 The x and y coordinates of some point of grid intersection (meant to configure offset).
10247 The width and height of the drawn cell.
10250 The thickness of the drawn cell.
10252 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10253 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10257 @subsection Examples
10261 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10263 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10267 Draw a white 3x3 grid with an opacity of 50%:
10269 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10273 @subsection Commands
10274 This filter supports same commands as options.
10275 The command accepts the same syntax of the corresponding option.
10277 If the specified expression is not valid, it is kept at its current
10283 Draw a text string or text from a specified file on top of a video, using the
10284 libfreetype library.
10286 To enable compilation of this filter, you need to configure FFmpeg with
10287 @code{--enable-libfreetype}.
10288 To enable default font fallback and the @var{font} option you need to
10289 configure FFmpeg with @code{--enable-libfontconfig}.
10290 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10291 @code{--enable-libfribidi}.
10295 It accepts the following parameters:
10300 Used to draw a box around text using the background color.
10301 The value must be either 1 (enable) or 0 (disable).
10302 The default value of @var{box} is 0.
10305 Set the width of the border to be drawn around the box using @var{boxcolor}.
10306 The default value of @var{boxborderw} is 0.
10309 The color to be used for drawing box around text. For the syntax of this
10310 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10312 The default value of @var{boxcolor} is "white".
10315 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10316 The default value of @var{line_spacing} is 0.
10319 Set the width of the border to be drawn around the text using @var{bordercolor}.
10320 The default value of @var{borderw} is 0.
10323 Set the color to be used for drawing border around text. For the syntax of this
10324 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10326 The default value of @var{bordercolor} is "black".
10329 Select how the @var{text} is expanded. Can be either @code{none},
10330 @code{strftime} (deprecated) or
10331 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10335 Set a start time for the count. Value is in microseconds. Only applied
10336 in the deprecated strftime expansion mode. To emulate in normal expansion
10337 mode use the @code{pts} function, supplying the start time (in seconds)
10338 as the second argument.
10341 If true, check and fix text coords to avoid clipping.
10344 The color to be used for drawing fonts. For the syntax of this option, check
10345 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10347 The default value of @var{fontcolor} is "black".
10349 @item fontcolor_expr
10350 String which is expanded the same way as @var{text} to obtain dynamic
10351 @var{fontcolor} value. By default this option has empty value and is not
10352 processed. When this option is set, it overrides @var{fontcolor} option.
10355 The font family to be used for drawing text. By default Sans.
10358 The font file to be used for drawing text. The path must be included.
10359 This parameter is mandatory if the fontconfig support is disabled.
10362 Draw the text applying alpha blending. The value can
10363 be a number between 0.0 and 1.0.
10364 The expression accepts the same variables @var{x, y} as well.
10365 The default value is 1.
10366 Please see @var{fontcolor_expr}.
10369 The font size to be used for drawing text.
10370 The default value of @var{fontsize} is 16.
10373 If set to 1, attempt to shape the text (for example, reverse the order of
10374 right-to-left text and join Arabic characters) before drawing it.
10375 Otherwise, just draw the text exactly as given.
10376 By default 1 (if supported).
10378 @item ft_load_flags
10379 The flags to be used for loading the fonts.
10381 The flags map the corresponding flags supported by libfreetype, and are
10382 a combination of the following values:
10389 @item vertical_layout
10390 @item force_autohint
10393 @item ignore_global_advance_width
10395 @item ignore_transform
10397 @item linear_design
10401 Default value is "default".
10403 For more information consult the documentation for the FT_LOAD_*
10407 The color to be used for drawing a shadow behind the drawn text. For the
10408 syntax of this option, check the @ref{color syntax,,"Color" section in the
10409 ffmpeg-utils manual,ffmpeg-utils}.
10411 The default value of @var{shadowcolor} is "black".
10415 The x and y offsets for the text shadow position with respect to the
10416 position of the text. They can be either positive or negative
10417 values. The default value for both is "0".
10420 The starting frame number for the n/frame_num variable. The default value
10424 The size in number of spaces to use for rendering the tab.
10425 Default value is 4.
10428 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10429 format. It can be used with or without text parameter. @var{timecode_rate}
10430 option must be specified.
10432 @item timecode_rate, rate, r
10433 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10434 integer. Minimum value is "1".
10435 Drop-frame timecode is supported for frame rates 30 & 60.
10438 If set to 1, the output of the timecode option will wrap around at 24 hours.
10439 Default is 0 (disabled).
10442 The text string to be drawn. The text must be a sequence of UTF-8
10443 encoded characters.
10444 This parameter is mandatory if no file is specified with the parameter
10448 A text file containing text to be drawn. The text must be a sequence
10449 of UTF-8 encoded characters.
10451 This parameter is mandatory if no text string is specified with the
10452 parameter @var{text}.
10454 If both @var{text} and @var{textfile} are specified, an error is thrown.
10457 If set to 1, the @var{textfile} will be reloaded before each frame.
10458 Be sure to update it atomically, or it may be read partially, or even fail.
10462 The expressions which specify the offsets where text will be drawn
10463 within the video frame. They are relative to the top/left border of the
10466 The default value of @var{x} and @var{y} is "0".
10468 See below for the list of accepted constants and functions.
10471 The parameters for @var{x} and @var{y} are expressions containing the
10472 following constants and functions:
10476 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10480 horizontal and vertical chroma subsample values. For example for the
10481 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10484 the height of each text line
10492 @item max_glyph_a, ascent
10493 the maximum distance from the baseline to the highest/upper grid
10494 coordinate used to place a glyph outline point, for all the rendered
10496 It is a positive value, due to the grid's orientation with the Y axis
10499 @item max_glyph_d, descent
10500 the maximum distance from the baseline to the lowest grid coordinate
10501 used to place a glyph outline point, for all the rendered glyphs.
10502 This is a negative value, due to the grid's orientation, with the Y axis
10506 maximum glyph height, that is the maximum height for all the glyphs
10507 contained in the rendered text, it is equivalent to @var{ascent} -
10511 maximum glyph width, that is the maximum width for all the glyphs
10512 contained in the rendered text
10515 the number of input frame, starting from 0
10517 @item rand(min, max)
10518 return a random number included between @var{min} and @var{max}
10521 The input sample aspect ratio.
10524 timestamp expressed in seconds, NAN if the input timestamp is unknown
10527 the height of the rendered text
10530 the width of the rendered text
10534 the x and y offset coordinates where the text is drawn.
10536 These parameters allow the @var{x} and @var{y} expressions to refer
10537 to each other, so you can for example specify @code{y=x/dar}.
10540 A one character description of the current frame's picture type.
10543 The current packet's position in the input file or stream
10544 (in bytes, from the start of the input). A value of -1 indicates
10545 this info is not available.
10548 The current packet's duration, in seconds.
10551 The current packet's size (in bytes).
10554 @anchor{drawtext_expansion}
10555 @subsection Text expansion
10557 If @option{expansion} is set to @code{strftime},
10558 the filter recognizes strftime() sequences in the provided text and
10559 expands them accordingly. Check the documentation of strftime(). This
10560 feature is deprecated.
10562 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10564 If @option{expansion} is set to @code{normal} (which is the default),
10565 the following expansion mechanism is used.
10567 The backslash character @samp{\}, followed by any character, always expands to
10568 the second character.
10570 Sequences of the form @code{%@{...@}} are expanded. The text between the
10571 braces is a function name, possibly followed by arguments separated by ':'.
10572 If the arguments contain special characters or delimiters (':' or '@}'),
10573 they should be escaped.
10575 Note that they probably must also be escaped as the value for the
10576 @option{text} option in the filter argument string and as the filter
10577 argument in the filtergraph description, and possibly also for the shell,
10578 that makes up to four levels of escaping; using a text file avoids these
10581 The following functions are available:
10586 The expression evaluation result.
10588 It must take one argument specifying the expression to be evaluated,
10589 which accepts the same constants and functions as the @var{x} and
10590 @var{y} values. Note that not all constants should be used, for
10591 example the text size is not known when evaluating the expression, so
10592 the constants @var{text_w} and @var{text_h} will have an undefined
10595 @item expr_int_format, eif
10596 Evaluate the expression's value and output as formatted integer.
10598 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10599 The second argument specifies the output format. Allowed values are @samp{x},
10600 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10601 @code{printf} function.
10602 The third parameter is optional and sets the number of positions taken by the output.
10603 It can be used to add padding with zeros from the left.
10606 The time at which the filter is running, expressed in UTC.
10607 It can accept an argument: a strftime() format string.
10610 The time at which the filter is running, expressed in the local time zone.
10611 It can accept an argument: a strftime() format string.
10614 Frame metadata. Takes one or two arguments.
10616 The first argument is mandatory and specifies the metadata key.
10618 The second argument is optional and specifies a default value, used when the
10619 metadata key is not found or empty.
10621 Available metadata can be identified by inspecting entries
10622 starting with TAG included within each frame section
10623 printed by running @code{ffprobe -show_frames}.
10625 String metadata generated in filters leading to
10626 the drawtext filter are also available.
10629 The frame number, starting from 0.
10632 A one character description of the current picture type.
10635 The timestamp of the current frame.
10636 It can take up to three arguments.
10638 The first argument is the format of the timestamp; it defaults to @code{flt}
10639 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10640 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10641 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10642 @code{localtime} stands for the timestamp of the frame formatted as
10643 local time zone time.
10645 The second argument is an offset added to the timestamp.
10647 If the format is set to @code{hms}, a third argument @code{24HH} may be
10648 supplied to present the hour part of the formatted timestamp in 24h format
10651 If the format is set to @code{localtime} or @code{gmtime},
10652 a third argument may be supplied: a strftime() format string.
10653 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10656 @subsection Commands
10658 This filter supports altering parameters via commands:
10661 Alter existing filter parameters.
10663 Syntax for the argument is the same as for filter invocation, e.g.
10666 fontsize=56:fontcolor=green:text='Hello World'
10669 Full filter invocation with sendcmd would look like this:
10672 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10676 If the entire argument can't be parsed or applied as valid values then the filter will
10677 continue with its existing parameters.
10679 @subsection Examples
10683 Draw "Test Text" with font FreeSerif, using the default values for the
10684 optional parameters.
10687 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10691 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10692 and y=50 (counting from the top-left corner of the screen), text is
10693 yellow with a red box around it. Both the text and the box have an
10697 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10698 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10701 Note that the double quotes are not necessary if spaces are not used
10702 within the parameter list.
10705 Show the text at the center of the video frame:
10707 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10711 Show the text at a random position, switching to a new position every 30 seconds:
10713 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)"
10717 Show a text line sliding from right to left in the last row of the video
10718 frame. The file @file{LONG_LINE} is assumed to contain a single line
10721 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10725 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10727 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10731 Draw a single green letter "g", at the center of the input video.
10732 The glyph baseline is placed at half screen height.
10734 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10738 Show text for 1 second every 3 seconds:
10740 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10744 Use fontconfig to set the font. Note that the colons need to be escaped.
10746 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10750 Draw "Test Text" with font size dependent on height of the video.
10752 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10756 Print the date of a real-time encoding (see strftime(3)):
10758 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10762 Show text fading in and out (appearing/disappearing):
10765 DS=1.0 # display start
10766 DE=10.0 # display end
10767 FID=1.5 # fade in duration
10768 FOD=5 # fade out duration
10769 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 @}"
10773 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10774 and the @option{fontsize} value are included in the @option{y} offset.
10776 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10777 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10781 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10782 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10783 must have option @option{-export_path_metadata 1} for the special metadata fields
10784 to be available for filters.
10786 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10791 For more information about libfreetype, check:
10792 @url{http://www.freetype.org/}.
10794 For more information about fontconfig, check:
10795 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10797 For more information about libfribidi, check:
10798 @url{http://fribidi.org/}.
10800 @section edgedetect
10802 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10804 The filter accepts the following options:
10809 Set low and high threshold values used by the Canny thresholding
10812 The high threshold selects the "strong" edge pixels, which are then
10813 connected through 8-connectivity with the "weak" edge pixels selected
10814 by the low threshold.
10816 @var{low} and @var{high} threshold values must be chosen in the range
10817 [0,1], and @var{low} should be lesser or equal to @var{high}.
10819 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10823 Define the drawing mode.
10827 Draw white/gray wires on black background.
10830 Mix the colors to create a paint/cartoon effect.
10833 Apply Canny edge detector on all selected planes.
10835 Default value is @var{wires}.
10838 Select planes for filtering. By default all available planes are filtered.
10841 @subsection Examples
10845 Standard edge detection with custom values for the hysteresis thresholding:
10847 edgedetect=low=0.1:high=0.4
10851 Painting effect without thresholding:
10853 edgedetect=mode=colormix:high=0
10859 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10861 For each input image, the filter will compute the optimal mapping from
10862 the input to the output given the codebook length, that is the number
10863 of distinct output colors.
10865 This filter accepts the following options.
10868 @item codebook_length, l
10869 Set codebook length. The value must be a positive integer, and
10870 represents the number of distinct output colors. Default value is 256.
10873 Set the maximum number of iterations to apply for computing the optimal
10874 mapping. The higher the value the better the result and the higher the
10875 computation time. Default value is 1.
10878 Set a random seed, must be an integer included between 0 and
10879 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10880 will try to use a good random seed on a best effort basis.
10883 Set pal8 output pixel format. This option does not work with codebook
10884 length greater than 256.
10889 Measure graylevel entropy in histogram of color channels of video frames.
10891 It accepts the following parameters:
10895 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10897 @var{diff} mode measures entropy of histogram delta values, absolute differences
10898 between neighbour histogram values.
10902 Set brightness, contrast, saturation and approximate gamma adjustment.
10904 The filter accepts the following options:
10908 Set the contrast expression. The value must be a float value in range
10909 @code{-1000.0} to @code{1000.0}. The default value is "1".
10912 Set the brightness expression. The value must be a float value in
10913 range @code{-1.0} to @code{1.0}. The default value is "0".
10916 Set the saturation expression. The value must be a float in
10917 range @code{0.0} to @code{3.0}. The default value is "1".
10920 Set the gamma expression. The value must be a float in range
10921 @code{0.1} to @code{10.0}. The default value is "1".
10924 Set the gamma expression for red. The value must be a float in
10925 range @code{0.1} to @code{10.0}. The default value is "1".
10928 Set the gamma expression for green. The value must be a float in range
10929 @code{0.1} to @code{10.0}. The default value is "1".
10932 Set the gamma expression for blue. The value must be a float in range
10933 @code{0.1} to @code{10.0}. The default value is "1".
10936 Set the gamma weight expression. It can be used to reduce the effect
10937 of a high gamma value on bright image areas, e.g. keep them from
10938 getting overamplified and just plain white. The value must be a float
10939 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10940 gamma correction all the way down while @code{1.0} leaves it at its
10941 full strength. Default is "1".
10944 Set when the expressions for brightness, contrast, saturation and
10945 gamma expressions are evaluated.
10947 It accepts the following values:
10950 only evaluate expressions once during the filter initialization or
10951 when a command is processed
10954 evaluate expressions for each incoming frame
10957 Default value is @samp{init}.
10960 The expressions accept the following parameters:
10963 frame count of the input frame starting from 0
10966 byte position of the corresponding packet in the input file, NAN if
10970 frame rate of the input video, NAN if the input frame rate is unknown
10973 timestamp expressed in seconds, NAN if the input timestamp is unknown
10976 @subsection Commands
10977 The filter supports the following commands:
10981 Set the contrast expression.
10984 Set the brightness expression.
10987 Set the saturation expression.
10990 Set the gamma expression.
10993 Set the gamma_r expression.
10996 Set gamma_g expression.
10999 Set gamma_b expression.
11002 Set gamma_weight expression.
11004 The command accepts the same syntax of the corresponding option.
11006 If the specified expression is not valid, it is kept at its current
11013 Apply erosion effect to the video.
11015 This filter replaces the pixel by the local(3x3) minimum.
11017 It accepts the following options:
11024 Limit the maximum change for each plane, default is 65535.
11025 If 0, plane will remain unchanged.
11028 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11031 Flags to local 3x3 coordinates maps like this:
11038 @subsection Commands
11040 This filter supports the all above options as @ref{commands}.
11042 @section extractplanes
11044 Extract color channel components from input video stream into
11045 separate grayscale video streams.
11047 The filter accepts the following option:
11051 Set plane(s) to extract.
11053 Available values for planes are:
11064 Choosing planes not available in the input will result in an error.
11065 That means you cannot select @code{r}, @code{g}, @code{b} planes
11066 with @code{y}, @code{u}, @code{v} planes at same time.
11069 @subsection Examples
11073 Extract luma, u and v color channel component from input video frame
11074 into 3 grayscale outputs:
11076 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
11082 Apply a fade-in/out effect to the input video.
11084 It accepts the following parameters:
11088 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11090 Default is @code{in}.
11092 @item start_frame, s
11093 Specify the number of the frame to start applying the fade
11094 effect at. Default is 0.
11097 The number of frames that the fade effect lasts. At the end of the
11098 fade-in effect, the output video will have the same intensity as the input video.
11099 At the end of the fade-out transition, the output video will be filled with the
11100 selected @option{color}.
11104 If set to 1, fade only alpha channel, if one exists on the input.
11105 Default value is 0.
11107 @item start_time, st
11108 Specify the timestamp (in seconds) of the frame to start to apply the fade
11109 effect. If both start_frame and start_time are specified, the fade will start at
11110 whichever comes last. Default is 0.
11113 The number of seconds for which the fade effect has to last. At the end of the
11114 fade-in effect the output video will have the same intensity as the input video,
11115 at the end of the fade-out transition the output video will be filled with the
11116 selected @option{color}.
11117 If both duration and nb_frames are specified, duration is used. Default is 0
11118 (nb_frames is used by default).
11121 Specify the color of the fade. Default is "black".
11124 @subsection Examples
11128 Fade in the first 30 frames of video:
11133 The command above is equivalent to:
11139 Fade out the last 45 frames of a 200-frame video:
11142 fade=type=out:start_frame=155:nb_frames=45
11146 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11148 fade=in:0:25, fade=out:975:25
11152 Make the first 5 frames yellow, then fade in from frame 5-24:
11154 fade=in:5:20:color=yellow
11158 Fade in alpha over first 25 frames of video:
11160 fade=in:0:25:alpha=1
11164 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11166 fade=t=in:st=5.5:d=0.5
11172 Denoise frames using 3D FFT (frequency domain filtering).
11174 The filter accepts the following options:
11178 Set the noise sigma constant. This sets denoising strength.
11179 Default value is 1. Allowed range is from 0 to 30.
11180 Using very high sigma with low overlap may give blocking artifacts.
11183 Set amount of denoising. By default all detected noise is reduced.
11184 Default value is 1. Allowed range is from 0 to 1.
11187 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11188 Actual size of block in pixels is 2 to power of @var{block}, so by default
11189 block size in pixels is 2^4 which is 16.
11192 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11195 Set number of previous frames to use for denoising. By default is set to 0.
11198 Set number of next frames to to use for denoising. By default is set to 0.
11201 Set planes which will be filtered, by default are all available filtered
11206 Apply arbitrary expressions to samples in frequency domain
11210 Adjust the dc value (gain) of the luma plane of the image. The filter
11211 accepts an integer value in range @code{0} to @code{1000}. The default
11212 value is set to @code{0}.
11215 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11216 filter accepts an integer value in range @code{0} to @code{1000}. The
11217 default value is set to @code{0}.
11220 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11221 filter accepts an integer value in range @code{0} to @code{1000}. The
11222 default value is set to @code{0}.
11225 Set the frequency domain weight expression for the luma plane.
11228 Set the frequency domain weight expression for the 1st chroma plane.
11231 Set the frequency domain weight expression for the 2nd chroma plane.
11234 Set when the expressions are evaluated.
11236 It accepts the following values:
11239 Only evaluate expressions once during the filter initialization.
11242 Evaluate expressions for each incoming frame.
11245 Default value is @samp{init}.
11247 The filter accepts the following variables:
11250 The coordinates of the current sample.
11254 The width and height of the image.
11257 The number of input frame, starting from 0.
11260 @subsection Examples
11266 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11272 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11278 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11284 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11291 Extract a single field from an interlaced image using stride
11292 arithmetic to avoid wasting CPU time. The output frames are marked as
11295 The filter accepts the following options:
11299 Specify whether to extract the top (if the value is @code{0} or
11300 @code{top}) or the bottom field (if the value is @code{1} or
11306 Create new frames by copying the top and bottom fields from surrounding frames
11307 supplied as numbers by the hint file.
11311 Set file containing hints: absolute/relative frame numbers.
11313 There must be one line for each frame in a clip. Each line must contain two
11314 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11315 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11316 is current frame number for @code{absolute} mode or out of [-1, 1] range
11317 for @code{relative} mode. First number tells from which frame to pick up top
11318 field and second number tells from which frame to pick up bottom field.
11320 If optionally followed by @code{+} output frame will be marked as interlaced,
11321 else if followed by @code{-} output frame will be marked as progressive, else
11322 it will be marked same as input frame.
11323 If optionally followed by @code{t} output frame will use only top field, or in
11324 case of @code{b} it will use only bottom field.
11325 If line starts with @code{#} or @code{;} that line is skipped.
11328 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11331 Example of first several lines of @code{hint} file for @code{relative} mode:
11333 0,0 - # first frame
11334 1,0 - # second frame, use third's frame top field and second's frame bottom field
11335 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11350 @section fieldmatch
11352 Field matching filter for inverse telecine. It is meant to reconstruct the
11353 progressive frames from a telecined stream. The filter does not drop duplicated
11354 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11355 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11357 The separation of the field matching and the decimation is notably motivated by
11358 the possibility of inserting a de-interlacing filter fallback between the two.
11359 If the source has mixed telecined and real interlaced content,
11360 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11361 But these remaining combed frames will be marked as interlaced, and thus can be
11362 de-interlaced by a later filter such as @ref{yadif} before decimation.
11364 In addition to the various configuration options, @code{fieldmatch} can take an
11365 optional second stream, activated through the @option{ppsrc} option. If
11366 enabled, the frames reconstruction will be based on the fields and frames from
11367 this second stream. This allows the first input to be pre-processed in order to
11368 help the various algorithms of the filter, while keeping the output lossless
11369 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11370 or brightness/contrast adjustments can help.
11372 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11373 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11374 which @code{fieldmatch} is based on. While the semantic and usage are very
11375 close, some behaviour and options names can differ.
11377 The @ref{decimate} filter currently only works for constant frame rate input.
11378 If your input has mixed telecined (30fps) and progressive content with a lower
11379 framerate like 24fps use the following filterchain to produce the necessary cfr
11380 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11382 The filter accepts the following options:
11386 Specify the assumed field order of the input stream. Available values are:
11390 Auto detect parity (use FFmpeg's internal parity value).
11392 Assume bottom field first.
11394 Assume top field first.
11397 Note that it is sometimes recommended not to trust the parity announced by the
11400 Default value is @var{auto}.
11403 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11404 sense that it won't risk creating jerkiness due to duplicate frames when
11405 possible, but if there are bad edits or blended fields it will end up
11406 outputting combed frames when a good match might actually exist. On the other
11407 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11408 but will almost always find a good frame if there is one. The other values are
11409 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11410 jerkiness and creating duplicate frames versus finding good matches in sections
11411 with bad edits, orphaned fields, blended fields, etc.
11413 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11415 Available values are:
11419 2-way matching (p/c)
11421 2-way matching, and trying 3rd match if still combed (p/c + n)
11423 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11425 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11426 still combed (p/c + n + u/b)
11428 3-way matching (p/c/n)
11430 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11431 detected as combed (p/c/n + u/b)
11434 The parenthesis at the end indicate the matches that would be used for that
11435 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11438 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11441 Default value is @var{pc_n}.
11444 Mark the main input stream as a pre-processed input, and enable the secondary
11445 input stream as the clean source to pick the fields from. See the filter
11446 introduction for more details. It is similar to the @option{clip2} feature from
11449 Default value is @code{0} (disabled).
11452 Set the field to match from. It is recommended to set this to the same value as
11453 @option{order} unless you experience matching failures with that setting. In
11454 certain circumstances changing the field that is used to match from can have a
11455 large impact on matching performance. Available values are:
11459 Automatic (same value as @option{order}).
11461 Match from the bottom field.
11463 Match from the top field.
11466 Default value is @var{auto}.
11469 Set whether or not chroma is included during the match comparisons. In most
11470 cases it is recommended to leave this enabled. You should set this to @code{0}
11471 only if your clip has bad chroma problems such as heavy rainbowing or other
11472 artifacts. Setting this to @code{0} could also be used to speed things up at
11473 the cost of some accuracy.
11475 Default value is @code{1}.
11479 These define an exclusion band which excludes the lines between @option{y0} and
11480 @option{y1} from being included in the field matching decision. An exclusion
11481 band can be used to ignore subtitles, a logo, or other things that may
11482 interfere with the matching. @option{y0} sets the starting scan line and
11483 @option{y1} sets the ending line; all lines in between @option{y0} and
11484 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11485 @option{y0} and @option{y1} to the same value will disable the feature.
11486 @option{y0} and @option{y1} defaults to @code{0}.
11489 Set the scene change detection threshold as a percentage of maximum change on
11490 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11491 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11492 @option{scthresh} is @code{[0.0, 100.0]}.
11494 Default value is @code{12.0}.
11497 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11498 account the combed scores of matches when deciding what match to use as the
11499 final match. Available values are:
11503 No final matching based on combed scores.
11505 Combed scores are only used when a scene change is detected.
11507 Use combed scores all the time.
11510 Default is @var{sc}.
11513 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11514 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11515 Available values are:
11519 No forced calculation.
11521 Force p/c/n calculations.
11523 Force p/c/n/u/b calculations.
11526 Default value is @var{none}.
11529 This is the area combing threshold used for combed frame detection. This
11530 essentially controls how "strong" or "visible" combing must be to be detected.
11531 Larger values mean combing must be more visible and smaller values mean combing
11532 can be less visible or strong and still be detected. Valid settings are from
11533 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11534 be detected as combed). This is basically a pixel difference value. A good
11535 range is @code{[8, 12]}.
11537 Default value is @code{9}.
11540 Sets whether or not chroma is considered in the combed frame decision. Only
11541 disable this if your source has chroma problems (rainbowing, etc.) that are
11542 causing problems for the combed frame detection with chroma enabled. Actually,
11543 using @option{chroma}=@var{0} is usually more reliable, except for the case
11544 where there is chroma only combing in the source.
11546 Default value is @code{0}.
11550 Respectively set the x-axis and y-axis size of the window used during combed
11551 frame detection. This has to do with the size of the area in which
11552 @option{combpel} pixels are required to be detected as combed for a frame to be
11553 declared combed. See the @option{combpel} parameter description for more info.
11554 Possible values are any number that is a power of 2 starting at 4 and going up
11557 Default value is @code{16}.
11560 The number of combed pixels inside any of the @option{blocky} by
11561 @option{blockx} size blocks on the frame for the frame to be detected as
11562 combed. While @option{cthresh} controls how "visible" the combing must be, this
11563 setting controls "how much" combing there must be in any localized area (a
11564 window defined by the @option{blockx} and @option{blocky} settings) on the
11565 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11566 which point no frames will ever be detected as combed). This setting is known
11567 as @option{MI} in TFM/VFM vocabulary.
11569 Default value is @code{80}.
11572 @anchor{p/c/n/u/b meaning}
11573 @subsection p/c/n/u/b meaning
11575 @subsubsection p/c/n
11577 We assume the following telecined stream:
11580 Top fields: 1 2 2 3 4
11581 Bottom fields: 1 2 3 4 4
11584 The numbers correspond to the progressive frame the fields relate to. Here, the
11585 first two frames are progressive, the 3rd and 4th are combed, and so on.
11587 When @code{fieldmatch} is configured to run a matching from bottom
11588 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11593 B 1 2 3 4 4 <-- matching reference
11602 As a result of the field matching, we can see that some frames get duplicated.
11603 To perform a complete inverse telecine, you need to rely on a decimation filter
11604 after this operation. See for instance the @ref{decimate} filter.
11606 The same operation now matching from top fields (@option{field}=@var{top})
11611 T 1 2 2 3 4 <-- matching reference
11621 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11622 basically, they refer to the frame and field of the opposite parity:
11625 @item @var{p} matches the field of the opposite parity in the previous frame
11626 @item @var{c} matches the field of the opposite parity in the current frame
11627 @item @var{n} matches the field of the opposite parity in the next frame
11632 The @var{u} and @var{b} matching are a bit special in the sense that they match
11633 from the opposite parity flag. In the following examples, we assume that we are
11634 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11635 'x' is placed above and below each matched fields.
11637 With bottom matching (@option{field}=@var{bottom}):
11642 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11643 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11651 With top matching (@option{field}=@var{top}):
11656 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11657 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11665 @subsection Examples
11667 Simple IVTC of a top field first telecined stream:
11669 fieldmatch=order=tff:combmatch=none, decimate
11672 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11674 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11677 @section fieldorder
11679 Transform the field order of the input video.
11681 It accepts the following parameters:
11686 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11687 for bottom field first.
11690 The default value is @samp{tff}.
11692 The transformation is done by shifting the picture content up or down
11693 by one line, and filling the remaining line with appropriate picture content.
11694 This method is consistent with most broadcast field order converters.
11696 If the input video is not flagged as being interlaced, or it is already
11697 flagged as being of the required output field order, then this filter does
11698 not alter the incoming video.
11700 It is very useful when converting to or from PAL DV material,
11701 which is bottom field first.
11705 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11708 @section fifo, afifo
11710 Buffer input images and send them when they are requested.
11712 It is mainly useful when auto-inserted by the libavfilter
11715 It does not take parameters.
11717 @section fillborders
11719 Fill borders of the input video, without changing video stream dimensions.
11720 Sometimes video can have garbage at the four edges and you may not want to
11721 crop video input to keep size multiple of some number.
11723 This filter accepts the following options:
11727 Number of pixels to fill from left border.
11730 Number of pixels to fill from right border.
11733 Number of pixels to fill from top border.
11736 Number of pixels to fill from bottom border.
11741 It accepts the following values:
11744 fill pixels using outermost pixels
11747 fill pixels using mirroring
11750 fill pixels with constant value
11753 Default is @var{smear}.
11756 Set color for pixels in fixed mode. Default is @var{black}.
11759 @subsection Commands
11760 This filter supports same @ref{commands} as options.
11761 The command accepts the same syntax of the corresponding option.
11763 If the specified expression is not valid, it is kept at its current
11768 Find a rectangular object
11770 It accepts the following options:
11774 Filepath of the object image, needs to be in gray8.
11777 Detection threshold, default is 0.5.
11780 Number of mipmaps, default is 3.
11782 @item xmin, ymin, xmax, ymax
11783 Specifies the rectangle in which to search.
11786 @subsection Examples
11790 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11792 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11798 Flood area with values of same pixel components with another values.
11800 It accepts the following options:
11803 Set pixel x coordinate.
11806 Set pixel y coordinate.
11809 Set source #0 component value.
11812 Set source #1 component value.
11815 Set source #2 component value.
11818 Set source #3 component value.
11821 Set destination #0 component value.
11824 Set destination #1 component value.
11827 Set destination #2 component value.
11830 Set destination #3 component value.
11836 Convert the input video to one of the specified pixel formats.
11837 Libavfilter will try to pick one that is suitable as input to
11840 It accepts the following parameters:
11844 A '|'-separated list of pixel format names, such as
11845 "pix_fmts=yuv420p|monow|rgb24".
11849 @subsection Examples
11853 Convert the input video to the @var{yuv420p} format
11855 format=pix_fmts=yuv420p
11858 Convert the input video to any of the formats in the list
11860 format=pix_fmts=yuv420p|yuv444p|yuv410p
11867 Convert the video to specified constant frame rate by duplicating or dropping
11868 frames as necessary.
11870 It accepts the following parameters:
11874 The desired output frame rate. The default is @code{25}.
11877 Assume the first PTS should be the given value, in seconds. This allows for
11878 padding/trimming at the start of stream. By default, no assumption is made
11879 about the first frame's expected PTS, so no padding or trimming is done.
11880 For example, this could be set to 0 to pad the beginning with duplicates of
11881 the first frame if a video stream starts after the audio stream or to trim any
11882 frames with a negative PTS.
11885 Timestamp (PTS) rounding method.
11887 Possible values are:
11894 round towards -infinity
11896 round towards +infinity
11900 The default is @code{near}.
11903 Action performed when reading the last frame.
11905 Possible values are:
11908 Use same timestamp rounding method as used for other frames.
11910 Pass through last frame if input duration has not been reached yet.
11912 The default is @code{round}.
11916 Alternatively, the options can be specified as a flat string:
11917 @var{fps}[:@var{start_time}[:@var{round}]].
11919 See also the @ref{setpts} filter.
11921 @subsection Examples
11925 A typical usage in order to set the fps to 25:
11931 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11933 fps=fps=film:round=near
11939 Pack two different video streams into a stereoscopic video, setting proper
11940 metadata on supported codecs. The two views should have the same size and
11941 framerate and processing will stop when the shorter video ends. Please note
11942 that you may conveniently adjust view properties with the @ref{scale} and
11945 It accepts the following parameters:
11949 The desired packing format. Supported values are:
11954 The views are next to each other (default).
11957 The views are on top of each other.
11960 The views are packed by line.
11963 The views are packed by column.
11966 The views are temporally interleaved.
11975 # Convert left and right views into a frame-sequential video
11976 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11978 # Convert views into a side-by-side video with the same output resolution as the input
11979 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
11984 Change the frame rate by interpolating new video output frames from the source
11987 This filter is not designed to function correctly with interlaced media. If
11988 you wish to change the frame rate of interlaced media then you are required
11989 to deinterlace before this filter and re-interlace after this filter.
11991 A description of the accepted options follows.
11995 Specify the output frames per second. This option can also be specified
11996 as a value alone. The default is @code{50}.
11999 Specify the start of a range where the output frame will be created as a
12000 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12001 the default is @code{15}.
12004 Specify the end of a range where the output frame will be created as a
12005 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12006 the default is @code{240}.
12009 Specify the level at which a scene change is detected as a value between
12010 0 and 100 to indicate a new scene; a low value reflects a low
12011 probability for the current frame to introduce a new scene, while a higher
12012 value means the current frame is more likely to be one.
12013 The default is @code{8.2}.
12016 Specify flags influencing the filter process.
12018 Available value for @var{flags} is:
12021 @item scene_change_detect, scd
12022 Enable scene change detection using the value of the option @var{scene}.
12023 This flag is enabled by default.
12029 Select one frame every N-th frame.
12031 This filter accepts the following option:
12034 Select frame after every @code{step} frames.
12035 Allowed values are positive integers higher than 0. Default value is @code{1}.
12038 @section freezedetect
12040 Detect frozen video.
12042 This filter logs a message and sets frame metadata when it detects that the
12043 input video has no significant change in content during a specified duration.
12044 Video freeze detection calculates the mean average absolute difference of all
12045 the components of video frames and compares it to a noise floor.
12047 The printed times and duration are expressed in seconds. The
12048 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12049 whose timestamp equals or exceeds the detection duration and it contains the
12050 timestamp of the first frame of the freeze. The
12051 @code{lavfi.freezedetect.freeze_duration} and
12052 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12055 The filter accepts the following options:
12059 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12060 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12064 Set freeze duration until notification (default is 2 seconds).
12067 @section freezeframes
12069 Freeze video frames.
12071 This filter freezes video frames using frame from 2nd input.
12073 The filter accepts the following options:
12077 Set number of first frame from which to start freeze.
12080 Set number of last frame from which to end freeze.
12083 Set number of frame from 2nd input which will be used instead of replaced frames.
12089 Apply a frei0r effect to the input video.
12091 To enable the compilation of this filter, you need to install the frei0r
12092 header and configure FFmpeg with @code{--enable-frei0r}.
12094 It accepts the following parameters:
12099 The name of the frei0r effect to load. If the environment variable
12100 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12101 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12102 Otherwise, the standard frei0r paths are searched, in this order:
12103 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12104 @file{/usr/lib/frei0r-1/}.
12106 @item filter_params
12107 A '|'-separated list of parameters to pass to the frei0r effect.
12111 A frei0r effect parameter can be a boolean (its value is either
12112 "y" or "n"), a double, a color (specified as
12113 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12114 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12115 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12116 a position (specified as @var{X}/@var{Y}, where
12117 @var{X} and @var{Y} are floating point numbers) and/or a string.
12119 The number and types of parameters depend on the loaded effect. If an
12120 effect parameter is not specified, the default value is set.
12122 @subsection Examples
12126 Apply the distort0r effect, setting the first two double parameters:
12128 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12132 Apply the colordistance effect, taking a color as the first parameter:
12134 frei0r=colordistance:0.2/0.3/0.4
12135 frei0r=colordistance:violet
12136 frei0r=colordistance:0x112233
12140 Apply the perspective effect, specifying the top left and top right image
12143 frei0r=perspective:0.2/0.2|0.8/0.2
12147 For more information, see
12148 @url{http://frei0r.dyne.org}
12150 @subsection Commands
12152 This filter supports the @option{filter_params} option as @ref{commands}.
12156 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12158 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12159 processing filter, one of them is performed once per block, not per pixel.
12160 This allows for much higher speed.
12162 The filter accepts the following options:
12166 Set quality. This option defines the number of levels for averaging. It accepts
12167 an integer in the range 4-5. Default value is @code{4}.
12170 Force a constant quantization parameter. It accepts an integer in range 0-63.
12171 If not set, the filter will use the QP from the video stream (if available).
12174 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12175 more details but also more artifacts, while higher values make the image smoother
12176 but also blurrier. Default value is @code{0} − PSNR optimal.
12178 @item use_bframe_qp
12179 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12180 option may cause flicker since the B-Frames have often larger QP. Default is
12181 @code{0} (not enabled).
12187 Apply Gaussian blur filter.
12189 The filter accepts the following options:
12193 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12196 Set number of steps for Gaussian approximation. Default is @code{1}.
12199 Set which planes to filter. By default all planes are filtered.
12202 Set vertical sigma, if negative it will be same as @code{sigma}.
12203 Default is @code{-1}.
12206 @subsection Commands
12207 This filter supports same commands as options.
12208 The command accepts the same syntax of the corresponding option.
12210 If the specified expression is not valid, it is kept at its current
12215 Apply generic equation to each pixel.
12217 The filter accepts the following options:
12220 @item lum_expr, lum
12221 Set the luminance expression.
12223 Set the chrominance blue expression.
12225 Set the chrominance red expression.
12226 @item alpha_expr, a
12227 Set the alpha expression.
12229 Set the red expression.
12230 @item green_expr, g
12231 Set the green expression.
12233 Set the blue expression.
12236 The colorspace is selected according to the specified options. If one
12237 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12238 options is specified, the filter will automatically select a YCbCr
12239 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12240 @option{blue_expr} options is specified, it will select an RGB
12243 If one of the chrominance expression is not defined, it falls back on the other
12244 one. If no alpha expression is specified it will evaluate to opaque value.
12245 If none of chrominance expressions are specified, they will evaluate
12246 to the luminance expression.
12248 The expressions can use the following variables and functions:
12252 The sequential number of the filtered frame, starting from @code{0}.
12256 The coordinates of the current sample.
12260 The width and height of the image.
12264 Width and height scale depending on the currently filtered plane. It is the
12265 ratio between the corresponding luma plane number of pixels and the current
12266 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12267 @code{0.5,0.5} for chroma planes.
12270 Time of the current frame, expressed in seconds.
12273 Return the value of the pixel at location (@var{x},@var{y}) of the current
12277 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12281 Return the value of the pixel at location (@var{x},@var{y}) of the
12282 blue-difference chroma plane. Return 0 if there is no such plane.
12285 Return the value of the pixel at location (@var{x},@var{y}) of the
12286 red-difference chroma plane. Return 0 if there is no such plane.
12291 Return the value of the pixel at location (@var{x},@var{y}) of the
12292 red/green/blue component. Return 0 if there is no such component.
12295 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12296 plane. Return 0 if there is no such plane.
12298 @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)
12299 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12300 sums of samples within a rectangle. See the functions without the sum postfix.
12302 @item interpolation
12303 Set one of interpolation methods:
12308 Default is bilinear.
12311 For functions, if @var{x} and @var{y} are outside the area, the value will be
12312 automatically clipped to the closer edge.
12314 Please note that this filter can use multiple threads in which case each slice
12315 will have its own expression state. If you want to use only a single expression
12316 state because your expressions depend on previous state then you should limit
12317 the number of filter threads to 1.
12319 @subsection Examples
12323 Flip the image horizontally:
12329 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12330 wavelength of 100 pixels:
12332 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12336 Generate a fancy enigmatic moving light:
12338 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
12342 Generate a quick emboss effect:
12344 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12348 Modify RGB components depending on pixel position:
12350 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12354 Create a radial gradient that is the same size as the input (also see
12355 the @ref{vignette} filter):
12357 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12363 Fix the banding artifacts that are sometimes introduced into nearly flat
12364 regions by truncation to 8-bit color depth.
12365 Interpolate the gradients that should go where the bands are, and
12368 It is designed for playback only. Do not use it prior to
12369 lossy compression, because compression tends to lose the dither and
12370 bring back the bands.
12372 It accepts the following parameters:
12377 The maximum amount by which the filter will change any one pixel. This is also
12378 the threshold for detecting nearly flat regions. Acceptable values range from
12379 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12383 The neighborhood to fit the gradient to. A larger radius makes for smoother
12384 gradients, but also prevents the filter from modifying the pixels near detailed
12385 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12386 values will be clipped to the valid range.
12390 Alternatively, the options can be specified as a flat string:
12391 @var{strength}[:@var{radius}]
12393 @subsection Examples
12397 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12403 Specify radius, omitting the strength (which will fall-back to the default
12411 @anchor{graphmonitor}
12412 @section graphmonitor
12413 Show various filtergraph stats.
12415 With this filter one can debug complete filtergraph.
12416 Especially issues with links filling with queued frames.
12418 The filter accepts the following options:
12422 Set video output size. Default is @var{hd720}.
12425 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12428 Set output mode, can be @var{fulll} or @var{compact}.
12429 In @var{compact} mode only filters with some queued frames have displayed stats.
12432 Set flags which enable which stats are shown in video.
12434 Available values for flags are:
12437 Display number of queued frames in each link.
12439 @item frame_count_in
12440 Display number of frames taken from filter.
12442 @item frame_count_out
12443 Display number of frames given out from filter.
12446 Display current filtered frame pts.
12449 Display current filtered frame time.
12452 Display time base for filter link.
12455 Display used format for filter link.
12458 Display video size or number of audio channels in case of audio used by filter link.
12461 Display video frame rate or sample rate in case of audio used by filter link.
12464 Display link output status.
12468 Set upper limit for video rate of output stream, Default value is @var{25}.
12469 This guarantee that output video frame rate will not be higher than this value.
12473 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12474 and corrects the scene colors accordingly.
12476 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12478 The filter accepts the following options:
12482 The order of differentiation to be applied on the scene. Must be chosen in the range
12483 [0,2] and default value is 1.
12486 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12487 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12488 max value instead of calculating Minkowski distance.
12491 The standard deviation of Gaussian blur to be applied on the scene. Must be
12492 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12493 can't be equal to 0 if @var{difford} is greater than 0.
12496 @subsection Examples
12502 greyedge=difford=1:minknorm=5:sigma=2
12508 greyedge=difford=1:minknorm=0:sigma=2
12516 Apply a Hald CLUT to a video stream.
12518 First input is the video stream to process, and second one is the Hald CLUT.
12519 The Hald CLUT input can be a simple picture or a complete video stream.
12521 The filter accepts the following options:
12525 Force termination when the shortest input terminates. Default is @code{0}.
12527 Continue applying the last CLUT after the end of the stream. A value of
12528 @code{0} disable the filter after the last frame of the CLUT is reached.
12529 Default is @code{1}.
12532 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12533 filters share the same internals).
12535 This filter also supports the @ref{framesync} options.
12537 More information about the Hald CLUT can be found on Eskil Steenberg's website
12538 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12540 @subsection Workflow examples
12542 @subsubsection Hald CLUT video stream
12544 Generate an identity Hald CLUT stream altered with various effects:
12546 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
12549 Note: make sure you use a lossless codec.
12551 Then use it with @code{haldclut} to apply it on some random stream:
12553 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12556 The Hald CLUT will be applied to the 10 first seconds (duration of
12557 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12558 to the remaining frames of the @code{mandelbrot} stream.
12560 @subsubsection Hald CLUT with preview
12562 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12563 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12564 biggest possible square starting at the top left of the picture. The remaining
12565 padding pixels (bottom or right) will be ignored. This area can be used to add
12566 a preview of the Hald CLUT.
12568 Typically, the following generated Hald CLUT will be supported by the
12569 @code{haldclut} filter:
12572 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12573 pad=iw+320 [padded_clut];
12574 smptebars=s=320x256, split [a][b];
12575 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12576 [main][b] overlay=W-320" -frames:v 1 clut.png
12579 It contains the original and a preview of the effect of the CLUT: SMPTE color
12580 bars are displayed on the right-top, and below the same color bars processed by
12583 Then, the effect of this Hald CLUT can be visualized with:
12585 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12590 Flip the input video horizontally.
12592 For example, to horizontally flip the input video with @command{ffmpeg}:
12594 ffmpeg -i in.avi -vf "hflip" out.avi
12598 This filter applies a global color histogram equalization on a
12601 It can be used to correct video that has a compressed range of pixel
12602 intensities. The filter redistributes the pixel intensities to
12603 equalize their distribution across the intensity range. It may be
12604 viewed as an "automatically adjusting contrast filter". This filter is
12605 useful only for correcting degraded or poorly captured source
12608 The filter accepts the following options:
12612 Determine the amount of equalization to be applied. As the strength
12613 is reduced, the distribution of pixel intensities more-and-more
12614 approaches that of the input frame. The value must be a float number
12615 in the range [0,1] and defaults to 0.200.
12618 Set the maximum intensity that can generated and scale the output
12619 values appropriately. The strength should be set as desired and then
12620 the intensity can be limited if needed to avoid washing-out. The value
12621 must be a float number in the range [0,1] and defaults to 0.210.
12624 Set the antibanding level. If enabled the filter will randomly vary
12625 the luminance of output pixels by a small amount to avoid banding of
12626 the histogram. Possible values are @code{none}, @code{weak} or
12627 @code{strong}. It defaults to @code{none}.
12633 Compute and draw a color distribution histogram for the input video.
12635 The computed histogram is a representation of the color component
12636 distribution in an image.
12638 Standard histogram displays the color components distribution in an image.
12639 Displays color graph for each color component. Shows distribution of
12640 the Y, U, V, A or R, G, B components, depending on input format, in the
12641 current frame. Below each graph a color component scale meter is shown.
12643 The filter accepts the following options:
12647 Set height of level. Default value is @code{200}.
12648 Allowed range is [50, 2048].
12651 Set height of color scale. Default value is @code{12}.
12652 Allowed range is [0, 40].
12656 It accepts the following values:
12659 Per color component graphs are placed below each other.
12662 Per color component graphs are placed side by side.
12665 Presents information identical to that in the @code{parade}, except
12666 that the graphs representing color components are superimposed directly
12669 Default is @code{stack}.
12672 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12673 Default is @code{linear}.
12676 Set what color components to display.
12677 Default is @code{7}.
12680 Set foreground opacity. Default is @code{0.7}.
12683 Set background opacity. Default is @code{0.5}.
12686 @subsection Examples
12691 Calculate and draw histogram:
12693 ffplay -i input -vf histogram
12701 This is a high precision/quality 3d denoise filter. It aims to reduce
12702 image noise, producing smooth images and making still images really
12703 still. It should enhance compressibility.
12705 It accepts the following optional parameters:
12709 A non-negative floating point number which specifies spatial luma strength.
12710 It defaults to 4.0.
12712 @item chroma_spatial
12713 A non-negative floating point number which specifies spatial chroma strength.
12714 It defaults to 3.0*@var{luma_spatial}/4.0.
12717 A floating point number which specifies luma temporal strength. It defaults to
12718 6.0*@var{luma_spatial}/4.0.
12721 A floating point number which specifies chroma temporal strength. It defaults to
12722 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12725 @subsection Commands
12726 This filter supports same @ref{commands} as options.
12727 The command accepts the same syntax of the corresponding option.
12729 If the specified expression is not valid, it is kept at its current
12732 @anchor{hwdownload}
12733 @section hwdownload
12735 Download hardware frames to system memory.
12737 The input must be in hardware frames, and the output a non-hardware format.
12738 Not all formats will be supported on the output - it may be necessary to insert
12739 an additional @option{format} filter immediately following in the graph to get
12740 the output in a supported format.
12744 Map hardware frames to system memory or to another device.
12746 This filter has several different modes of operation; which one is used depends
12747 on the input and output formats:
12750 Hardware frame input, normal frame output
12752 Map the input frames to system memory and pass them to the output. If the
12753 original hardware frame is later required (for example, after overlaying
12754 something else on part of it), the @option{hwmap} filter can be used again
12755 in the next mode to retrieve it.
12757 Normal frame input, hardware frame output
12759 If the input is actually a software-mapped hardware frame, then unmap it -
12760 that is, return the original hardware frame.
12762 Otherwise, a device must be provided. Create new hardware surfaces on that
12763 device for the output, then map them back to the software format at the input
12764 and give those frames to the preceding filter. This will then act like the
12765 @option{hwupload} filter, but may be able to avoid an additional copy when
12766 the input is already in a compatible format.
12768 Hardware frame input and output
12770 A device must be supplied for the output, either directly or with the
12771 @option{derive_device} option. The input and output devices must be of
12772 different types and compatible - the exact meaning of this is
12773 system-dependent, but typically it means that they must refer to the same
12774 underlying hardware context (for example, refer to the same graphics card).
12776 If the input frames were originally created on the output device, then unmap
12777 to retrieve the original frames.
12779 Otherwise, map the frames to the output device - create new hardware frames
12780 on the output corresponding to the frames on the input.
12783 The following additional parameters are accepted:
12787 Set the frame mapping mode. Some combination of:
12790 The mapped frame should be readable.
12792 The mapped frame should be writeable.
12794 The mapping will always overwrite the entire frame.
12796 This may improve performance in some cases, as the original contents of the
12797 frame need not be loaded.
12799 The mapping must not involve any copying.
12801 Indirect mappings to copies of frames are created in some cases where either
12802 direct mapping is not possible or it would have unexpected properties.
12803 Setting this flag ensures that the mapping is direct and will fail if that is
12806 Defaults to @var{read+write} if not specified.
12808 @item derive_device @var{type}
12809 Rather than using the device supplied at initialisation, instead derive a new
12810 device of type @var{type} from the device the input frames exist on.
12813 In a hardware to hardware mapping, map in reverse - create frames in the sink
12814 and map them back to the source. This may be necessary in some cases where
12815 a mapping in one direction is required but only the opposite direction is
12816 supported by the devices being used.
12818 This option is dangerous - it may break the preceding filter in undefined
12819 ways if there are any additional constraints on that filter's output.
12820 Do not use it without fully understanding the implications of its use.
12826 Upload system memory frames to hardware surfaces.
12828 The device to upload to must be supplied when the filter is initialised. If
12829 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12830 option or with the @option{derive_device} option. The input and output devices
12831 must be of different types and compatible - the exact meaning of this is
12832 system-dependent, but typically it means that they must refer to the same
12833 underlying hardware context (for example, refer to the same graphics card).
12835 The following additional parameters are accepted:
12838 @item derive_device @var{type}
12839 Rather than using the device supplied at initialisation, instead derive a new
12840 device of type @var{type} from the device the input frames exist on.
12843 @anchor{hwupload_cuda}
12844 @section hwupload_cuda
12846 Upload system memory frames to a CUDA device.
12848 It accepts the following optional parameters:
12852 The number of the CUDA device to use
12857 Apply a high-quality magnification filter designed for pixel art. This filter
12858 was originally created by Maxim Stepin.
12860 It accepts the following option:
12864 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12865 @code{hq3x} and @code{4} for @code{hq4x}.
12866 Default is @code{3}.
12870 Stack input videos horizontally.
12872 All streams must be of same pixel format and of same height.
12874 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12875 to create same output.
12877 The filter accepts the following option:
12881 Set number of input streams. Default is 2.
12884 If set to 1, force the output to terminate when the shortest input
12885 terminates. Default value is 0.
12890 Modify the hue and/or the saturation of the input.
12892 It accepts the following parameters:
12896 Specify the hue angle as a number of degrees. It accepts an expression,
12897 and defaults to "0".
12900 Specify the saturation in the [-10,10] range. It accepts an expression and
12904 Specify the hue angle as a number of radians. It accepts an
12905 expression, and defaults to "0".
12908 Specify the brightness in the [-10,10] range. It accepts an expression and
12912 @option{h} and @option{H} are mutually exclusive, and can't be
12913 specified at the same time.
12915 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12916 expressions containing the following constants:
12920 frame count of the input frame starting from 0
12923 presentation timestamp of the input frame expressed in time base units
12926 frame rate of the input video, NAN if the input frame rate is unknown
12929 timestamp expressed in seconds, NAN if the input timestamp is unknown
12932 time base of the input video
12935 @subsection Examples
12939 Set the hue to 90 degrees and the saturation to 1.0:
12945 Same command but expressing the hue in radians:
12951 Rotate hue and make the saturation swing between 0
12952 and 2 over a period of 1 second:
12954 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12958 Apply a 3 seconds saturation fade-in effect starting at 0:
12960 hue="s=min(t/3\,1)"
12963 The general fade-in expression can be written as:
12965 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12969 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12971 hue="s=max(0\, min(1\, (8-t)/3))"
12974 The general fade-out expression can be written as:
12976 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12981 @subsection Commands
12983 This filter supports the following commands:
12989 Modify the hue and/or the saturation and/or brightness of the input video.
12990 The command accepts the same syntax of the corresponding option.
12992 If the specified expression is not valid, it is kept at its current
12996 @section hysteresis
12998 Grow first stream into second stream by connecting components.
12999 This makes it possible to build more robust edge masks.
13001 This filter accepts the following options:
13005 Set which planes will be processed as bitmap, unprocessed planes will be
13006 copied from first stream.
13007 By default value 0xf, all planes will be processed.
13010 Set threshold which is used in filtering. If pixel component value is higher than
13011 this value filter algorithm for connecting components is activated.
13012 By default value is 0.
13015 The @code{hysteresis} filter also supports the @ref{framesync} options.
13019 Detect video interlacing type.
13021 This filter tries to detect if the input frames are interlaced, progressive,
13022 top or bottom field first. It will also try to detect fields that are
13023 repeated between adjacent frames (a sign of telecine).
13025 Single frame detection considers only immediately adjacent frames when classifying each frame.
13026 Multiple frame detection incorporates the classification history of previous frames.
13028 The filter will log these metadata values:
13031 @item single.current_frame
13032 Detected type of current frame using single-frame detection. One of:
13033 ``tff'' (top field first), ``bff'' (bottom field first),
13034 ``progressive'', or ``undetermined''
13037 Cumulative number of frames detected as top field first using single-frame detection.
13040 Cumulative number of frames detected as top field first using multiple-frame detection.
13043 Cumulative number of frames detected as bottom field first using single-frame detection.
13045 @item multiple.current_frame
13046 Detected type of current frame using multiple-frame detection. One of:
13047 ``tff'' (top field first), ``bff'' (bottom field first),
13048 ``progressive'', or ``undetermined''
13051 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13053 @item single.progressive
13054 Cumulative number of frames detected as progressive using single-frame detection.
13056 @item multiple.progressive
13057 Cumulative number of frames detected as progressive using multiple-frame detection.
13059 @item single.undetermined
13060 Cumulative number of frames that could not be classified using single-frame detection.
13062 @item multiple.undetermined
13063 Cumulative number of frames that could not be classified using multiple-frame detection.
13065 @item repeated.current_frame
13066 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13068 @item repeated.neither
13069 Cumulative number of frames with no repeated field.
13072 Cumulative number of frames with the top field repeated from the previous frame's top field.
13074 @item repeated.bottom
13075 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13078 The filter accepts the following options:
13082 Set interlacing threshold.
13084 Set progressive threshold.
13086 Threshold for repeated field detection.
13088 Number of frames after which a given frame's contribution to the
13089 statistics is halved (i.e., it contributes only 0.5 to its
13090 classification). The default of 0 means that all frames seen are given
13091 full weight of 1.0 forever.
13092 @item analyze_interlaced_flag
13093 When this is not 0 then idet will use the specified number of frames to determine
13094 if the interlaced flag is accurate, it will not count undetermined frames.
13095 If the flag is found to be accurate it will be used without any further
13096 computations, if it is found to be inaccurate it will be cleared without any
13097 further computations. This allows inserting the idet filter as a low computational
13098 method to clean up the interlaced flag
13103 Deinterleave or interleave fields.
13105 This filter allows one to process interlaced images fields without
13106 deinterlacing them. Deinterleaving splits the input frame into 2
13107 fields (so called half pictures). Odd lines are moved to the top
13108 half of the output image, even lines to the bottom half.
13109 You can process (filter) them independently and then re-interleave them.
13111 The filter accepts the following options:
13115 @item chroma_mode, c
13116 @item alpha_mode, a
13117 Available values for @var{luma_mode}, @var{chroma_mode} and
13118 @var{alpha_mode} are:
13124 @item deinterleave, d
13125 Deinterleave fields, placing one above the other.
13127 @item interleave, i
13128 Interleave fields. Reverse the effect of deinterleaving.
13130 Default value is @code{none}.
13132 @item luma_swap, ls
13133 @item chroma_swap, cs
13134 @item alpha_swap, as
13135 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13138 @subsection Commands
13140 This filter supports the all above options as @ref{commands}.
13144 Apply inflate effect to the video.
13146 This filter replaces the pixel by the local(3x3) average by taking into account
13147 only values higher than the pixel.
13149 It accepts the following options:
13156 Limit the maximum change for each plane, default is 65535.
13157 If 0, plane will remain unchanged.
13160 @subsection Commands
13162 This filter supports the all above options as @ref{commands}.
13166 Simple interlacing filter from progressive contents. This interleaves upper (or
13167 lower) lines from odd frames with lower (or upper) lines from even frames,
13168 halving the frame rate and preserving image height.
13171 Original Original New Frame
13172 Frame 'j' Frame 'j+1' (tff)
13173 ========== =========== ==================
13174 Line 0 --------------------> Frame 'j' Line 0
13175 Line 1 Line 1 ----> Frame 'j+1' Line 1
13176 Line 2 ---------------------> Frame 'j' Line 2
13177 Line 3 Line 3 ----> Frame 'j+1' Line 3
13179 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13182 It accepts the following optional parameters:
13186 This determines whether the interlaced frame is taken from the even
13187 (tff - default) or odd (bff) lines of the progressive frame.
13190 Vertical lowpass filter to avoid twitter interlacing and
13191 reduce moire patterns.
13195 Disable vertical lowpass filter
13198 Enable linear filter (default)
13201 Enable complex filter. This will slightly less reduce twitter and moire
13202 but better retain detail and subjective sharpness impression.
13209 Deinterlace input video by applying Donald Graft's adaptive kernel
13210 deinterling. Work on interlaced parts of a video to produce
13211 progressive frames.
13213 The description of the accepted parameters follows.
13217 Set the threshold which affects the filter's tolerance when
13218 determining if a pixel line must be processed. It must be an integer
13219 in the range [0,255] and defaults to 10. A value of 0 will result in
13220 applying the process on every pixels.
13223 Paint pixels exceeding the threshold value to white if set to 1.
13227 Set the fields order. Swap fields if set to 1, leave fields alone if
13231 Enable additional sharpening if set to 1. Default is 0.
13234 Enable twoway sharpening if set to 1. Default is 0.
13237 @subsection Examples
13241 Apply default values:
13243 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13247 Enable additional sharpening:
13253 Paint processed pixels in white:
13261 Slowly update darker pixels.
13263 This filter makes short flashes of light appear longer.
13264 This filter accepts the following options:
13268 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13271 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13274 @section lenscorrection
13276 Correct radial lens distortion
13278 This filter can be used to correct for radial distortion as can result from the use
13279 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13280 one can use tools available for example as part of opencv or simply trial-and-error.
13281 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13282 and extract the k1 and k2 coefficients from the resulting matrix.
13284 Note that effectively the same filter is available in the open-source tools Krita and
13285 Digikam from the KDE project.
13287 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13288 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13289 brightness distribution, so you may want to use both filters together in certain
13290 cases, though you will have to take care of ordering, i.e. whether vignetting should
13291 be applied before or after lens correction.
13293 @subsection Options
13295 The filter accepts the following options:
13299 Relative x-coordinate of the focal point of the image, and thereby the center of the
13300 distortion. This value has a range [0,1] and is expressed as fractions of the image
13301 width. Default is 0.5.
13303 Relative y-coordinate of the focal point of the image, and thereby the center of the
13304 distortion. This value has a range [0,1] and is expressed as fractions of the image
13305 height. Default is 0.5.
13307 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13308 no correction. Default is 0.
13310 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13311 0 means no correction. Default is 0.
13314 The formula that generates the correction is:
13316 @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)
13318 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13319 distances from the focal point in the source and target images, respectively.
13323 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13325 The @code{lensfun} filter requires the camera make, camera model, and lens model
13326 to apply the lens correction. The filter will load the lensfun database and
13327 query it to find the corresponding camera and lens entries in the database. As
13328 long as these entries can be found with the given options, the filter can
13329 perform corrections on frames. Note that incomplete strings will result in the
13330 filter choosing the best match with the given options, and the filter will
13331 output the chosen camera and lens models (logged with level "info"). You must
13332 provide the make, camera model, and lens model as they are required.
13334 The filter accepts the following options:
13338 The make of the camera (for example, "Canon"). This option is required.
13341 The model of the camera (for example, "Canon EOS 100D"). This option is
13345 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13346 option is required.
13349 The type of correction to apply. The following values are valid options:
13353 Enables fixing lens vignetting.
13356 Enables fixing lens geometry. This is the default.
13359 Enables fixing chromatic aberrations.
13362 Enables fixing lens vignetting and lens geometry.
13365 Enables fixing lens vignetting and chromatic aberrations.
13368 Enables fixing both lens geometry and chromatic aberrations.
13371 Enables all possible corrections.
13375 The focal length of the image/video (zoom; expected constant for video). For
13376 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13377 range should be chosen when using that lens. Default 18.
13380 The aperture of the image/video (expected constant for video). Note that
13381 aperture is only used for vignetting correction. Default 3.5.
13383 @item focus_distance
13384 The focus distance of the image/video (expected constant for video). Note that
13385 focus distance is only used for vignetting and only slightly affects the
13386 vignetting correction process. If unknown, leave it at the default value (which
13390 The scale factor which is applied after transformation. After correction the
13391 video is no longer necessarily rectangular. This parameter controls how much of
13392 the resulting image is visible. The value 0 means that a value will be chosen
13393 automatically such that there is little or no unmapped area in the output
13394 image. 1.0 means that no additional scaling is done. Lower values may result
13395 in more of the corrected image being visible, while higher values may avoid
13396 unmapped areas in the output.
13398 @item target_geometry
13399 The target geometry of the output image/video. The following values are valid
13403 @item rectilinear (default)
13406 @item equirectangular
13407 @item fisheye_orthographic
13408 @item fisheye_stereographic
13409 @item fisheye_equisolid
13410 @item fisheye_thoby
13413 Apply the reverse of image correction (instead of correcting distortion, apply
13416 @item interpolation
13417 The type of interpolation used when correcting distortion. The following values
13422 @item linear (default)
13427 @subsection Examples
13431 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13432 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13436 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
13440 Apply the same as before, but only for the first 5 seconds of video.
13443 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
13450 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13451 score between two input videos.
13453 The obtained VMAF score is printed through the logging system.
13455 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13456 After installing the library it can be enabled using:
13457 @code{./configure --enable-libvmaf}.
13458 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13460 The filter has following options:
13464 Set the model path which is to be used for SVM.
13465 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13468 Set the file path to be used to store logs.
13471 Set the format of the log file (csv, json or xml).
13473 @item enable_transform
13474 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13475 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13476 Default value: @code{false}
13479 Invokes the phone model which will generate VMAF scores higher than in the
13480 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13481 Default value: @code{false}
13484 Enables computing psnr along with vmaf.
13485 Default value: @code{false}
13488 Enables computing ssim along with vmaf.
13489 Default value: @code{false}
13492 Enables computing ms_ssim along with vmaf.
13493 Default value: @code{false}
13496 Set the pool method to be used for computing vmaf.
13497 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13500 Set number of threads to be used when computing vmaf.
13501 Default value: @code{0}, which makes use of all available logical processors.
13504 Set interval for frame subsampling used when computing vmaf.
13505 Default value: @code{1}
13507 @item enable_conf_interval
13508 Enables confidence interval.
13509 Default value: @code{false}
13512 This filter also supports the @ref{framesync} options.
13514 @subsection Examples
13517 On the below examples the input file @file{main.mpg} being processed is
13518 compared with the reference file @file{ref.mpg}.
13521 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13525 Example with options:
13527 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13531 Example with options and different containers:
13533 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 -
13539 Limits the pixel components values to the specified range [min, max].
13541 The filter accepts the following options:
13545 Lower bound. Defaults to the lowest allowed value for the input.
13548 Upper bound. Defaults to the highest allowed value for the input.
13551 Specify which planes will be processed. Defaults to all available.
13554 @subsection Commands
13556 This filter supports the all above options as @ref{commands}.
13562 The filter accepts the following options:
13566 Set the number of loops. Setting this value to -1 will result in infinite loops.
13570 Set maximal size in number of frames. Default is 0.
13573 Set first frame of loop. Default is 0.
13576 @subsection Examples
13580 Loop single first frame infinitely:
13582 loop=loop=-1:size=1:start=0
13586 Loop single first frame 10 times:
13588 loop=loop=10:size=1:start=0
13592 Loop 10 first frames 5 times:
13594 loop=loop=5:size=10:start=0
13600 Apply a 1D LUT to an input video.
13602 The filter accepts the following options:
13606 Set the 1D LUT file name.
13608 Currently supported formats:
13617 Select interpolation mode.
13619 Available values are:
13623 Use values from the nearest defined point.
13625 Interpolate values using the linear interpolation.
13627 Interpolate values using the cosine interpolation.
13629 Interpolate values using the cubic interpolation.
13631 Interpolate values using the spline interpolation.
13638 Apply a 3D LUT to an input video.
13640 The filter accepts the following options:
13644 Set the 3D LUT file name.
13646 Currently supported formats:
13660 Select interpolation mode.
13662 Available values are:
13666 Use values from the nearest defined point.
13668 Interpolate values using the 8 points defining a cube.
13670 Interpolate values using a tetrahedron.
13676 Turn certain luma values into transparency.
13678 The filter accepts the following options:
13682 Set the luma which will be used as base for transparency.
13683 Default value is @code{0}.
13686 Set the range of luma values to be keyed out.
13687 Default value is @code{0.01}.
13690 Set the range of softness. Default value is @code{0}.
13691 Use this to control gradual transition from zero to full transparency.
13694 @subsection Commands
13695 This filter supports same @ref{commands} as options.
13696 The command accepts the same syntax of the corresponding option.
13698 If the specified expression is not valid, it is kept at its current
13701 @section lut, lutrgb, lutyuv
13703 Compute a look-up table for binding each pixel component input value
13704 to an output value, and apply it to the input video.
13706 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13707 to an RGB input video.
13709 These filters accept the following parameters:
13712 set first pixel component expression
13714 set second pixel component expression
13716 set third pixel component expression
13718 set fourth pixel component expression, corresponds to the alpha component
13721 set red component expression
13723 set green component expression
13725 set blue component expression
13727 alpha component expression
13730 set Y/luminance component expression
13732 set U/Cb component expression
13734 set V/Cr component expression
13737 Each of them specifies the expression to use for computing the lookup table for
13738 the corresponding pixel component values.
13740 The exact component associated to each of the @var{c*} options depends on the
13743 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13744 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13746 The expressions can contain the following constants and functions:
13751 The input width and height.
13754 The input value for the pixel component.
13757 The input value, clipped to the @var{minval}-@var{maxval} range.
13760 The maximum value for the pixel component.
13763 The minimum value for the pixel component.
13766 The negated value for the pixel component value, clipped to the
13767 @var{minval}-@var{maxval} range; it corresponds to the expression
13768 "maxval-clipval+minval".
13771 The computed value in @var{val}, clipped to the
13772 @var{minval}-@var{maxval} range.
13774 @item gammaval(gamma)
13775 The computed gamma correction value of the pixel component value,
13776 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13778 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13782 All expressions default to "val".
13784 @subsection Examples
13788 Negate input video:
13790 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13791 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13794 The above is the same as:
13796 lutrgb="r=negval:g=negval:b=negval"
13797 lutyuv="y=negval:u=negval:v=negval"
13807 Remove chroma components, turning the video into a graytone image:
13809 lutyuv="u=128:v=128"
13813 Apply a luma burning effect:
13819 Remove green and blue components:
13825 Set a constant alpha channel value on input:
13827 format=rgba,lutrgb=a="maxval-minval/2"
13831 Correct luminance gamma by a factor of 0.5:
13833 lutyuv=y=gammaval(0.5)
13837 Discard least significant bits of luma:
13839 lutyuv=y='bitand(val, 128+64+32)'
13843 Technicolor like effect:
13845 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13849 @section lut2, tlut2
13851 The @code{lut2} filter takes two input streams and outputs one
13854 The @code{tlut2} (time lut2) filter takes two consecutive frames
13855 from one single stream.
13857 This filter accepts the following parameters:
13860 set first pixel component expression
13862 set second pixel component expression
13864 set third pixel component expression
13866 set fourth pixel component expression, corresponds to the alpha component
13869 set output bit depth, only available for @code{lut2} filter. By default is 0,
13870 which means bit depth is automatically picked from first input format.
13873 The @code{lut2} filter also supports the @ref{framesync} options.
13875 Each of them specifies the expression to use for computing the lookup table for
13876 the corresponding pixel component values.
13878 The exact component associated to each of the @var{c*} options depends on the
13881 The expressions can contain the following constants:
13886 The input width and height.
13889 The first input value for the pixel component.
13892 The second input value for the pixel component.
13895 The first input video bit depth.
13898 The second input video bit depth.
13901 All expressions default to "x".
13903 @subsection Examples
13907 Highlight differences between two RGB video streams:
13909 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)'
13913 Highlight differences between two YUV video streams:
13915 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)'
13919 Show max difference between two video streams:
13921 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)))'
13925 @section maskedclamp
13927 Clamp the first input stream with the second input and third input stream.
13929 Returns the value of first stream to be between second input
13930 stream - @code{undershoot} and third input stream + @code{overshoot}.
13932 This filter accepts the following options:
13935 Default value is @code{0}.
13938 Default value is @code{0}.
13941 Set which planes will be processed as bitmap, unprocessed planes will be
13942 copied from first stream.
13943 By default value 0xf, all planes will be processed.
13948 Merge the second and third input stream into output stream using absolute differences
13949 between second input stream and first input stream and absolute difference between
13950 third input stream and first input stream. The picked value will be from second input
13951 stream if second absolute difference is greater than first one or from third input stream
13954 This filter accepts the following options:
13957 Set which planes will be processed as bitmap, unprocessed planes will be
13958 copied from first stream.
13959 By default value 0xf, all planes will be processed.
13962 @section maskedmerge
13964 Merge the first input stream with the second input stream using per pixel
13965 weights in the third input stream.
13967 A value of 0 in the third stream pixel component means that pixel component
13968 from first stream is returned unchanged, while maximum value (eg. 255 for
13969 8-bit videos) means that pixel component from second stream is returned
13970 unchanged. Intermediate values define the amount of merging between both
13971 input stream's pixel components.
13973 This filter accepts the following options:
13976 Set which planes will be processed as bitmap, unprocessed planes will be
13977 copied from first stream.
13978 By default value 0xf, all planes will be processed.
13983 Merge the second and third input stream into output stream using absolute differences
13984 between second input stream and first input stream and absolute difference between
13985 third input stream and first input stream. The picked value will be from second input
13986 stream if second absolute difference is less than first one or from third input stream
13989 This filter accepts the following options:
13992 Set which planes will be processed as bitmap, unprocessed planes will be
13993 copied from first stream.
13994 By default value 0xf, all planes will be processed.
13997 @section maskedthreshold
13998 Pick pixels comparing absolute difference of two video streams with fixed
14001 If absolute difference between pixel component of first and second video
14002 stream is equal or lower than user supplied threshold than pixel component
14003 from first video stream is picked, otherwise pixel component from second
14004 video stream is picked.
14006 This filter accepts the following options:
14009 Set threshold used when picking pixels from absolute difference from two input
14013 Set which planes will be processed as bitmap, unprocessed planes will be
14014 copied from second stream.
14015 By default value 0xf, all planes will be processed.
14019 Create mask from input video.
14021 For example it is useful to create motion masks after @code{tblend} filter.
14023 This filter accepts the following options:
14027 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14030 Set high threshold. Any pixel component higher than this value will be set to max value
14031 allowed for current pixel format.
14034 Set planes to filter, by default all available planes are filtered.
14037 Fill all frame pixels with this value.
14040 Set max average pixel value for frame. If sum of all pixel components is higher that this
14041 average, output frame will be completely filled with value set by @var{fill} option.
14042 Typically useful for scene changes when used in combination with @code{tblend} filter.
14047 Apply motion-compensation deinterlacing.
14049 It needs one field per frame as input and must thus be used together
14050 with yadif=1/3 or equivalent.
14052 This filter accepts the following options:
14055 Set the deinterlacing mode.
14057 It accepts one of the following values:
14062 use iterative motion estimation
14064 like @samp{slow}, but use multiple reference frames.
14066 Default value is @samp{fast}.
14069 Set the picture field parity assumed for the input video. It must be
14070 one of the following values:
14074 assume top field first
14076 assume bottom field first
14079 Default value is @samp{bff}.
14082 Set per-block quantization parameter (QP) used by the internal
14085 Higher values should result in a smoother motion vector field but less
14086 optimal individual vectors. Default value is 1.
14091 Pick median pixel from certain rectangle defined by radius.
14093 This filter accepts the following options:
14097 Set horizontal radius size. Default value is @code{1}.
14098 Allowed range is integer from 1 to 127.
14101 Set which planes to process. Default is @code{15}, which is all available planes.
14104 Set vertical radius size. Default value is @code{0}.
14105 Allowed range is integer from 0 to 127.
14106 If it is 0, value will be picked from horizontal @code{radius} option.
14109 Set median percentile. Default value is @code{0.5}.
14110 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14111 minimum values, and @code{1} maximum values.
14114 @subsection Commands
14115 This filter supports same @ref{commands} as options.
14116 The command accepts the same syntax of the corresponding option.
14118 If the specified expression is not valid, it is kept at its current
14121 @section mergeplanes
14123 Merge color channel components from several video streams.
14125 The filter accepts up to 4 input streams, and merge selected input
14126 planes to the output video.
14128 This filter accepts the following options:
14131 Set input to output plane mapping. Default is @code{0}.
14133 The mappings is specified as a bitmap. It should be specified as a
14134 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14135 mapping for the first plane of the output stream. 'A' sets the number of
14136 the input stream to use (from 0 to 3), and 'a' the plane number of the
14137 corresponding input to use (from 0 to 3). The rest of the mappings is
14138 similar, 'Bb' describes the mapping for the output stream second
14139 plane, 'Cc' describes the mapping for the output stream third plane and
14140 'Dd' describes the mapping for the output stream fourth plane.
14143 Set output pixel format. Default is @code{yuva444p}.
14146 @subsection Examples
14150 Merge three gray video streams of same width and height into single video stream:
14152 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14156 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14158 [a0][a1]mergeplanes=0x00010210:yuva444p
14162 Swap Y and A plane in yuva444p stream:
14164 format=yuva444p,mergeplanes=0x03010200:yuva444p
14168 Swap U and V plane in yuv420p stream:
14170 format=yuv420p,mergeplanes=0x000201:yuv420p
14174 Cast a rgb24 clip to yuv444p:
14176 format=rgb24,mergeplanes=0x000102:yuv444p
14182 Estimate and export motion vectors using block matching algorithms.
14183 Motion vectors are stored in frame side data to be used by other filters.
14185 This filter accepts the following options:
14188 Specify the motion estimation method. Accepts one of the following values:
14192 Exhaustive search algorithm.
14194 Three step search algorithm.
14196 Two dimensional logarithmic search algorithm.
14198 New three step search algorithm.
14200 Four step search algorithm.
14202 Diamond search algorithm.
14204 Hexagon-based search algorithm.
14206 Enhanced predictive zonal search algorithm.
14208 Uneven multi-hexagon search algorithm.
14210 Default value is @samp{esa}.
14213 Macroblock size. Default @code{16}.
14216 Search parameter. Default @code{7}.
14219 @section midequalizer
14221 Apply Midway Image Equalization effect using two video streams.
14223 Midway Image Equalization adjusts a pair of images to have the same
14224 histogram, while maintaining their dynamics as much as possible. It's
14225 useful for e.g. matching exposures from a pair of stereo cameras.
14227 This filter has two inputs and one output, which must be of same pixel format, but
14228 may be of different sizes. The output of filter is first input adjusted with
14229 midway histogram of both inputs.
14231 This filter accepts the following option:
14235 Set which planes to process. Default is @code{15}, which is all available planes.
14238 @section minterpolate
14240 Convert the video to specified frame rate using motion interpolation.
14242 This filter accepts the following options:
14245 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}.
14248 Motion interpolation mode. Following values are accepted:
14251 Duplicate previous or next frame for interpolating new ones.
14253 Blend source frames. Interpolated frame is mean of previous and next frames.
14255 Motion compensated interpolation. Following options are effective when this mode is selected:
14259 Motion compensation mode. Following values are accepted:
14262 Overlapped block motion compensation.
14264 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14266 Default mode is @samp{obmc}.
14269 Motion estimation mode. Following values are accepted:
14272 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14274 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14276 Default mode is @samp{bilat}.
14279 The algorithm to be used for motion estimation. Following values are accepted:
14282 Exhaustive search algorithm.
14284 Three step search algorithm.
14286 Two dimensional logarithmic search algorithm.
14288 New three step search algorithm.
14290 Four step search algorithm.
14292 Diamond search algorithm.
14294 Hexagon-based search algorithm.
14296 Enhanced predictive zonal search algorithm.
14298 Uneven multi-hexagon search algorithm.
14300 Default algorithm is @samp{epzs}.
14303 Macroblock size. Default @code{16}.
14306 Motion estimation search parameter. Default @code{32}.
14309 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).
14314 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:
14317 Disable scene change detection.
14319 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14321 Default method is @samp{fdiff}.
14323 @item scd_threshold
14324 Scene change detection threshold. Default is @code{10.}.
14329 Mix several video input streams into one video stream.
14331 A description of the accepted options follows.
14335 The number of inputs. If unspecified, it defaults to 2.
14338 Specify weight of each input video stream as sequence.
14339 Each weight is separated by space. If number of weights
14340 is smaller than number of @var{frames} last specified
14341 weight will be used for all remaining unset weights.
14344 Specify scale, if it is set it will be multiplied with sum
14345 of each weight multiplied with pixel values to give final destination
14346 pixel value. By default @var{scale} is auto scaled to sum of weights.
14349 Specify how end of stream is determined.
14352 The duration of the longest input. (default)
14355 The duration of the shortest input.
14358 The duration of the first input.
14362 @section mpdecimate
14364 Drop frames that do not differ greatly from the previous frame in
14365 order to reduce frame rate.
14367 The main use of this filter is for very-low-bitrate encoding
14368 (e.g. streaming over dialup modem), but it could in theory be used for
14369 fixing movies that were inverse-telecined incorrectly.
14371 A description of the accepted options follows.
14375 Set the maximum number of consecutive frames which can be dropped (if
14376 positive), or the minimum interval between dropped frames (if
14377 negative). If the value is 0, the frame is dropped disregarding the
14378 number of previous sequentially dropped frames.
14380 Default value is 0.
14385 Set the dropping threshold values.
14387 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14388 represent actual pixel value differences, so a threshold of 64
14389 corresponds to 1 unit of difference for each pixel, or the same spread
14390 out differently over the block.
14392 A frame is a candidate for dropping if no 8x8 blocks differ by more
14393 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14394 meaning the whole image) differ by more than a threshold of @option{lo}.
14396 Default value for @option{hi} is 64*12, default value for @option{lo} is
14397 64*5, and default value for @option{frac} is 0.33.
14403 Negate (invert) the input video.
14405 It accepts the following option:
14410 With value 1, it negates the alpha component, if present. Default value is 0.
14416 Denoise frames using Non-Local Means algorithm.
14418 Each pixel is adjusted by looking for other pixels with similar contexts. This
14419 context similarity is defined by comparing their surrounding patches of size
14420 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14423 Note that the research area defines centers for patches, which means some
14424 patches will be made of pixels outside that research area.
14426 The filter accepts the following options.
14430 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14433 Set patch size. Default is 7. Must be odd number in range [0, 99].
14436 Same as @option{p} but for chroma planes.
14438 The default value is @var{0} and means automatic.
14441 Set research size. Default is 15. Must be odd number in range [0, 99].
14444 Same as @option{r} but for chroma planes.
14446 The default value is @var{0} and means automatic.
14451 Deinterlace video using neural network edge directed interpolation.
14453 This filter accepts the following options:
14457 Mandatory option, without binary file filter can not work.
14458 Currently file can be found here:
14459 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14462 Set which frames to deinterlace, by default it is @code{all}.
14463 Can be @code{all} or @code{interlaced}.
14466 Set mode of operation.
14468 Can be one of the following:
14472 Use frame flags, both fields.
14474 Use frame flags, single field.
14476 Use top field only.
14478 Use bottom field only.
14480 Use both fields, top first.
14482 Use both fields, bottom first.
14486 Set which planes to process, by default filter process all frames.
14489 Set size of local neighborhood around each pixel, used by the predictor neural
14492 Can be one of the following:
14505 Set the number of neurons in predictor neural network.
14506 Can be one of the following:
14517 Controls the number of different neural network predictions that are blended
14518 together to compute the final output value. Can be @code{fast}, default or
14522 Set which set of weights to use in the predictor.
14523 Can be one of the following:
14527 weights trained to minimize absolute error
14529 weights trained to minimize squared error
14533 Controls whether or not the prescreener neural network is used to decide
14534 which pixels should be processed by the predictor neural network and which
14535 can be handled by simple cubic interpolation.
14536 The prescreener is trained to know whether cubic interpolation will be
14537 sufficient for a pixel or whether it should be predicted by the predictor nn.
14538 The computational complexity of the prescreener nn is much less than that of
14539 the predictor nn. Since most pixels can be handled by cubic interpolation,
14540 using the prescreener generally results in much faster processing.
14541 The prescreener is pretty accurate, so the difference between using it and not
14542 using it is almost always unnoticeable.
14544 Can be one of the following:
14552 Default is @code{new}.
14555 Set various debugging flags.
14560 Force libavfilter not to use any of the specified pixel formats for the
14561 input to the next filter.
14563 It accepts the following parameters:
14567 A '|'-separated list of pixel format names, such as
14568 pix_fmts=yuv420p|monow|rgb24".
14572 @subsection Examples
14576 Force libavfilter to use a format different from @var{yuv420p} for the
14577 input to the vflip filter:
14579 noformat=pix_fmts=yuv420p,vflip
14583 Convert the input video to any of the formats not contained in the list:
14585 noformat=yuv420p|yuv444p|yuv410p
14591 Add noise on video input frame.
14593 The filter accepts the following options:
14601 Set noise seed for specific pixel component or all pixel components in case
14602 of @var{all_seed}. Default value is @code{123457}.
14604 @item all_strength, alls
14605 @item c0_strength, c0s
14606 @item c1_strength, c1s
14607 @item c2_strength, c2s
14608 @item c3_strength, c3s
14609 Set noise strength for specific pixel component or all pixel components in case
14610 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14612 @item all_flags, allf
14613 @item c0_flags, c0f
14614 @item c1_flags, c1f
14615 @item c2_flags, c2f
14616 @item c3_flags, c3f
14617 Set pixel component flags or set flags for all components if @var{all_flags}.
14618 Available values for component flags are:
14621 averaged temporal noise (smoother)
14623 mix random noise with a (semi)regular pattern
14625 temporal noise (noise pattern changes between frames)
14627 uniform noise (gaussian otherwise)
14631 @subsection Examples
14633 Add temporal and uniform noise to input video:
14635 noise=alls=20:allf=t+u
14640 Normalize RGB video (aka histogram stretching, contrast stretching).
14641 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14643 For each channel of each frame, the filter computes the input range and maps
14644 it linearly to the user-specified output range. The output range defaults
14645 to the full dynamic range from pure black to pure white.
14647 Temporal smoothing can be used on the input range to reduce flickering (rapid
14648 changes in brightness) caused when small dark or bright objects enter or leave
14649 the scene. This is similar to the auto-exposure (automatic gain control) on a
14650 video camera, and, like a video camera, it may cause a period of over- or
14651 under-exposure of the video.
14653 The R,G,B channels can be normalized independently, which may cause some
14654 color shifting, or linked together as a single channel, which prevents
14655 color shifting. Linked normalization preserves hue. Independent normalization
14656 does not, so it can be used to remove some color casts. Independent and linked
14657 normalization can be combined in any ratio.
14659 The normalize filter accepts the following options:
14664 Colors which define the output range. The minimum input value is mapped to
14665 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14666 The defaults are black and white respectively. Specifying white for
14667 @var{blackpt} and black for @var{whitept} will give color-inverted,
14668 normalized video. Shades of grey can be used to reduce the dynamic range
14669 (contrast). Specifying saturated colors here can create some interesting
14673 The number of previous frames to use for temporal smoothing. The input range
14674 of each channel is smoothed using a rolling average over the current frame
14675 and the @var{smoothing} previous frames. The default is 0 (no temporal
14679 Controls the ratio of independent (color shifting) channel normalization to
14680 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14681 independent. Defaults to 1.0 (fully independent).
14684 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14685 expensive no-op. Defaults to 1.0 (full strength).
14689 @subsection Commands
14690 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14691 The command accepts the same syntax of the corresponding option.
14693 If the specified expression is not valid, it is kept at its current
14696 @subsection Examples
14698 Stretch video contrast to use the full dynamic range, with no temporal
14699 smoothing; may flicker depending on the source content:
14701 normalize=blackpt=black:whitept=white:smoothing=0
14704 As above, but with 50 frames of temporal smoothing; flicker should be
14705 reduced, depending on the source content:
14707 normalize=blackpt=black:whitept=white:smoothing=50
14710 As above, but with hue-preserving linked channel normalization:
14712 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14715 As above, but with half strength:
14717 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14720 Map the darkest input color to red, the brightest input color to cyan:
14722 normalize=blackpt=red:whitept=cyan
14727 Pass the video source unchanged to the output.
14730 Optical Character Recognition
14732 This filter uses Tesseract for optical character recognition. To enable
14733 compilation of this filter, you need to configure FFmpeg with
14734 @code{--enable-libtesseract}.
14736 It accepts the following options:
14740 Set datapath to tesseract data. Default is to use whatever was
14741 set at installation.
14744 Set language, default is "eng".
14747 Set character whitelist.
14750 Set character blacklist.
14753 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14754 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14758 Apply a video transform using libopencv.
14760 To enable this filter, install the libopencv library and headers and
14761 configure FFmpeg with @code{--enable-libopencv}.
14763 It accepts the following parameters:
14768 The name of the libopencv filter to apply.
14770 @item filter_params
14771 The parameters to pass to the libopencv filter. If not specified, the default
14772 values are assumed.
14776 Refer to the official libopencv documentation for more precise
14778 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14780 Several libopencv filters are supported; see the following subsections.
14785 Dilate an image by using a specific structuring element.
14786 It corresponds to the libopencv function @code{cvDilate}.
14788 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14790 @var{struct_el} represents a structuring element, and has the syntax:
14791 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14793 @var{cols} and @var{rows} represent the number of columns and rows of
14794 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14795 point, and @var{shape} the shape for the structuring element. @var{shape}
14796 must be "rect", "cross", "ellipse", or "custom".
14798 If the value for @var{shape} is "custom", it must be followed by a
14799 string of the form "=@var{filename}". The file with name
14800 @var{filename} is assumed to represent a binary image, with each
14801 printable character corresponding to a bright pixel. When a custom
14802 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14803 or columns and rows of the read file are assumed instead.
14805 The default value for @var{struct_el} is "3x3+0x0/rect".
14807 @var{nb_iterations} specifies the number of times the transform is
14808 applied to the image, and defaults to 1.
14812 # Use the default values
14815 # Dilate using a structuring element with a 5x5 cross, iterating two times
14816 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14818 # Read the shape from the file diamond.shape, iterating two times.
14819 # The file diamond.shape may contain a pattern of characters like this
14825 # The specified columns and rows are ignored
14826 # but the anchor point coordinates are not
14827 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14832 Erode an image by using a specific structuring element.
14833 It corresponds to the libopencv function @code{cvErode}.
14835 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14836 with the same syntax and semantics as the @ref{dilate} filter.
14840 Smooth the input video.
14842 The filter takes the following parameters:
14843 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14845 @var{type} is the type of smooth filter to apply, and must be one of
14846 the following values: "blur", "blur_no_scale", "median", "gaussian",
14847 or "bilateral". The default value is "gaussian".
14849 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14850 depends on the smooth type. @var{param1} and
14851 @var{param2} accept integer positive values or 0. @var{param3} and
14852 @var{param4} accept floating point values.
14854 The default value for @var{param1} is 3. The default value for the
14855 other parameters is 0.
14857 These parameters correspond to the parameters assigned to the
14858 libopencv function @code{cvSmooth}.
14860 @section oscilloscope
14862 2D Video Oscilloscope.
14864 Useful to measure spatial impulse, step responses, chroma delays, etc.
14866 It accepts the following parameters:
14870 Set scope center x position.
14873 Set scope center y position.
14876 Set scope size, relative to frame diagonal.
14879 Set scope tilt/rotation.
14885 Set trace center x position.
14888 Set trace center y position.
14891 Set trace width, relative to width of frame.
14894 Set trace height, relative to height of frame.
14897 Set which components to trace. By default it traces first three components.
14900 Draw trace grid. By default is enabled.
14903 Draw some statistics. By default is enabled.
14906 Draw scope. By default is enabled.
14909 @subsection Commands
14910 This filter supports same @ref{commands} as options.
14911 The command accepts the same syntax of the corresponding option.
14913 If the specified expression is not valid, it is kept at its current
14916 @subsection Examples
14920 Inspect full first row of video frame.
14922 oscilloscope=x=0.5:y=0:s=1
14926 Inspect full last row of video frame.
14928 oscilloscope=x=0.5:y=1:s=1
14932 Inspect full 5th line of video frame of height 1080.
14934 oscilloscope=x=0.5:y=5/1080:s=1
14938 Inspect full last column of video frame.
14940 oscilloscope=x=1:y=0.5:s=1:t=1
14948 Overlay one video on top of another.
14950 It takes two inputs and has one output. The first input is the "main"
14951 video on which the second input is overlaid.
14953 It accepts the following parameters:
14955 A description of the accepted options follows.
14960 Set the expression for the x and y coordinates of the overlaid video
14961 on the main video. Default value is "0" for both expressions. In case
14962 the expression is invalid, it is set to a huge value (meaning that the
14963 overlay will not be displayed within the output visible area).
14966 See @ref{framesync}.
14969 Set when the expressions for @option{x}, and @option{y} are evaluated.
14971 It accepts the following values:
14974 only evaluate expressions once during the filter initialization or
14975 when a command is processed
14978 evaluate expressions for each incoming frame
14981 Default value is @samp{frame}.
14984 See @ref{framesync}.
14987 Set the format for the output video.
14989 It accepts the following values:
14992 force YUV420 output
14995 force YUV420p10 output
14998 force YUV422 output
15001 force YUV422p10 output
15004 force YUV444 output
15007 force packed RGB output
15010 force planar RGB output
15013 automatically pick format
15016 Default value is @samp{yuv420}.
15019 See @ref{framesync}.
15022 Set format of alpha of the overlaid video, it can be @var{straight} or
15023 @var{premultiplied}. Default is @var{straight}.
15026 The @option{x}, and @option{y} expressions can contain the following
15032 The main input width and height.
15036 The overlay input width and height.
15040 The computed values for @var{x} and @var{y}. They are evaluated for
15045 horizontal and vertical chroma subsample values of the output
15046 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15050 the number of input frame, starting from 0
15053 the position in the file of the input frame, NAN if unknown
15056 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15060 This filter also supports the @ref{framesync} options.
15062 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15063 when evaluation is done @emph{per frame}, and will evaluate to NAN
15064 when @option{eval} is set to @samp{init}.
15066 Be aware that frames are taken from each input video in timestamp
15067 order, hence, if their initial timestamps differ, it is a good idea
15068 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15069 have them begin in the same zero timestamp, as the example for
15070 the @var{movie} filter does.
15072 You can chain together more overlays but you should test the
15073 efficiency of such approach.
15075 @subsection Commands
15077 This filter supports the following commands:
15081 Modify the x and y of the overlay input.
15082 The command accepts the same syntax of the corresponding option.
15084 If the specified expression is not valid, it is kept at its current
15088 @subsection Examples
15092 Draw the overlay at 10 pixels from the bottom right corner of the main
15095 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15098 Using named options the example above becomes:
15100 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15104 Insert a transparent PNG logo in the bottom left corner of the input,
15105 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15107 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15111 Insert 2 different transparent PNG logos (second logo on bottom
15112 right corner) using the @command{ffmpeg} tool:
15114 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
15118 Add a transparent color layer on top of the main video; @code{WxH}
15119 must specify the size of the main input to the overlay filter:
15121 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15125 Play an original video and a filtered version (here with the deshake
15126 filter) side by side using the @command{ffplay} tool:
15128 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15131 The above command is the same as:
15133 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15137 Make a sliding overlay appearing from the left to the right top part of the
15138 screen starting since time 2:
15140 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15144 Compose output by putting two input videos side to side:
15146 ffmpeg -i left.avi -i right.avi -filter_complex "
15147 nullsrc=size=200x100 [background];
15148 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15149 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15150 [background][left] overlay=shortest=1 [background+left];
15151 [background+left][right] overlay=shortest=1:x=100 [left+right]
15156 Mask 10-20 seconds of a video by applying the delogo filter to a section
15158 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15159 -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]'
15164 Chain several overlays in cascade:
15166 nullsrc=s=200x200 [bg];
15167 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15168 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15169 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15170 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15171 [in3] null, [mid2] overlay=100:100 [out0]
15176 @anchor{overlay_cuda}
15177 @section overlay_cuda
15179 Overlay one video on top of another.
15181 This is the CUDA variant of the @ref{overlay} filter.
15182 It only accepts CUDA frames. The underlying input pixel formats have to match.
15184 It takes two inputs and has one output. The first input is the "main"
15185 video on which the second input is overlaid.
15187 It accepts the following parameters:
15192 Set the x and y coordinates of the overlaid video on the main video.
15193 Default value is "0" for both expressions.
15196 See @ref{framesync}.
15199 See @ref{framesync}.
15202 See @ref{framesync}.
15206 This filter also supports the @ref{framesync} options.
15210 Apply Overcomplete Wavelet denoiser.
15212 The filter accepts the following options:
15218 Larger depth values will denoise lower frequency components more, but
15219 slow down filtering.
15221 Must be an int in the range 8-16, default is @code{8}.
15223 @item luma_strength, ls
15226 Must be a double value in the range 0-1000, default is @code{1.0}.
15228 @item chroma_strength, cs
15229 Set chroma strength.
15231 Must be a double value in the range 0-1000, default is @code{1.0}.
15237 Add paddings to the input image, and place the original input at the
15238 provided @var{x}, @var{y} coordinates.
15240 It accepts the following parameters:
15245 Specify an expression for the size of the output image with the
15246 paddings added. If the value for @var{width} or @var{height} is 0, the
15247 corresponding input size is used for the output.
15249 The @var{width} expression can reference the value set by the
15250 @var{height} expression, and vice versa.
15252 The default value of @var{width} and @var{height} is 0.
15256 Specify the offsets to place the input image at within the padded area,
15257 with respect to the top/left border of the output image.
15259 The @var{x} expression can reference the value set by the @var{y}
15260 expression, and vice versa.
15262 The default value of @var{x} and @var{y} is 0.
15264 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15265 so the input image is centered on the padded area.
15268 Specify the color of the padded area. For the syntax of this option,
15269 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15270 manual,ffmpeg-utils}.
15272 The default value of @var{color} is "black".
15275 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15277 It accepts the following values:
15281 Only evaluate expressions once during the filter initialization or when
15282 a command is processed.
15285 Evaluate expressions for each incoming frame.
15289 Default value is @samp{init}.
15292 Pad to aspect instead to a resolution.
15296 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15297 options are expressions containing the following constants:
15302 The input video width and height.
15306 These are the same as @var{in_w} and @var{in_h}.
15310 The output width and height (the size of the padded area), as
15311 specified by the @var{width} and @var{height} expressions.
15315 These are the same as @var{out_w} and @var{out_h}.
15319 The x and y offsets as specified by the @var{x} and @var{y}
15320 expressions, or NAN if not yet specified.
15323 same as @var{iw} / @var{ih}
15326 input sample aspect ratio
15329 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15333 The horizontal and vertical chroma subsample values. For example for the
15334 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15337 @subsection Examples
15341 Add paddings with the color "violet" to the input video. The output video
15342 size is 640x480, and the top-left corner of the input video is placed at
15345 pad=640:480:0:40:violet
15348 The example above is equivalent to the following command:
15350 pad=width=640:height=480:x=0:y=40:color=violet
15354 Pad the input to get an output with dimensions increased by 3/2,
15355 and put the input video at the center of the padded area:
15357 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15361 Pad the input to get a squared output with size equal to the maximum
15362 value between the input width and height, and put the input video at
15363 the center of the padded area:
15365 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15369 Pad the input to get a final w/h ratio of 16:9:
15371 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15375 In case of anamorphic video, in order to set the output display aspect
15376 correctly, it is necessary to use @var{sar} in the expression,
15377 according to the relation:
15379 (ih * X / ih) * sar = output_dar
15380 X = output_dar / sar
15383 Thus the previous example needs to be modified to:
15385 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15389 Double the output size and put the input video in the bottom-right
15390 corner of the output padded area:
15392 pad="2*iw:2*ih:ow-iw:oh-ih"
15396 @anchor{palettegen}
15397 @section palettegen
15399 Generate one palette for a whole video stream.
15401 It accepts the following options:
15405 Set the maximum number of colors to quantize in the palette.
15406 Note: the palette will still contain 256 colors; the unused palette entries
15409 @item reserve_transparent
15410 Create a palette of 255 colors maximum and reserve the last one for
15411 transparency. Reserving the transparency color is useful for GIF optimization.
15412 If not set, the maximum of colors in the palette will be 256. You probably want
15413 to disable this option for a standalone image.
15416 @item transparency_color
15417 Set the color that will be used as background for transparency.
15420 Set statistics mode.
15422 It accepts the following values:
15425 Compute full frame histograms.
15427 Compute histograms only for the part that differs from previous frame. This
15428 might be relevant to give more importance to the moving part of your input if
15429 the background is static.
15431 Compute new histogram for each frame.
15434 Default value is @var{full}.
15437 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15438 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15439 color quantization of the palette. This information is also visible at
15440 @var{info} logging level.
15442 @subsection Examples
15446 Generate a representative palette of a given video using @command{ffmpeg}:
15448 ffmpeg -i input.mkv -vf palettegen palette.png
15452 @section paletteuse
15454 Use a palette to downsample an input video stream.
15456 The filter takes two inputs: one video stream and a palette. The palette must
15457 be a 256 pixels image.
15459 It accepts the following options:
15463 Select dithering mode. Available algorithms are:
15466 Ordered 8x8 bayer dithering (deterministic)
15468 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15469 Note: this dithering is sometimes considered "wrong" and is included as a
15471 @item floyd_steinberg
15472 Floyd and Steingberg dithering (error diffusion)
15474 Frankie Sierra dithering v2 (error diffusion)
15476 Frankie Sierra dithering v2 "Lite" (error diffusion)
15479 Default is @var{sierra2_4a}.
15482 When @var{bayer} dithering is selected, this option defines the scale of the
15483 pattern (how much the crosshatch pattern is visible). A low value means more
15484 visible pattern for less banding, and higher value means less visible pattern
15485 at the cost of more banding.
15487 The option must be an integer value in the range [0,5]. Default is @var{2}.
15490 If set, define the zone to process
15494 Only the changing rectangle will be reprocessed. This is similar to GIF
15495 cropping/offsetting compression mechanism. This option can be useful for speed
15496 if only a part of the image is changing, and has use cases such as limiting the
15497 scope of the error diffusal @option{dither} to the rectangle that bounds the
15498 moving scene (it leads to more deterministic output if the scene doesn't change
15499 much, and as a result less moving noise and better GIF compression).
15502 Default is @var{none}.
15505 Take new palette for each output frame.
15507 @item alpha_threshold
15508 Sets the alpha threshold for transparency. Alpha values above this threshold
15509 will be treated as completely opaque, and values below this threshold will be
15510 treated as completely transparent.
15512 The option must be an integer value in the range [0,255]. Default is @var{128}.
15515 @subsection Examples
15519 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15520 using @command{ffmpeg}:
15522 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15526 @section perspective
15528 Correct perspective of video not recorded perpendicular to the screen.
15530 A description of the accepted parameters follows.
15541 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15542 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15543 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15544 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15545 then the corners of the source will be sent to the specified coordinates.
15547 The expressions can use the following variables:
15552 the width and height of video frame.
15556 Output frame count.
15559 @item interpolation
15560 Set interpolation for perspective correction.
15562 It accepts the following values:
15568 Default value is @samp{linear}.
15571 Set interpretation of coordinate options.
15573 It accepts the following values:
15577 Send point in the source specified by the given coordinates to
15578 the corners of the destination.
15580 @item 1, destination
15582 Send the corners of the source to the point in the destination specified
15583 by the given coordinates.
15585 Default value is @samp{source}.
15589 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15591 It accepts the following values:
15594 only evaluate expressions once during the filter initialization or
15595 when a command is processed
15598 evaluate expressions for each incoming frame
15601 Default value is @samp{init}.
15606 Delay interlaced video by one field time so that the field order changes.
15608 The intended use is to fix PAL movies that have been captured with the
15609 opposite field order to the film-to-video transfer.
15611 A description of the accepted parameters follows.
15617 It accepts the following values:
15620 Capture field order top-first, transfer bottom-first.
15621 Filter will delay the bottom field.
15624 Capture field order bottom-first, transfer top-first.
15625 Filter will delay the top field.
15628 Capture and transfer with the same field order. This mode only exists
15629 for the documentation of the other options to refer to, but if you
15630 actually select it, the filter will faithfully do nothing.
15633 Capture field order determined automatically by field flags, transfer
15635 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15636 basis using field flags. If no field information is available,
15637 then this works just like @samp{u}.
15640 Capture unknown or varying, transfer opposite.
15641 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15642 analyzing the images and selecting the alternative that produces best
15643 match between the fields.
15646 Capture top-first, transfer unknown or varying.
15647 Filter selects among @samp{t} and @samp{p} using image analysis.
15650 Capture bottom-first, transfer unknown or varying.
15651 Filter selects among @samp{b} and @samp{p} using image analysis.
15654 Capture determined by field flags, transfer unknown or varying.
15655 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15656 image analysis. If no field information is available, then this works just
15657 like @samp{U}. This is the default mode.
15660 Both capture and transfer unknown or varying.
15661 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15665 @section photosensitivity
15666 Reduce various flashes in video, so to help users with epilepsy.
15668 It accepts the following options:
15671 Set how many frames to use when filtering. Default is 30.
15674 Set detection threshold factor. Default is 1.
15678 Set how many pixels to skip when sampling frames. Default is 1.
15679 Allowed range is from 1 to 1024.
15682 Leave frames unchanged. Default is disabled.
15685 @section pixdesctest
15687 Pixel format descriptor test filter, mainly useful for internal
15688 testing. The output video should be equal to the input video.
15692 format=monow, pixdesctest
15695 can be used to test the monowhite pixel format descriptor definition.
15699 Display sample values of color channels. Mainly useful for checking color
15700 and levels. Minimum supported resolution is 640x480.
15702 The filters accept the following options:
15706 Set scope X position, relative offset on X axis.
15709 Set scope Y position, relative offset on Y axis.
15718 Set window opacity. This window also holds statistics about pixel area.
15721 Set window X position, relative offset on X axis.
15724 Set window Y position, relative offset on Y axis.
15729 Enable the specified chain of postprocessing subfilters using libpostproc. This
15730 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15731 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15732 Each subfilter and some options have a short and a long name that can be used
15733 interchangeably, i.e. dr/dering are the same.
15735 The filters accept the following options:
15739 Set postprocessing subfilters string.
15742 All subfilters share common options to determine their scope:
15746 Honor the quality commands for this subfilter.
15749 Do chrominance filtering, too (default).
15752 Do luminance filtering only (no chrominance).
15755 Do chrominance filtering only (no luminance).
15758 These options can be appended after the subfilter name, separated by a '|'.
15760 Available subfilters are:
15763 @item hb/hdeblock[|difference[|flatness]]
15764 Horizontal deblocking filter
15767 Difference factor where higher values mean more deblocking (default: @code{32}).
15769 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15772 @item vb/vdeblock[|difference[|flatness]]
15773 Vertical deblocking filter
15776 Difference factor where higher values mean more deblocking (default: @code{32}).
15778 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15781 @item ha/hadeblock[|difference[|flatness]]
15782 Accurate horizontal deblocking filter
15785 Difference factor where higher values mean more deblocking (default: @code{32}).
15787 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15790 @item va/vadeblock[|difference[|flatness]]
15791 Accurate vertical deblocking filter
15794 Difference factor where higher values mean more deblocking (default: @code{32}).
15796 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15800 The horizontal and vertical deblocking filters share the difference and
15801 flatness values so you cannot set different horizontal and vertical
15805 @item h1/x1hdeblock
15806 Experimental horizontal deblocking filter
15808 @item v1/x1vdeblock
15809 Experimental vertical deblocking filter
15814 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15817 larger -> stronger filtering
15819 larger -> stronger filtering
15821 larger -> stronger filtering
15824 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15827 Stretch luminance to @code{0-255}.
15830 @item lb/linblenddeint
15831 Linear blend deinterlacing filter that deinterlaces the given block by
15832 filtering all lines with a @code{(1 2 1)} filter.
15834 @item li/linipoldeint
15835 Linear interpolating deinterlacing filter that deinterlaces the given block by
15836 linearly interpolating every second line.
15838 @item ci/cubicipoldeint
15839 Cubic interpolating deinterlacing filter deinterlaces the given block by
15840 cubically interpolating every second line.
15842 @item md/mediandeint
15843 Median deinterlacing filter that deinterlaces the given block by applying a
15844 median filter to every second line.
15846 @item fd/ffmpegdeint
15847 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15848 second line with a @code{(-1 4 2 4 -1)} filter.
15851 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15852 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15854 @item fq/forceQuant[|quantizer]
15855 Overrides the quantizer table from the input with the constant quantizer you
15863 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15866 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15869 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15872 @subsection Examples
15876 Apply horizontal and vertical deblocking, deringing and automatic
15877 brightness/contrast:
15883 Apply default filters without brightness/contrast correction:
15889 Apply default filters and temporal denoiser:
15891 pp=default/tmpnoise|1|2|3
15895 Apply deblocking on luminance only, and switch vertical deblocking on or off
15896 automatically depending on available CPU time:
15903 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15904 similar to spp = 6 with 7 point DCT, where only the center sample is
15907 The filter accepts the following options:
15911 Force a constant quantization parameter. It accepts an integer in range
15912 0 to 63. If not set, the filter will use the QP from the video stream
15916 Set thresholding mode. Available modes are:
15920 Set hard thresholding.
15922 Set soft thresholding (better de-ringing effect, but likely blurrier).
15924 Set medium thresholding (good results, default).
15928 @section premultiply
15929 Apply alpha premultiply effect to input video stream using first plane
15930 of second stream as alpha.
15932 Both streams must have same dimensions and same pixel format.
15934 The filter accepts the following option:
15938 Set which planes will be processed, unprocessed planes will be copied.
15939 By default value 0xf, all planes will be processed.
15942 Do not require 2nd input for processing, instead use alpha plane from input stream.
15946 Apply prewitt operator to input video stream.
15948 The filter accepts the following option:
15952 Set which planes will be processed, unprocessed planes will be copied.
15953 By default value 0xf, all planes will be processed.
15956 Set value which will be multiplied with filtered result.
15959 Set value which will be added to filtered result.
15962 @subsection Commands
15964 This filter supports the all above options as @ref{commands}.
15966 @section pseudocolor
15968 Alter frame colors in video with pseudocolors.
15970 This filter accepts the following options:
15974 set pixel first component expression
15977 set pixel second component expression
15980 set pixel third component expression
15983 set pixel fourth component expression, corresponds to the alpha component
15986 set component to use as base for altering colors
15989 Each of them specifies the expression to use for computing the lookup table for
15990 the corresponding pixel component values.
15992 The expressions can contain the following constants and functions:
15997 The input width and height.
16000 The input value for the pixel component.
16002 @item ymin, umin, vmin, amin
16003 The minimum allowed component value.
16005 @item ymax, umax, vmax, amax
16006 The maximum allowed component value.
16009 All expressions default to "val".
16011 @subsection Examples
16015 Change too high luma values to gradient:
16017 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'"
16023 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16024 Ratio) between two input videos.
16026 This filter takes in input two input videos, the first input is
16027 considered the "main" source and is passed unchanged to the
16028 output. The second input is used as a "reference" video for computing
16031 Both video inputs must have the same resolution and pixel format for
16032 this filter to work correctly. Also it assumes that both inputs
16033 have the same number of frames, which are compared one by one.
16035 The obtained average PSNR is printed through the logging system.
16037 The filter stores the accumulated MSE (mean squared error) of each
16038 frame, and at the end of the processing it is averaged across all frames
16039 equally, and the following formula is applied to obtain the PSNR:
16042 PSNR = 10*log10(MAX^2/MSE)
16045 Where MAX is the average of the maximum values of each component of the
16048 The description of the accepted parameters follows.
16051 @item stats_file, f
16052 If specified the filter will use the named file to save the PSNR of
16053 each individual frame. When filename equals "-" the data is sent to
16056 @item stats_version
16057 Specifies which version of the stats file format to use. Details of
16058 each format are written below.
16059 Default value is 1.
16061 @item stats_add_max
16062 Determines whether the max value is output to the stats log.
16063 Default value is 0.
16064 Requires stats_version >= 2. If this is set and stats_version < 2,
16065 the filter will return an error.
16068 This filter also supports the @ref{framesync} options.
16070 The file printed if @var{stats_file} is selected, contains a sequence of
16071 key/value pairs of the form @var{key}:@var{value} for each compared
16074 If a @var{stats_version} greater than 1 is specified, a header line precedes
16075 the list of per-frame-pair stats, with key value pairs following the frame
16076 format with the following parameters:
16079 @item psnr_log_version
16080 The version of the log file format. Will match @var{stats_version}.
16083 A comma separated list of the per-frame-pair parameters included in
16087 A description of each shown per-frame-pair parameter follows:
16091 sequential number of the input frame, starting from 1
16094 Mean Square Error pixel-by-pixel average difference of the compared
16095 frames, averaged over all the image components.
16097 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16098 Mean Square Error pixel-by-pixel average difference of the compared
16099 frames for the component specified by the suffix.
16101 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16102 Peak Signal to Noise ratio of the compared frames for the component
16103 specified by the suffix.
16105 @item max_avg, max_y, max_u, max_v
16106 Maximum allowed value for each channel, and average over all
16110 @subsection Examples
16115 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16116 [main][ref] psnr="stats_file=stats.log" [out]
16119 On this example the input file being processed is compared with the
16120 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16121 is stored in @file{stats.log}.
16124 Another example with different containers:
16126 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 -
16133 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16134 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16137 The pullup filter is designed to take advantage of future context in making
16138 its decisions. This filter is stateless in the sense that it does not lock
16139 onto a pattern to follow, but it instead looks forward to the following
16140 fields in order to identify matches and rebuild progressive frames.
16142 To produce content with an even framerate, insert the fps filter after
16143 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16144 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16146 The filter accepts the following options:
16153 These options set the amount of "junk" to ignore at the left, right, top, and
16154 bottom of the image, respectively. Left and right are in units of 8 pixels,
16155 while top and bottom are in units of 2 lines.
16156 The default is 8 pixels on each side.
16159 Set the strict breaks. Setting this option to 1 will reduce the chances of
16160 filter generating an occasional mismatched frame, but it may also cause an
16161 excessive number of frames to be dropped during high motion sequences.
16162 Conversely, setting it to -1 will make filter match fields more easily.
16163 This may help processing of video where there is slight blurring between
16164 the fields, but may also cause there to be interlaced frames in the output.
16165 Default value is @code{0}.
16168 Set the metric plane to use. It accepts the following values:
16174 Use chroma blue plane.
16177 Use chroma red plane.
16180 This option may be set to use chroma plane instead of the default luma plane
16181 for doing filter's computations. This may improve accuracy on very clean
16182 source material, but more likely will decrease accuracy, especially if there
16183 is chroma noise (rainbow effect) or any grayscale video.
16184 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16185 load and make pullup usable in realtime on slow machines.
16188 For best results (without duplicated frames in the output file) it is
16189 necessary to change the output frame rate. For example, to inverse
16190 telecine NTSC input:
16192 ffmpeg -i input -vf pullup -r 24000/1001 ...
16197 Change video quantization parameters (QP).
16199 The filter accepts the following option:
16203 Set expression for quantization parameter.
16206 The expression is evaluated through the eval API and can contain, among others,
16207 the following constants:
16211 1 if index is not 129, 0 otherwise.
16214 Sequential index starting from -129 to 128.
16217 @subsection Examples
16221 Some equation like:
16229 Flush video frames from internal cache of frames into a random order.
16230 No frame is discarded.
16231 Inspired by @ref{frei0r} nervous filter.
16235 Set size in number of frames of internal cache, in range from @code{2} to
16236 @code{512}. Default is @code{30}.
16239 Set seed for random number generator, must be an integer included between
16240 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16241 less than @code{0}, the filter will try to use a good random seed on a
16245 @section readeia608
16247 Read closed captioning (EIA-608) information from the top lines of a video frame.
16249 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16250 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16251 with EIA-608 data (starting from 0). A description of each metadata value follows:
16254 @item lavfi.readeia608.X.cc
16255 The two bytes stored as EIA-608 data (printed in hexadecimal).
16257 @item lavfi.readeia608.X.line
16258 The number of the line on which the EIA-608 data was identified and read.
16261 This filter accepts the following options:
16265 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16268 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16271 Set the ratio of width reserved for sync code detection.
16272 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16275 Enable checking the parity bit. In the event of a parity error, the filter will output
16276 @code{0x00} for that character. Default is false.
16279 Lowpass lines prior to further processing. Default is enabled.
16282 @subsection Commands
16284 This filter supports the all above options as @ref{commands}.
16286 @subsection Examples
16290 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16292 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
16298 Read vertical interval timecode (VITC) information from the top lines of a
16301 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16302 timecode value, if a valid timecode has been detected. Further metadata key
16303 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16304 timecode data has been found or not.
16306 This filter accepts the following options:
16310 Set the maximum number of lines to scan for VITC data. If the value is set to
16311 @code{-1} the full video frame is scanned. Default is @code{45}.
16314 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16315 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16318 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16319 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16322 @subsection Examples
16326 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16327 draw @code{--:--:--:--} as a placeholder:
16329 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16335 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16337 Destination pixel at position (X, Y) will be picked from source (x, y) position
16338 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16339 value for pixel will be used for destination pixel.
16341 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16342 will have Xmap/Ymap video stream dimensions.
16343 Xmap and Ymap input video streams are 16bit depth, single channel.
16347 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16348 Default is @code{color}.
16351 Specify the color of the unmapped pixels. For the syntax of this option,
16352 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16353 manual,ffmpeg-utils}. Default color is @code{black}.
16356 @section removegrain
16358 The removegrain filter is a spatial denoiser for progressive video.
16362 Set mode for the first plane.
16365 Set mode for the second plane.
16368 Set mode for the third plane.
16371 Set mode for the fourth plane.
16374 Range of mode is from 0 to 24. Description of each mode follows:
16378 Leave input plane unchanged. Default.
16381 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16384 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16387 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16390 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16391 This is equivalent to a median filter.
16394 Line-sensitive clipping giving the minimal change.
16397 Line-sensitive clipping, intermediate.
16400 Line-sensitive clipping, intermediate.
16403 Line-sensitive clipping, intermediate.
16406 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16409 Replaces the target pixel with the closest neighbour.
16412 [1 2 1] horizontal and vertical kernel blur.
16418 Bob mode, interpolates top field from the line where the neighbours
16419 pixels are the closest.
16422 Bob mode, interpolates bottom field from the line where the neighbours
16423 pixels are the closest.
16426 Bob mode, interpolates top field. Same as 13 but with a more complicated
16427 interpolation formula.
16430 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16431 interpolation formula.
16434 Clips the pixel with the minimum and maximum of respectively the maximum and
16435 minimum of each pair of opposite neighbour pixels.
16438 Line-sensitive clipping using opposite neighbours whose greatest distance from
16439 the current pixel is minimal.
16442 Replaces the pixel with the average of its 8 neighbours.
16445 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16448 Clips pixels using the averages of opposite neighbour.
16451 Same as mode 21 but simpler and faster.
16454 Small edge and halo removal, but reputed useless.
16460 @section removelogo
16462 Suppress a TV station logo, using an image file to determine which
16463 pixels comprise the logo. It works by filling in the pixels that
16464 comprise the logo with neighboring pixels.
16466 The filter accepts the following options:
16470 Set the filter bitmap file, which can be any image format supported by
16471 libavformat. The width and height of the image file must match those of the
16472 video stream being processed.
16475 Pixels in the provided bitmap image with a value of zero are not
16476 considered part of the logo, non-zero pixels are considered part of
16477 the logo. If you use white (255) for the logo and black (0) for the
16478 rest, you will be safe. For making the filter bitmap, it is
16479 recommended to take a screen capture of a black frame with the logo
16480 visible, and then using a threshold filter followed by the erode
16481 filter once or twice.
16483 If needed, little splotches can be fixed manually. Remember that if
16484 logo pixels are not covered, the filter quality will be much
16485 reduced. Marking too many pixels as part of the logo does not hurt as
16486 much, but it will increase the amount of blurring needed to cover over
16487 the image and will destroy more information than necessary, and extra
16488 pixels will slow things down on a large logo.
16490 @section repeatfields
16492 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16493 fields based on its value.
16497 Reverse a video clip.
16499 Warning: This filter requires memory to buffer the entire clip, so trimming
16502 @subsection Examples
16506 Take the first 5 seconds of a clip, and reverse it.
16513 Shift R/G/B/A pixels horizontally and/or vertically.
16515 The filter accepts the following options:
16518 Set amount to shift red horizontally.
16520 Set amount to shift red vertically.
16522 Set amount to shift green horizontally.
16524 Set amount to shift green vertically.
16526 Set amount to shift blue horizontally.
16528 Set amount to shift blue vertically.
16530 Set amount to shift alpha horizontally.
16532 Set amount to shift alpha vertically.
16534 Set edge mode, can be @var{smear}, default, or @var{warp}.
16537 @subsection Commands
16539 This filter supports the all above options as @ref{commands}.
16542 Apply roberts cross operator to input video stream.
16544 The filter accepts the following option:
16548 Set which planes will be processed, unprocessed planes will be copied.
16549 By default value 0xf, all planes will be processed.
16552 Set value which will be multiplied with filtered result.
16555 Set value which will be added to filtered result.
16558 @subsection Commands
16560 This filter supports the all above options as @ref{commands}.
16564 Rotate video by an arbitrary angle expressed in radians.
16566 The filter accepts the following options:
16568 A description of the optional parameters follows.
16571 Set an expression for the angle by which to rotate the input video
16572 clockwise, expressed as a number of radians. A negative value will
16573 result in a counter-clockwise rotation. By default it is set to "0".
16575 This expression is evaluated for each frame.
16578 Set the output width expression, default value is "iw".
16579 This expression is evaluated just once during configuration.
16582 Set the output height expression, default value is "ih".
16583 This expression is evaluated just once during configuration.
16586 Enable bilinear interpolation if set to 1, a value of 0 disables
16587 it. Default value is 1.
16590 Set the color used to fill the output area not covered by the rotated
16591 image. For the general syntax of this option, check the
16592 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16593 If the special value "none" is selected then no
16594 background is printed (useful for example if the background is never shown).
16596 Default value is "black".
16599 The expressions for the angle and the output size can contain the
16600 following constants and functions:
16604 sequential number of the input frame, starting from 0. It is always NAN
16605 before the first frame is filtered.
16608 time in seconds of the input frame, it is set to 0 when the filter is
16609 configured. It is always NAN before the first frame is filtered.
16613 horizontal and vertical chroma subsample values. For example for the
16614 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16618 the input video width and height
16622 the output width and height, that is the size of the padded area as
16623 specified by the @var{width} and @var{height} expressions
16627 the minimal width/height required for completely containing the input
16628 video rotated by @var{a} radians.
16630 These are only available when computing the @option{out_w} and
16631 @option{out_h} expressions.
16634 @subsection Examples
16638 Rotate the input by PI/6 radians clockwise:
16644 Rotate the input by PI/6 radians counter-clockwise:
16650 Rotate the input by 45 degrees clockwise:
16656 Apply a constant rotation with period T, starting from an angle of PI/3:
16658 rotate=PI/3+2*PI*t/T
16662 Make the input video rotation oscillating with a period of T
16663 seconds and an amplitude of A radians:
16665 rotate=A*sin(2*PI/T*t)
16669 Rotate the video, output size is chosen so that the whole rotating
16670 input video is always completely contained in the output:
16672 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16676 Rotate the video, reduce the output size so that no background is ever
16679 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16683 @subsection Commands
16685 The filter supports the following commands:
16689 Set the angle expression.
16690 The command accepts the same syntax of the corresponding option.
16692 If the specified expression is not valid, it is kept at its current
16698 Apply Shape Adaptive Blur.
16700 The filter accepts the following options:
16703 @item luma_radius, lr
16704 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16705 value is 1.0. A greater value will result in a more blurred image, and
16706 in slower processing.
16708 @item luma_pre_filter_radius, lpfr
16709 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16712 @item luma_strength, ls
16713 Set luma maximum difference between pixels to still be considered, must
16714 be a value in the 0.1-100.0 range, default value is 1.0.
16716 @item chroma_radius, cr
16717 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16718 greater value will result in a more blurred image, and in slower
16721 @item chroma_pre_filter_radius, cpfr
16722 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16724 @item chroma_strength, cs
16725 Set chroma maximum difference between pixels to still be considered,
16726 must be a value in the -0.9-100.0 range.
16729 Each chroma option value, if not explicitly specified, is set to the
16730 corresponding luma option value.
16735 Scale (resize) the input video, using the libswscale library.
16737 The scale filter forces the output display aspect ratio to be the same
16738 of the input, by changing the output sample aspect ratio.
16740 If the input image format is different from the format requested by
16741 the next filter, the scale filter will convert the input to the
16744 @subsection Options
16745 The filter accepts the following options, or any of the options
16746 supported by the libswscale scaler.
16748 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16749 the complete list of scaler options.
16754 Set the output video dimension expression. Default value is the input
16757 If the @var{width} or @var{w} value is 0, the input width is used for
16758 the output. If the @var{height} or @var{h} value is 0, the input height
16759 is used for the output.
16761 If one and only one of the values is -n with n >= 1, the scale filter
16762 will use a value that maintains the aspect ratio of the input image,
16763 calculated from the other specified dimension. After that it will,
16764 however, make sure that the calculated dimension is divisible by n and
16765 adjust the value if necessary.
16767 If both values are -n with n >= 1, the behavior will be identical to
16768 both values being set to 0 as previously detailed.
16770 See below for the list of accepted constants for use in the dimension
16774 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16778 Only evaluate expressions once during the filter initialization or when a command is processed.
16781 Evaluate expressions for each incoming frame.
16785 Default value is @samp{init}.
16789 Set the interlacing mode. It accepts the following values:
16793 Force interlaced aware scaling.
16796 Do not apply interlaced scaling.
16799 Select interlaced aware scaling depending on whether the source frames
16800 are flagged as interlaced or not.
16803 Default value is @samp{0}.
16806 Set libswscale scaling flags. See
16807 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16808 complete list of values. If not explicitly specified the filter applies
16812 @item param0, param1
16813 Set libswscale input parameters for scaling algorithms that need them. See
16814 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16815 complete documentation. If not explicitly specified the filter applies
16821 Set the video size. For the syntax of this option, check the
16822 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16824 @item in_color_matrix
16825 @item out_color_matrix
16826 Set in/output YCbCr color space type.
16828 This allows the autodetected value to be overridden as well as allows forcing
16829 a specific value used for the output and encoder.
16831 If not specified, the color space type depends on the pixel format.
16837 Choose automatically.
16840 Format conforming to International Telecommunication Union (ITU)
16841 Recommendation BT.709.
16844 Set color space conforming to the United States Federal Communications
16845 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16850 Set color space conforming to:
16854 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16857 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16860 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16865 Set color space conforming to SMPTE ST 240:1999.
16868 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16873 Set in/output YCbCr sample range.
16875 This allows the autodetected value to be overridden as well as allows forcing
16876 a specific value used for the output and encoder. If not specified, the
16877 range depends on the pixel format. Possible values:
16881 Choose automatically.
16884 Set full range (0-255 in case of 8-bit luma).
16886 @item mpeg/limited/tv
16887 Set "MPEG" range (16-235 in case of 8-bit luma).
16890 @item force_original_aspect_ratio
16891 Enable decreasing or increasing output video width or height if necessary to
16892 keep the original aspect ratio. Possible values:
16896 Scale the video as specified and disable this feature.
16899 The output video dimensions will automatically be decreased if needed.
16902 The output video dimensions will automatically be increased if needed.
16906 One useful instance of this option is that when you know a specific device's
16907 maximum allowed resolution, you can use this to limit the output video to
16908 that, while retaining the aspect ratio. For example, device A allows
16909 1280x720 playback, and your video is 1920x800. Using this option (set it to
16910 decrease) and specifying 1280x720 to the command line makes the output
16913 Please note that this is a different thing than specifying -1 for @option{w}
16914 or @option{h}, you still need to specify the output resolution for this option
16917 @item force_divisible_by
16918 Ensures that both the output dimensions, width and height, are divisible by the
16919 given integer when used together with @option{force_original_aspect_ratio}. This
16920 works similar to using @code{-n} in the @option{w} and @option{h} options.
16922 This option respects the value set for @option{force_original_aspect_ratio},
16923 increasing or decreasing the resolution accordingly. The video's aspect ratio
16924 may be slightly modified.
16926 This option can be handy if you need to have a video fit within or exceed
16927 a defined resolution using @option{force_original_aspect_ratio} but also have
16928 encoder restrictions on width or height divisibility.
16932 The values of the @option{w} and @option{h} options are expressions
16933 containing the following constants:
16938 The input width and height
16942 These are the same as @var{in_w} and @var{in_h}.
16946 The output (scaled) width and height
16950 These are the same as @var{out_w} and @var{out_h}
16953 The same as @var{iw} / @var{ih}
16956 input sample aspect ratio
16959 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16963 horizontal and vertical input chroma subsample values. For example for the
16964 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16968 horizontal and vertical output chroma subsample values. For example for the
16969 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16972 The (sequential) number of the input frame, starting from 0.
16973 Only available with @code{eval=frame}.
16976 The presentation timestamp of the input frame, expressed as a number of
16977 seconds. Only available with @code{eval=frame}.
16980 The position (byte offset) of the frame in the input stream, or NaN if
16981 this information is unavailable and/or meaningless (for example in case of synthetic video).
16982 Only available with @code{eval=frame}.
16985 @subsection Examples
16989 Scale the input video to a size of 200x100
16994 This is equivalent to:
17005 Specify a size abbreviation for the output size:
17010 which can also be written as:
17016 Scale the input to 2x:
17018 scale=w=2*iw:h=2*ih
17022 The above is the same as:
17024 scale=2*in_w:2*in_h
17028 Scale the input to 2x with forced interlaced scaling:
17030 scale=2*iw:2*ih:interl=1
17034 Scale the input to half size:
17036 scale=w=iw/2:h=ih/2
17040 Increase the width, and set the height to the same size:
17046 Seek Greek harmony:
17053 Increase the height, and set the width to 3/2 of the height:
17055 scale=w=3/2*oh:h=3/5*ih
17059 Increase the size, making the size a multiple of the chroma
17062 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17066 Increase the width to a maximum of 500 pixels,
17067 keeping the same aspect ratio as the input:
17069 scale=w='min(500\, iw*3/2):h=-1'
17073 Make pixels square by combining scale and setsar:
17075 scale='trunc(ih*dar):ih',setsar=1/1
17079 Make pixels square by combining scale and setsar,
17080 making sure the resulting resolution is even (required by some codecs):
17082 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17086 @subsection Commands
17088 This filter supports the following commands:
17092 Set the output video dimension expression.
17093 The command accepts the same syntax of the corresponding option.
17095 If the specified expression is not valid, it is kept at its current
17101 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17102 format conversion on CUDA video frames. Setting the output width and height
17103 works in the same way as for the @var{scale} filter.
17105 The following additional options are accepted:
17108 The pixel format of the output CUDA frames. If set to the string "same" (the
17109 default), the input format will be kept. Note that automatic format negotiation
17110 and conversion is not yet supported for hardware frames
17113 The interpolation algorithm used for resizing. One of the following:
17120 @item cubic2p_bspline
17121 2-parameter cubic (B=1, C=0)
17123 @item cubic2p_catmullrom
17124 2-parameter cubic (B=0, C=1/2)
17126 @item cubic2p_b05c03
17127 2-parameter cubic (B=1/2, C=3/10)
17135 @item force_original_aspect_ratio
17136 Enable decreasing or increasing output video width or height if necessary to
17137 keep the original aspect ratio. Possible values:
17141 Scale the video as specified and disable this feature.
17144 The output video dimensions will automatically be decreased if needed.
17147 The output video dimensions will automatically be increased if needed.
17151 One useful instance of this option is that when you know a specific device's
17152 maximum allowed resolution, you can use this to limit the output video to
17153 that, while retaining the aspect ratio. For example, device A allows
17154 1280x720 playback, and your video is 1920x800. Using this option (set it to
17155 decrease) and specifying 1280x720 to the command line makes the output
17158 Please note that this is a different thing than specifying -1 for @option{w}
17159 or @option{h}, you still need to specify the output resolution for this option
17162 @item force_divisible_by
17163 Ensures that both the output dimensions, width and height, are divisible by the
17164 given integer when used together with @option{force_original_aspect_ratio}. This
17165 works similar to using @code{-n} in the @option{w} and @option{h} options.
17167 This option respects the value set for @option{force_original_aspect_ratio},
17168 increasing or decreasing the resolution accordingly. The video's aspect ratio
17169 may be slightly modified.
17171 This option can be handy if you need to have a video fit within or exceed
17172 a defined resolution using @option{force_original_aspect_ratio} but also have
17173 encoder restrictions on width or height divisibility.
17179 Scale (resize) the input video, based on a reference video.
17181 See the scale filter for available options, scale2ref supports the same but
17182 uses the reference video instead of the main input as basis. scale2ref also
17183 supports the following additional constants for the @option{w} and
17184 @option{h} options:
17189 The main input video's width and height
17192 The same as @var{main_w} / @var{main_h}
17195 The main input video's sample aspect ratio
17197 @item main_dar, mdar
17198 The main input video's display aspect ratio. Calculated from
17199 @code{(main_w / main_h) * main_sar}.
17203 The main input video's horizontal and vertical chroma subsample values.
17204 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17208 The (sequential) number of the main input frame, starting from 0.
17209 Only available with @code{eval=frame}.
17212 The presentation timestamp of the main input frame, expressed as a number of
17213 seconds. Only available with @code{eval=frame}.
17216 The position (byte offset) of the frame in the main input stream, or NaN if
17217 this information is unavailable and/or meaningless (for example in case of synthetic video).
17218 Only available with @code{eval=frame}.
17221 @subsection Examples
17225 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17227 'scale2ref[b][a];[a][b]overlay'
17231 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17233 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17237 @subsection Commands
17239 This filter supports the following commands:
17243 Set the output video dimension expression.
17244 The command accepts the same syntax of the corresponding option.
17246 If the specified expression is not valid, it is kept at its current
17251 Scroll input video horizontally and/or vertically by constant speed.
17253 The filter accepts the following options:
17255 @item horizontal, h
17256 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17257 Negative values changes scrolling direction.
17260 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17261 Negative values changes scrolling direction.
17264 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17267 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17270 @subsection Commands
17272 This filter supports the following @ref{commands}:
17274 @item horizontal, h
17275 Set the horizontal scrolling speed.
17277 Set the vertical scrolling speed.
17283 Detect video scene change.
17285 This filter sets frame metadata with mafd between frame, the scene score, and
17286 forward the frame to the next filter, so they can use these metadata to detect
17287 scene change or others.
17289 In addition, this filter logs a message and sets frame metadata when it detects
17290 a scene change by @option{threshold}.
17292 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17294 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17295 to detect scene change.
17297 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17298 detect scene change with @option{threshold}.
17300 The filter accepts the following options:
17304 Set the scene change detection threshold as a percentage of maximum change. Good
17305 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17308 Default value is @code{10.}.
17311 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17312 You can enable it if you want to get snapshot of scene change frames only.
17315 @anchor{selectivecolor}
17316 @section selectivecolor
17318 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17319 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17320 by the "purity" of the color (that is, how saturated it already is).
17322 This filter is similar to the Adobe Photoshop Selective Color tool.
17324 The filter accepts the following options:
17327 @item correction_method
17328 Select color correction method.
17330 Available values are:
17333 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17336 Specified adjustments are relative to the original component value.
17338 Default is @code{absolute}.
17340 Adjustments for red pixels (pixels where the red component is the maximum)
17342 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17344 Adjustments for green pixels (pixels where the green component is the maximum)
17346 Adjustments for cyan pixels (pixels where the red component is the minimum)
17348 Adjustments for blue pixels (pixels where the blue component is the maximum)
17350 Adjustments for magenta pixels (pixels where the green component is the minimum)
17352 Adjustments for white pixels (pixels where all components are greater than 128)
17354 Adjustments for all pixels except pure black and pure white
17356 Adjustments for black pixels (pixels where all components are lesser than 128)
17358 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17361 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17362 4 space separated floating point adjustment values in the [-1,1] range,
17363 respectively to adjust the amount of cyan, magenta, yellow and black for the
17364 pixels of its range.
17366 @subsection Examples
17370 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17371 increase magenta by 27% in blue areas:
17373 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17377 Use a Photoshop selective color preset:
17379 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17383 @anchor{separatefields}
17384 @section separatefields
17386 The @code{separatefields} takes a frame-based video input and splits
17387 each frame into its components fields, producing a new half height clip
17388 with twice the frame rate and twice the frame count.
17390 This filter use field-dominance information in frame to decide which
17391 of each pair of fields to place first in the output.
17392 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17394 @section setdar, setsar
17396 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17399 This is done by changing the specified Sample (aka Pixel) Aspect
17400 Ratio, according to the following equation:
17402 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17405 Keep in mind that the @code{setdar} filter does not modify the pixel
17406 dimensions of the video frame. Also, the display aspect ratio set by
17407 this filter may be changed by later filters in the filterchain,
17408 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17411 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17412 the filter output video.
17414 Note that as a consequence of the application of this filter, the
17415 output display aspect ratio will change according to the equation
17418 Keep in mind that the sample aspect ratio set by the @code{setsar}
17419 filter may be changed by later filters in the filterchain, e.g. if
17420 another "setsar" or a "setdar" filter is applied.
17422 It accepts the following parameters:
17425 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17426 Set the aspect ratio used by the filter.
17428 The parameter can be a floating point number string, an expression, or
17429 a string of the form @var{num}:@var{den}, where @var{num} and
17430 @var{den} are the numerator and denominator of the aspect ratio. If
17431 the parameter is not specified, it is assumed the value "0".
17432 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17436 Set the maximum integer value to use for expressing numerator and
17437 denominator when reducing the expressed aspect ratio to a rational.
17438 Default value is @code{100}.
17442 The parameter @var{sar} is an expression containing
17443 the following constants:
17447 These are approximated values for the mathematical constants e
17448 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17451 The input width and height.
17454 These are the same as @var{w} / @var{h}.
17457 The input sample aspect ratio.
17460 The input display aspect ratio. It is the same as
17461 (@var{w} / @var{h}) * @var{sar}.
17464 Horizontal and vertical chroma subsample values. For example, for the
17465 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17468 @subsection Examples
17473 To change the display aspect ratio to 16:9, specify one of the following:
17480 To change the sample aspect ratio to 10:11, specify:
17486 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17487 1000 in the aspect ratio reduction, use the command:
17489 setdar=ratio=16/9:max=1000
17497 Force field for the output video frame.
17499 The @code{setfield} filter marks the interlace type field for the
17500 output frames. It does not change the input frame, but only sets the
17501 corresponding property, which affects how the frame is treated by
17502 following filters (e.g. @code{fieldorder} or @code{yadif}).
17504 The filter accepts the following options:
17509 Available values are:
17513 Keep the same field property.
17516 Mark the frame as bottom-field-first.
17519 Mark the frame as top-field-first.
17522 Mark the frame as progressive.
17529 Force frame parameter for the output video frame.
17531 The @code{setparams} filter marks interlace and color range for the
17532 output frames. It does not change the input frame, but only sets the
17533 corresponding property, which affects how the frame is treated by
17538 Available values are:
17542 Keep the same field property (default).
17545 Mark the frame as bottom-field-first.
17548 Mark the frame as top-field-first.
17551 Mark the frame as progressive.
17555 Available values are:
17559 Keep the same color range property (default).
17561 @item unspecified, unknown
17562 Mark the frame as unspecified color range.
17564 @item limited, tv, mpeg
17565 Mark the frame as limited range.
17567 @item full, pc, jpeg
17568 Mark the frame as full range.
17571 @item color_primaries
17572 Set the color primaries.
17573 Available values are:
17577 Keep the same color primaries property (default).
17594 Set the color transfer.
17595 Available values are:
17599 Keep the same color trc property (default).
17621 Set the colorspace.
17622 Available values are:
17626 Keep the same colorspace property (default).
17639 @item chroma-derived-nc
17640 @item chroma-derived-c
17647 Show a line containing various information for each input video frame.
17648 The input video is not modified.
17650 This filter supports the following options:
17654 Calculate checksums of each plane. By default enabled.
17657 The shown line contains a sequence of key/value pairs of the form
17658 @var{key}:@var{value}.
17660 The following values are shown in the output:
17664 The (sequential) number of the input frame, starting from 0.
17667 The Presentation TimeStamp of the input frame, expressed as a number of
17668 time base units. The time base unit depends on the filter input pad.
17671 The Presentation TimeStamp of the input frame, expressed as a number of
17675 The position of the frame in the input stream, or -1 if this information is
17676 unavailable and/or meaningless (for example in case of synthetic video).
17679 The pixel format name.
17682 The sample aspect ratio of the input frame, expressed in the form
17683 @var{num}/@var{den}.
17686 The size of the input frame. For the syntax of this option, check the
17687 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17690 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17691 for bottom field first).
17694 This is 1 if the frame is a key frame, 0 otherwise.
17697 The picture type of the input frame ("I" for an I-frame, "P" for a
17698 P-frame, "B" for a B-frame, or "?" for an unknown type).
17699 Also refer to the documentation of the @code{AVPictureType} enum and of
17700 the @code{av_get_picture_type_char} function defined in
17701 @file{libavutil/avutil.h}.
17704 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17706 @item plane_checksum
17707 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17708 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17711 The mean value of pixels in each plane of the input frame, expressed in the form
17712 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17715 The standard deviation of pixel values in each plane of the input frame, expressed
17716 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17720 @section showpalette
17722 Displays the 256 colors palette of each frame. This filter is only relevant for
17723 @var{pal8} pixel format frames.
17725 It accepts the following option:
17729 Set the size of the box used to represent one palette color entry. Default is
17730 @code{30} (for a @code{30x30} pixel box).
17733 @section shuffleframes
17735 Reorder and/or duplicate and/or drop video frames.
17737 It accepts the following parameters:
17741 Set the destination indexes of input frames.
17742 This is space or '|' separated list of indexes that maps input frames to output
17743 frames. Number of indexes also sets maximal value that each index may have.
17744 '-1' index have special meaning and that is to drop frame.
17747 The first frame has the index 0. The default is to keep the input unchanged.
17749 @subsection Examples
17753 Swap second and third frame of every three frames of the input:
17755 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17759 Swap 10th and 1st frame of every ten frames of the input:
17761 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17765 @section shuffleplanes
17767 Reorder and/or duplicate video planes.
17769 It accepts the following parameters:
17774 The index of the input plane to be used as the first output plane.
17777 The index of the input plane to be used as the second output plane.
17780 The index of the input plane to be used as the third output plane.
17783 The index of the input plane to be used as the fourth output plane.
17787 The first plane has the index 0. The default is to keep the input unchanged.
17789 @subsection Examples
17793 Swap the second and third planes of the input:
17795 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17799 @anchor{signalstats}
17800 @section signalstats
17801 Evaluate various visual metrics that assist in determining issues associated
17802 with the digitization of analog video media.
17804 By default the filter will log these metadata values:
17808 Display the minimal Y value contained within the input frame. Expressed in
17812 Display the Y value at the 10% percentile within the input frame. Expressed in
17816 Display the average Y value within the input frame. Expressed in range of
17820 Display the Y value at the 90% percentile within the input frame. Expressed in
17824 Display the maximum Y value contained within the input frame. Expressed in
17828 Display the minimal U value contained within the input frame. Expressed in
17832 Display the U value at the 10% percentile within the input frame. Expressed in
17836 Display the average U value within the input frame. Expressed in range of
17840 Display the U value at the 90% percentile within the input frame. Expressed in
17844 Display the maximum U value contained within the input frame. Expressed in
17848 Display the minimal V value contained within the input frame. Expressed in
17852 Display the V value at the 10% percentile within the input frame. Expressed in
17856 Display the average V value within the input frame. Expressed in range of
17860 Display the V value at the 90% percentile within the input frame. Expressed in
17864 Display the maximum V value contained within the input frame. Expressed in
17868 Display the minimal saturation value contained within the input frame.
17869 Expressed in range of [0-~181.02].
17872 Display the saturation value at the 10% percentile within the input frame.
17873 Expressed in range of [0-~181.02].
17876 Display the average saturation value within the input frame. Expressed in range
17880 Display the saturation value at the 90% percentile within the input frame.
17881 Expressed in range of [0-~181.02].
17884 Display the maximum saturation value contained within the input frame.
17885 Expressed in range of [0-~181.02].
17888 Display the median value for hue within the input frame. Expressed in range of
17892 Display the average value for hue within the input frame. Expressed in range of
17896 Display the average of sample value difference between all values of the Y
17897 plane in the current frame and corresponding values of the previous input frame.
17898 Expressed in range of [0-255].
17901 Display the average of sample value difference between all values of the U
17902 plane in the current frame and corresponding values of the previous input frame.
17903 Expressed in range of [0-255].
17906 Display the average of sample value difference between all values of the V
17907 plane in the current frame and corresponding values of the previous input frame.
17908 Expressed in range of [0-255].
17911 Display bit depth of Y plane in current frame.
17912 Expressed in range of [0-16].
17915 Display bit depth of U plane in current frame.
17916 Expressed in range of [0-16].
17919 Display bit depth of V plane in current frame.
17920 Expressed in range of [0-16].
17923 The filter accepts the following options:
17929 @option{stat} specify an additional form of image analysis.
17930 @option{out} output video with the specified type of pixel highlighted.
17932 Both options accept the following values:
17936 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17937 unlike the neighboring pixels of the same field. Examples of temporal outliers
17938 include the results of video dropouts, head clogs, or tape tracking issues.
17941 Identify @var{vertical line repetition}. Vertical line repetition includes
17942 similar rows of pixels within a frame. In born-digital video vertical line
17943 repetition is common, but this pattern is uncommon in video digitized from an
17944 analog source. When it occurs in video that results from the digitization of an
17945 analog source it can indicate concealment from a dropout compensator.
17948 Identify pixels that fall outside of legal broadcast range.
17952 Set the highlight color for the @option{out} option. The default color is
17956 @subsection Examples
17960 Output data of various video metrics:
17962 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17966 Output specific data about the minimum and maximum values of the Y plane per frame:
17968 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17972 Playback video while highlighting pixels that are outside of broadcast range in red.
17974 ffplay example.mov -vf signalstats="out=brng:color=red"
17978 Playback video with signalstats metadata drawn over the frame.
17980 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17983 The contents of signalstat_drawtext.txt used in the command are:
17986 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17987 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17988 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17989 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17997 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17998 input. In this case the matching between the inputs can be calculated additionally.
17999 The filter always passes through the first input. The signature of each stream can
18000 be written into a file.
18002 It accepts the following options:
18006 Enable or disable the matching process.
18008 Available values are:
18012 Disable the calculation of a matching (default).
18014 Calculate the matching for the whole video and output whether the whole video
18015 matches or only parts.
18017 Calculate only until a matching is found or the video ends. Should be faster in
18022 Set the number of inputs. The option value must be a non negative integer.
18023 Default value is 1.
18026 Set the path to which the output is written. If there is more than one input,
18027 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18028 integer), that will be replaced with the input number. If no filename is
18029 specified, no output will be written. This is the default.
18032 Choose the output format.
18034 Available values are:
18038 Use the specified binary representation (default).
18040 Use the specified xml representation.
18044 Set threshold to detect one word as similar. The option value must be an integer
18045 greater than zero. The default value is 9000.
18048 Set threshold to detect all words as similar. The option value must be an integer
18049 greater than zero. The default value is 60000.
18052 Set threshold to detect frames as similar. The option value must be an integer
18053 greater than zero. The default value is 116.
18056 Set the minimum length of a sequence in frames to recognize it as matching
18057 sequence. The option value must be a non negative integer value.
18058 The default value is 0.
18061 Set the minimum relation, that matching frames to all frames must have.
18062 The option value must be a double value between 0 and 1. The default value is 0.5.
18065 @subsection Examples
18069 To calculate the signature of an input video and store it in signature.bin:
18071 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18075 To detect whether two videos match and store the signatures in XML format in
18076 signature0.xml and signature1.xml:
18078 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 -
18086 Blur the input video without impacting the outlines.
18088 It accepts the following options:
18091 @item luma_radius, lr
18092 Set the luma radius. The option value must be a float number in
18093 the range [0.1,5.0] that specifies the variance of the gaussian filter
18094 used to blur the image (slower if larger). Default value is 1.0.
18096 @item luma_strength, ls
18097 Set the luma strength. The option value must be a float number
18098 in the range [-1.0,1.0] that configures the blurring. A value included
18099 in [0.0,1.0] will blur the image whereas a value included in
18100 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18102 @item luma_threshold, lt
18103 Set the luma threshold used as a coefficient to determine
18104 whether a pixel should be blurred or not. The option value must be an
18105 integer in the range [-30,30]. A value of 0 will filter all the image,
18106 a value included in [0,30] will filter flat areas and a value included
18107 in [-30,0] will filter edges. Default value is 0.
18109 @item chroma_radius, cr
18110 Set the chroma radius. The option value must be a float number in
18111 the range [0.1,5.0] that specifies the variance of the gaussian filter
18112 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18114 @item chroma_strength, cs
18115 Set the chroma strength. The option value must be a float number
18116 in the range [-1.0,1.0] that configures the blurring. A value included
18117 in [0.0,1.0] will blur the image whereas a value included in
18118 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18120 @item chroma_threshold, ct
18121 Set the chroma threshold used as a coefficient to determine
18122 whether a pixel should be blurred or not. The option value must be an
18123 integer in the range [-30,30]. A value of 0 will filter all the image,
18124 a value included in [0,30] will filter flat areas and a value included
18125 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18128 If a chroma option is not explicitly set, the corresponding luma value
18132 Apply sobel operator to input video stream.
18134 The filter accepts the following option:
18138 Set which planes will be processed, unprocessed planes will be copied.
18139 By default value 0xf, all planes will be processed.
18142 Set value which will be multiplied with filtered result.
18145 Set value which will be added to filtered result.
18148 @subsection Commands
18150 This filter supports the all above options as @ref{commands}.
18155 Apply a simple postprocessing filter that compresses and decompresses the image
18156 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18157 and average the results.
18159 The filter accepts the following options:
18163 Set quality. This option defines the number of levels for averaging. It accepts
18164 an integer in the range 0-6. If set to @code{0}, the filter will have no
18165 effect. A value of @code{6} means the higher quality. For each increment of
18166 that value the speed drops by a factor of approximately 2. Default value is
18170 Force a constant quantization parameter. If not set, the filter will use the QP
18171 from the video stream (if available).
18174 Set thresholding mode. Available modes are:
18178 Set hard thresholding (default).
18180 Set soft thresholding (better de-ringing effect, but likely blurrier).
18183 @item use_bframe_qp
18184 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18185 option may cause flicker since the B-Frames have often larger QP. Default is
18186 @code{0} (not enabled).
18189 @subsection Commands
18191 This filter supports the following commands:
18193 @item quality, level
18194 Set quality level. The value @code{max} can be used to set the maximum level,
18195 currently @code{6}.
18201 Scale the input by applying one of the super-resolution methods based on
18202 convolutional neural networks. Supported models:
18206 Super-Resolution Convolutional Neural Network model (SRCNN).
18207 See @url{https://arxiv.org/abs/1501.00092}.
18210 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18211 See @url{https://arxiv.org/abs/1609.05158}.
18214 Training scripts as well as scripts for model file (.pb) saving can be found at
18215 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18216 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18218 Native model files (.model) can be generated from TensorFlow model
18219 files (.pb) by using tools/python/convert.py
18221 The filter accepts the following options:
18225 Specify which DNN backend to use for model loading and execution. This option accepts
18226 the following values:
18230 Native implementation of DNN loading and execution.
18233 TensorFlow backend. To enable this backend you
18234 need to install the TensorFlow for C library (see
18235 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18236 @code{--enable-libtensorflow}
18239 Default value is @samp{native}.
18242 Set path to model file specifying network architecture and its parameters.
18243 Note that different backends use different file formats. TensorFlow backend
18244 can load files for both formats, while native backend can load files for only
18248 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18249 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18250 input upscaled using bicubic upscaling with proper scale factor.
18253 This feature can also be finished with @ref{dnn_processing} filter.
18257 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18259 This filter takes in input two input videos, the first input is
18260 considered the "main" source and is passed unchanged to the
18261 output. The second input is used as a "reference" video for computing
18264 Both video inputs must have the same resolution and pixel format for
18265 this filter to work correctly. Also it assumes that both inputs
18266 have the same number of frames, which are compared one by one.
18268 The filter stores the calculated SSIM of each frame.
18270 The description of the accepted parameters follows.
18273 @item stats_file, f
18274 If specified the filter will use the named file to save the SSIM of
18275 each individual frame. When filename equals "-" the data is sent to
18279 The file printed if @var{stats_file} is selected, contains a sequence of
18280 key/value pairs of the form @var{key}:@var{value} for each compared
18283 A description of each shown parameter follows:
18287 sequential number of the input frame, starting from 1
18289 @item Y, U, V, R, G, B
18290 SSIM of the compared frames for the component specified by the suffix.
18293 SSIM of the compared frames for the whole frame.
18296 Same as above but in dB representation.
18299 This filter also supports the @ref{framesync} options.
18301 @subsection Examples
18306 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18307 [main][ref] ssim="stats_file=stats.log" [out]
18310 On this example the input file being processed is compared with the
18311 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18312 is stored in @file{stats.log}.
18315 Another example with both psnr and ssim at same time:
18317 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18321 Another example with different containers:
18323 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 -
18329 Convert between different stereoscopic image formats.
18331 The filters accept the following options:
18335 Set stereoscopic image format of input.
18337 Available values for input image formats are:
18340 side by side parallel (left eye left, right eye right)
18343 side by side crosseye (right eye left, left eye right)
18346 side by side parallel with half width resolution
18347 (left eye left, right eye right)
18350 side by side crosseye with half width resolution
18351 (right eye left, left eye right)
18355 above-below (left eye above, right eye below)
18359 above-below (right eye above, left eye below)
18363 above-below with half height resolution
18364 (left eye above, right eye below)
18368 above-below with half height resolution
18369 (right eye above, left eye below)
18372 alternating frames (left eye first, right eye second)
18375 alternating frames (right eye first, left eye second)
18378 interleaved rows (left eye has top row, right eye starts on next row)
18381 interleaved rows (right eye has top row, left eye starts on next row)
18384 interleaved columns, left eye first
18387 interleaved columns, right eye first
18389 Default value is @samp{sbsl}.
18393 Set stereoscopic image format of output.
18397 side by side parallel (left eye left, right eye right)
18400 side by side crosseye (right eye left, left eye right)
18403 side by side parallel with half width resolution
18404 (left eye left, right eye right)
18407 side by side crosseye with half width resolution
18408 (right eye left, left eye right)
18412 above-below (left eye above, right eye below)
18416 above-below (right eye above, left eye below)
18420 above-below with half height resolution
18421 (left eye above, right eye below)
18425 above-below with half height resolution
18426 (right eye above, left eye below)
18429 alternating frames (left eye first, right eye second)
18432 alternating frames (right eye first, left eye second)
18435 interleaved rows (left eye has top row, right eye starts on next row)
18438 interleaved rows (right eye has top row, left eye starts on next row)
18441 anaglyph red/blue gray
18442 (red filter on left eye, blue filter on right eye)
18445 anaglyph red/green gray
18446 (red filter on left eye, green filter on right eye)
18449 anaglyph red/cyan gray
18450 (red filter on left eye, cyan filter on right eye)
18453 anaglyph red/cyan half colored
18454 (red filter on left eye, cyan filter on right eye)
18457 anaglyph red/cyan color
18458 (red filter on left eye, cyan filter on right eye)
18461 anaglyph red/cyan color optimized with the least squares projection of dubois
18462 (red filter on left eye, cyan filter on right eye)
18465 anaglyph green/magenta gray
18466 (green filter on left eye, magenta filter on right eye)
18469 anaglyph green/magenta half colored
18470 (green filter on left eye, magenta filter on right eye)
18473 anaglyph green/magenta colored
18474 (green filter on left eye, magenta filter on right eye)
18477 anaglyph green/magenta color optimized with the least squares projection of dubois
18478 (green filter on left eye, magenta filter on right eye)
18481 anaglyph yellow/blue gray
18482 (yellow filter on left eye, blue filter on right eye)
18485 anaglyph yellow/blue half colored
18486 (yellow filter on left eye, blue filter on right eye)
18489 anaglyph yellow/blue colored
18490 (yellow filter on left eye, blue filter on right eye)
18493 anaglyph yellow/blue color optimized with the least squares projection of dubois
18494 (yellow filter on left eye, blue filter on right eye)
18497 mono output (left eye only)
18500 mono output (right eye only)
18503 checkerboard, left eye first
18506 checkerboard, right eye first
18509 interleaved columns, left eye first
18512 interleaved columns, right eye first
18518 Default value is @samp{arcd}.
18521 @subsection Examples
18525 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18531 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18537 @section streamselect, astreamselect
18538 Select video or audio streams.
18540 The filter accepts the following options:
18544 Set number of inputs. Default is 2.
18547 Set input indexes to remap to outputs.
18550 @subsection Commands
18552 The @code{streamselect} and @code{astreamselect} filter supports the following
18557 Set input indexes to remap to outputs.
18560 @subsection Examples
18564 Select first 5 seconds 1st stream and rest of time 2nd stream:
18566 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18570 Same as above, but for audio:
18572 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18579 Draw subtitles on top of input video using the libass library.
18581 To enable compilation of this filter you need to configure FFmpeg with
18582 @code{--enable-libass}. This filter also requires a build with libavcodec and
18583 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18584 Alpha) subtitles format.
18586 The filter accepts the following options:
18590 Set the filename of the subtitle file to read. It must be specified.
18592 @item original_size
18593 Specify the size of the original video, the video for which the ASS file
18594 was composed. For the syntax of this option, check the
18595 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18596 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18597 correctly scale the fonts if the aspect ratio has been changed.
18600 Set a directory path containing fonts that can be used by the filter.
18601 These fonts will be used in addition to whatever the font provider uses.
18604 Process alpha channel, by default alpha channel is untouched.
18607 Set subtitles input character encoding. @code{subtitles} filter only. Only
18608 useful if not UTF-8.
18610 @item stream_index, si
18611 Set subtitles stream index. @code{subtitles} filter only.
18614 Override default style or script info parameters of the subtitles. It accepts a
18615 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18618 If the first key is not specified, it is assumed that the first value
18619 specifies the @option{filename}.
18621 For example, to render the file @file{sub.srt} on top of the input
18622 video, use the command:
18627 which is equivalent to:
18629 subtitles=filename=sub.srt
18632 To render the default subtitles stream from file @file{video.mkv}, use:
18634 subtitles=video.mkv
18637 To render the second subtitles stream from that file, use:
18639 subtitles=video.mkv:si=1
18642 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18643 @code{DejaVu Serif}, use:
18645 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18648 @section super2xsai
18650 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18651 Interpolate) pixel art scaling algorithm.
18653 Useful for enlarging pixel art images without reducing sharpness.
18657 Swap two rectangular objects in video.
18659 This filter accepts the following options:
18669 Set 1st rect x coordinate.
18672 Set 1st rect y coordinate.
18675 Set 2nd rect x coordinate.
18678 Set 2nd rect y coordinate.
18680 All expressions are evaluated once for each frame.
18683 The all options are expressions containing the following constants:
18688 The input width and height.
18691 same as @var{w} / @var{h}
18694 input sample aspect ratio
18697 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18700 The number of the input frame, starting from 0.
18703 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18706 the position in the file of the input frame, NAN if unknown
18713 Blend successive video frames.
18719 Apply telecine process to the video.
18721 This filter accepts the following options:
18730 The default value is @code{top}.
18734 A string of numbers representing the pulldown pattern you wish to apply.
18735 The default value is @code{23}.
18739 Some typical patterns:
18744 24p: 2332 (preferred)
18751 24p: 222222222223 ("Euro pulldown")
18756 @section thistogram
18758 Compute and draw a color distribution histogram for the input video across time.
18760 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18761 at certain time, this filter shows also past histograms of number of frames defined
18762 by @code{width} option.
18764 The computed histogram is a representation of the color component
18765 distribution in an image.
18767 The filter accepts the following options:
18771 Set width of single color component output. Default value is @code{0}.
18772 Value of @code{0} means width will be picked from input video.
18773 This also set number of passed histograms to keep.
18774 Allowed range is [0, 8192].
18776 @item display_mode, d
18778 It accepts the following values:
18781 Per color component graphs are placed below each other.
18784 Per color component graphs are placed side by side.
18787 Presents information identical to that in the @code{parade}, except
18788 that the graphs representing color components are superimposed directly
18791 Default is @code{stack}.
18793 @item levels_mode, m
18794 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18795 Default is @code{linear}.
18797 @item components, c
18798 Set what color components to display.
18799 Default is @code{7}.
18802 Set background opacity. Default is @code{0.9}.
18805 Show envelope. Default is disabled.
18808 Set envelope color. Default is @code{gold}.
18813 Available values for slide is:
18816 Draw new frame when right border is reached.
18819 Replace old columns with new ones.
18822 Scroll from right to left.
18825 Scroll from left to right.
18828 Draw single picture.
18831 Default is @code{replace}.
18836 Apply threshold effect to video stream.
18838 This filter needs four video streams to perform thresholding.
18839 First stream is stream we are filtering.
18840 Second stream is holding threshold values, third stream is holding min values,
18841 and last, fourth stream is holding max values.
18843 The filter accepts the following option:
18847 Set which planes will be processed, unprocessed planes will be copied.
18848 By default value 0xf, all planes will be processed.
18851 For example if first stream pixel's component value is less then threshold value
18852 of pixel component from 2nd threshold stream, third stream value will picked,
18853 otherwise fourth stream pixel component value will be picked.
18855 Using color source filter one can perform various types of thresholding:
18857 @subsection Examples
18861 Binary threshold, using gray color as threshold:
18863 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18867 Inverted binary threshold, using gray color as threshold:
18869 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18873 Truncate binary threshold, using gray color as threshold:
18875 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18879 Threshold to zero, using gray color as threshold:
18881 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18885 Inverted threshold to zero, using gray color as threshold:
18887 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18892 Select the most representative frame in a given sequence of consecutive frames.
18894 The filter accepts the following options:
18898 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18899 will pick one of them, and then handle the next batch of @var{n} frames until
18900 the end. Default is @code{100}.
18903 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18904 value will result in a higher memory usage, so a high value is not recommended.
18906 @subsection Examples
18910 Extract one picture each 50 frames:
18916 Complete example of a thumbnail creation with @command{ffmpeg}:
18918 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18925 Tile several successive frames together.
18927 The @ref{untile} filter can do the reverse.
18929 The filter accepts the following options:
18934 Set the grid size (i.e. the number of lines and columns). For the syntax of
18935 this option, check the
18936 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18939 Set the maximum number of frames to render in the given area. It must be less
18940 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18941 the area will be used.
18944 Set the outer border margin in pixels.
18947 Set the inner border thickness (i.e. the number of pixels between frames). For
18948 more advanced padding options (such as having different values for the edges),
18949 refer to the pad video filter.
18952 Specify the color of the unused area. For the syntax of this option, check the
18953 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18954 The default value of @var{color} is "black".
18957 Set the number of frames to overlap when tiling several successive frames together.
18958 The value must be between @code{0} and @var{nb_frames - 1}.
18961 Set the number of frames to initially be empty before displaying first output frame.
18962 This controls how soon will one get first output frame.
18963 The value must be between @code{0} and @var{nb_frames - 1}.
18966 @subsection Examples
18970 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18972 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18974 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18975 duplicating each output frame to accommodate the originally detected frame
18979 Display @code{5} pictures in an area of @code{3x2} frames,
18980 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18981 mixed flat and named options:
18983 tile=3x2:nb_frames=5:padding=7:margin=2
18987 @section tinterlace
18989 Perform various types of temporal field interlacing.
18991 Frames are counted starting from 1, so the first input frame is
18994 The filter accepts the following options:
18999 Specify the mode of the interlacing. This option can also be specified
19000 as a value alone. See below for a list of values for this option.
19002 Available values are:
19006 Move odd frames into the upper field, even into the lower field,
19007 generating a double height frame at half frame rate.
19011 Frame 1 Frame 2 Frame 3 Frame 4
19013 11111 22222 33333 44444
19014 11111 22222 33333 44444
19015 11111 22222 33333 44444
19016 11111 22222 33333 44444
19030 Only output odd frames, even frames are dropped, generating a frame with
19031 unchanged height at half frame rate.
19036 Frame 1 Frame 2 Frame 3 Frame 4
19038 11111 22222 33333 44444
19039 11111 22222 33333 44444
19040 11111 22222 33333 44444
19041 11111 22222 33333 44444
19051 Only output even frames, odd frames are dropped, generating a frame with
19052 unchanged height at half frame rate.
19057 Frame 1 Frame 2 Frame 3 Frame 4
19059 11111 22222 33333 44444
19060 11111 22222 33333 44444
19061 11111 22222 33333 44444
19062 11111 22222 33333 44444
19072 Expand each frame to full height, but pad alternate lines with black,
19073 generating a frame with double height at the same input frame rate.
19078 Frame 1 Frame 2 Frame 3 Frame 4
19080 11111 22222 33333 44444
19081 11111 22222 33333 44444
19082 11111 22222 33333 44444
19083 11111 22222 33333 44444
19086 11111 ..... 33333 .....
19087 ..... 22222 ..... 44444
19088 11111 ..... 33333 .....
19089 ..... 22222 ..... 44444
19090 11111 ..... 33333 .....
19091 ..... 22222 ..... 44444
19092 11111 ..... 33333 .....
19093 ..... 22222 ..... 44444
19097 @item interleave_top, 4
19098 Interleave the upper field from odd frames with the lower field from
19099 even frames, generating a frame with unchanged height at half frame rate.
19104 Frame 1 Frame 2 Frame 3 Frame 4
19106 11111<- 22222 33333<- 44444
19107 11111 22222<- 33333 44444<-
19108 11111<- 22222 33333<- 44444
19109 11111 22222<- 33333 44444<-
19119 @item interleave_bottom, 5
19120 Interleave the lower field from odd frames with the upper field from
19121 even frames, generating a frame with unchanged height at half frame rate.
19126 Frame 1 Frame 2 Frame 3 Frame 4
19128 11111 22222<- 33333 44444<-
19129 11111<- 22222 33333<- 44444
19130 11111 22222<- 33333 44444<-
19131 11111<- 22222 33333<- 44444
19141 @item interlacex2, 6
19142 Double frame rate with unchanged height. Frames are inserted each
19143 containing the second temporal field from the previous input frame and
19144 the first temporal field from the next input frame. This mode relies on
19145 the top_field_first flag. Useful for interlaced video displays with no
19146 field synchronisation.
19151 Frame 1 Frame 2 Frame 3 Frame 4
19153 11111 22222 33333 44444
19154 11111 22222 33333 44444
19155 11111 22222 33333 44444
19156 11111 22222 33333 44444
19159 11111 22222 22222 33333 33333 44444 44444
19160 11111 11111 22222 22222 33333 33333 44444
19161 11111 22222 22222 33333 33333 44444 44444
19162 11111 11111 22222 22222 33333 33333 44444
19167 Move odd frames into the upper field, even into the lower field,
19168 generating a double height frame at same frame rate.
19173 Frame 1 Frame 2 Frame 3 Frame 4
19175 11111 22222 33333 44444
19176 11111 22222 33333 44444
19177 11111 22222 33333 44444
19178 11111 22222 33333 44444
19181 11111 33333 33333 55555
19182 22222 22222 44444 44444
19183 11111 33333 33333 55555
19184 22222 22222 44444 44444
19185 11111 33333 33333 55555
19186 22222 22222 44444 44444
19187 11111 33333 33333 55555
19188 22222 22222 44444 44444
19193 Numeric values are deprecated but are accepted for backward
19194 compatibility reasons.
19196 Default mode is @code{merge}.
19199 Specify flags influencing the filter process.
19201 Available value for @var{flags} is:
19204 @item low_pass_filter, vlpf
19205 Enable linear vertical low-pass filtering in the filter.
19206 Vertical low-pass filtering is required when creating an interlaced
19207 destination from a progressive source which contains high-frequency
19208 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19211 @item complex_filter, cvlpf
19212 Enable complex vertical low-pass filtering.
19213 This will slightly less reduce interlace 'twitter' and Moire
19214 patterning but better retain detail and subjective sharpness impression.
19217 Bypass already interlaced frames, only adjust the frame rate.
19220 Vertical low-pass filtering and bypassing already interlaced frames can only be
19221 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19226 Pick median pixels from several successive input video frames.
19228 The filter accepts the following options:
19232 Set radius of median filter.
19233 Default is 1. Allowed range is from 1 to 127.
19236 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19239 Set median percentile. Default value is @code{0.5}.
19240 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19241 minimum values, and @code{1} maximum values.
19246 Mix successive video frames.
19248 A description of the accepted options follows.
19252 The number of successive frames to mix. If unspecified, it defaults to 3.
19255 Specify weight of each input video frame.
19256 Each weight is separated by space. If number of weights is smaller than
19257 number of @var{frames} last specified weight will be used for all remaining
19261 Specify scale, if it is set it will be multiplied with sum
19262 of each weight multiplied with pixel values to give final destination
19263 pixel value. By default @var{scale} is auto scaled to sum of weights.
19266 @subsection Examples
19270 Average 7 successive frames:
19272 tmix=frames=7:weights="1 1 1 1 1 1 1"
19276 Apply simple temporal convolution:
19278 tmix=frames=3:weights="-1 3 -1"
19282 Similar as above but only showing temporal differences:
19284 tmix=frames=3:weights="-1 2 -1":scale=1
19290 Tone map colors from different dynamic ranges.
19292 This filter expects data in single precision floating point, as it needs to
19293 operate on (and can output) out-of-range values. Another filter, such as
19294 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19296 The tonemapping algorithms implemented only work on linear light, so input
19297 data should be linearized beforehand (and possibly correctly tagged).
19300 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19303 @subsection Options
19304 The filter accepts the following options.
19308 Set the tone map algorithm to use.
19310 Possible values are:
19313 Do not apply any tone map, only desaturate overbright pixels.
19316 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19317 in-range values, while distorting out-of-range values.
19320 Stretch the entire reference gamut to a linear multiple of the display.
19323 Fit a logarithmic transfer between the tone curves.
19326 Preserve overall image brightness with a simple curve, using nonlinear
19327 contrast, which results in flattening details and degrading color accuracy.
19330 Preserve both dark and bright details better than @var{reinhard}, at the cost
19331 of slightly darkening everything. Use it when detail preservation is more
19332 important than color and brightness accuracy.
19335 Smoothly map out-of-range values, while retaining contrast and colors for
19336 in-range material as much as possible. Use it when color accuracy is more
19337 important than detail preservation.
19343 Tune the tone mapping algorithm.
19345 This affects the following algorithms:
19351 Specifies the scale factor to use while stretching.
19355 Specifies the exponent of the function.
19359 Specify an extra linear coefficient to multiply into the signal before clipping.
19363 Specify the local contrast coefficient at the display peak.
19364 Default to 0.5, which means that in-gamut values will be about half as bright
19371 Specify the transition point from linear to mobius transform. Every value
19372 below this point is guaranteed to be mapped 1:1. The higher the value, the
19373 more accurate the result will be, at the cost of losing bright details.
19374 Default to 0.3, which due to the steep initial slope still preserves in-range
19375 colors fairly accurately.
19379 Apply desaturation for highlights that exceed this level of brightness. The
19380 higher the parameter, the more color information will be preserved. This
19381 setting helps prevent unnaturally blown-out colors for super-highlights, by
19382 (smoothly) turning into white instead. This makes images feel more natural,
19383 at the cost of reducing information about out-of-range colors.
19385 The default of 2.0 is somewhat conservative and will mostly just apply to
19386 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19388 This option works only if the input frame has a supported color tag.
19391 Override signal/nominal/reference peak with this value. Useful when the
19392 embedded peak information in display metadata is not reliable or when tone
19393 mapping from a lower range to a higher range.
19398 Temporarily pad video frames.
19400 The filter accepts the following options:
19404 Specify number of delay frames before input video stream. Default is 0.
19407 Specify number of padding frames after input video stream.
19408 Set to -1 to pad indefinitely. Default is 0.
19411 Set kind of frames added to beginning of stream.
19412 Can be either @var{add} or @var{clone}.
19413 With @var{add} frames of solid-color are added.
19414 With @var{clone} frames are clones of first frame.
19415 Default is @var{add}.
19418 Set kind of frames added to end of stream.
19419 Can be either @var{add} or @var{clone}.
19420 With @var{add} frames of solid-color are added.
19421 With @var{clone} frames are clones of last frame.
19422 Default is @var{add}.
19424 @item start_duration, stop_duration
19425 Specify the duration of the start/stop delay. See
19426 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19427 for the accepted syntax.
19428 These options override @var{start} and @var{stop}. Default is 0.
19431 Specify the color of the padded area. For the syntax of this option,
19432 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19433 manual,ffmpeg-utils}.
19435 The default value of @var{color} is "black".
19441 Transpose rows with columns in the input video and optionally flip it.
19443 It accepts the following parameters:
19448 Specify the transposition direction.
19450 Can assume the following values:
19452 @item 0, 4, cclock_flip
19453 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19461 Rotate by 90 degrees clockwise, that is:
19469 Rotate by 90 degrees counterclockwise, that is:
19476 @item 3, 7, clock_flip
19477 Rotate by 90 degrees clockwise and vertically flip, that is:
19485 For values between 4-7, the transposition is only done if the input
19486 video geometry is portrait and not landscape. These values are
19487 deprecated, the @code{passthrough} option should be used instead.
19489 Numerical values are deprecated, and should be dropped in favor of
19490 symbolic constants.
19493 Do not apply the transposition if the input geometry matches the one
19494 specified by the specified value. It accepts the following values:
19497 Always apply transposition.
19499 Preserve portrait geometry (when @var{height} >= @var{width}).
19501 Preserve landscape geometry (when @var{width} >= @var{height}).
19504 Default value is @code{none}.
19507 For example to rotate by 90 degrees clockwise and preserve portrait
19510 transpose=dir=1:passthrough=portrait
19513 The command above can also be specified as:
19515 transpose=1:portrait
19518 @section transpose_npp
19520 Transpose rows with columns in the input video and optionally flip it.
19521 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19523 It accepts the following parameters:
19528 Specify the transposition direction.
19530 Can assume the following values:
19533 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19536 Rotate by 90 degrees clockwise.
19539 Rotate by 90 degrees counterclockwise.
19542 Rotate by 90 degrees clockwise and vertically flip.
19546 Do not apply the transposition if the input geometry matches the one
19547 specified by the specified value. It accepts the following values:
19550 Always apply transposition. (default)
19552 Preserve portrait geometry (when @var{height} >= @var{width}).
19554 Preserve landscape geometry (when @var{width} >= @var{height}).
19560 Trim the input so that the output contains one continuous subpart of the input.
19562 It accepts the following parameters:
19565 Specify the time of the start of the kept section, i.e. the frame with the
19566 timestamp @var{start} will be the first frame in the output.
19569 Specify the time of the first frame that will be dropped, i.e. the frame
19570 immediately preceding the one with the timestamp @var{end} will be the last
19571 frame in the output.
19574 This is the same as @var{start}, except this option sets the start timestamp
19575 in timebase units instead of seconds.
19578 This is the same as @var{end}, except this option sets the end timestamp
19579 in timebase units instead of seconds.
19582 The maximum duration of the output in seconds.
19585 The number of the first frame that should be passed to the output.
19588 The number of the first frame that should be dropped.
19591 @option{start}, @option{end}, and @option{duration} are expressed as time
19592 duration specifications; see
19593 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19594 for the accepted syntax.
19596 Note that the first two sets of the start/end options and the @option{duration}
19597 option look at the frame timestamp, while the _frame variants simply count the
19598 frames that pass through the filter. Also note that this filter does not modify
19599 the timestamps. If you wish for the output timestamps to start at zero, insert a
19600 setpts filter after the trim filter.
19602 If multiple start or end options are set, this filter tries to be greedy and
19603 keep all the frames that match at least one of the specified constraints. To keep
19604 only the part that matches all the constraints at once, chain multiple trim
19607 The defaults are such that all the input is kept. So it is possible to set e.g.
19608 just the end values to keep everything before the specified time.
19613 Drop everything except the second minute of input:
19615 ffmpeg -i INPUT -vf trim=60:120
19619 Keep only the first second:
19621 ffmpeg -i INPUT -vf trim=duration=1
19626 @section unpremultiply
19627 Apply alpha unpremultiply effect to input video stream using first plane
19628 of second stream as alpha.
19630 Both streams must have same dimensions and same pixel format.
19632 The filter accepts the following option:
19636 Set which planes will be processed, unprocessed planes will be copied.
19637 By default value 0xf, all planes will be processed.
19639 If the format has 1 or 2 components, then luma is bit 0.
19640 If the format has 3 or 4 components:
19641 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19642 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19643 If present, the alpha channel is always the last bit.
19646 Do not require 2nd input for processing, instead use alpha plane from input stream.
19652 Sharpen or blur the input video.
19654 It accepts the following parameters:
19657 @item luma_msize_x, lx
19658 Set the luma matrix horizontal size. It must be an odd integer between
19659 3 and 23. The default value is 5.
19661 @item luma_msize_y, ly
19662 Set the luma matrix vertical size. It must be an odd integer between 3
19663 and 23. The default value is 5.
19665 @item luma_amount, la
19666 Set the luma effect strength. It must be a floating point number, reasonable
19667 values lay between -1.5 and 1.5.
19669 Negative values will blur the input video, while positive values will
19670 sharpen it, a value of zero will disable the effect.
19672 Default value is 1.0.
19674 @item chroma_msize_x, cx
19675 Set the chroma matrix horizontal size. It must be an odd integer
19676 between 3 and 23. The default value is 5.
19678 @item chroma_msize_y, cy
19679 Set the chroma matrix vertical size. It must be an odd integer
19680 between 3 and 23. The default value is 5.
19682 @item chroma_amount, ca
19683 Set the chroma effect strength. It must be a floating point number, reasonable
19684 values lay between -1.5 and 1.5.
19686 Negative values will blur the input video, while positive values will
19687 sharpen it, a value of zero will disable the effect.
19689 Default value is 0.0.
19693 All parameters are optional and default to the equivalent of the
19694 string '5:5:1.0:5:5:0.0'.
19696 @subsection Examples
19700 Apply strong luma sharpen effect:
19702 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19706 Apply a strong blur of both luma and chroma parameters:
19708 unsharp=7:7:-2:7:7:-2
19715 Decompose a video made of tiled images into the individual images.
19717 The frame rate of the output video is the frame rate of the input video
19718 multiplied by the number of tiles.
19720 This filter does the reverse of @ref{tile}.
19722 The filter accepts the following options:
19727 Set the grid size (i.e. the number of lines and columns). For the syntax of
19728 this option, check the
19729 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19732 @subsection Examples
19736 Produce a 1-second video from a still image file made of 25 frames stacked
19737 vertically, like an analogic film reel:
19739 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19745 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19746 the image at several (or - in the case of @option{quality} level @code{8} - all)
19747 shifts and average the results.
19749 The way this differs from the behavior of spp is that uspp actually encodes &
19750 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19751 DCT similar to MJPEG.
19753 The filter accepts the following options:
19757 Set quality. This option defines the number of levels for averaging. It accepts
19758 an integer in the range 0-8. If set to @code{0}, the filter will have no
19759 effect. A value of @code{8} means the higher quality. For each increment of
19760 that value the speed drops by a factor of approximately 2. Default value is
19764 Force a constant quantization parameter. If not set, the filter will use the QP
19765 from the video stream (if available).
19770 Convert 360 videos between various formats.
19772 The filter accepts the following options:
19778 Set format of the input/output video.
19786 Equirectangular projection.
19791 Cubemap with 3x2/6x1/1x6 layout.
19793 Format specific options:
19798 Set padding proportion for the input/output cubemap. Values in decimals.
19805 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)
19808 Default value is @b{@samp{0}}.
19809 Maximum value is @b{@samp{0.1}}.
19813 Set fixed padding for the input/output cubemap. Values in pixels.
19815 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19819 Set order of faces for the input/output cubemap. Choose one direction for each position.
19821 Designation of directions:
19837 Default value is @b{@samp{rludfb}}.
19841 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19843 Designation of angles:
19846 0 degrees clockwise
19848 90 degrees clockwise
19850 180 degrees clockwise
19852 270 degrees clockwise
19855 Default value is @b{@samp{000000}}.
19859 Equi-Angular Cubemap.
19866 Format specific options:
19871 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19873 If diagonal field of view is set it overrides horizontal and vertical field of view.
19878 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19880 If diagonal field of view is set it overrides horizontal and vertical field of view.
19886 Format specific options:
19891 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19893 If diagonal field of view is set it overrides horizontal and vertical field of view.
19898 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19900 If diagonal field of view is set it overrides horizontal and vertical field of view.
19906 Facebook's 360 formats.
19909 Stereographic format.
19911 Format specific options:
19916 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19918 If diagonal field of view is set it overrides horizontal and vertical field of view.
19923 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19925 If diagonal field of view is set it overrides horizontal and vertical field of view.
19932 Ball format, gives significant distortion toward the back.
19935 Hammer-Aitoff map projection format.
19938 Sinusoidal map projection format.
19941 Fisheye projection.
19943 Format specific options:
19948 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19950 If diagonal field of view is set it overrides horizontal and vertical field of view.
19955 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19957 If diagonal field of view is set it overrides horizontal and vertical field of view.
19961 Pannini projection.
19963 Format specific options:
19966 Set output pannini parameter.
19969 Set input pannini parameter.
19973 Cylindrical projection.
19975 Format specific options:
19980 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19982 If diagonal field of view is set it overrides horizontal and vertical field of view.
19987 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19989 If diagonal field of view is set it overrides horizontal and vertical field of view.
19993 Perspective projection. @i{(output only)}
19995 Format specific options:
19998 Set perspective parameter.
20002 Tetrahedron projection.
20005 Truncated square pyramid projection.
20009 Half equirectangular projection.
20014 Format specific options:
20019 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20021 If diagonal field of view is set it overrides horizontal and vertical field of view.
20026 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20028 If diagonal field of view is set it overrides horizontal and vertical field of view.
20032 Orthographic format.
20034 Format specific options:
20039 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20041 If diagonal field of view is set it overrides horizontal and vertical field of view.
20046 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20048 If diagonal field of view is set it overrides horizontal and vertical field of view.
20052 Octahedron projection.
20056 Set interpolation method.@*
20057 @i{Note: more complex interpolation methods require much more memory to run.}
20067 Bilinear interpolation.
20069 Lagrange9 interpolation.
20072 Bicubic interpolation.
20075 Lanczos interpolation.
20078 Spline16 interpolation.
20081 Gaussian interpolation.
20083 Mitchell interpolation.
20086 Default value is @b{@samp{line}}.
20090 Set the output video resolution.
20092 Default resolution depends on formats.
20096 Set the input/output stereo format.
20107 Default value is @b{@samp{2d}} for input and output format.
20112 Set rotation for the output video. Values in degrees.
20115 Set rotation order for the output video. Choose one item for each position.
20126 Default value is @b{@samp{ypr}}.
20131 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20135 Set if input video is flipped horizontally/vertically. Boolean values.
20138 Set if input video is transposed. Boolean value, by default disabled.
20141 Set if output video needs to be transposed. Boolean value, by default disabled.
20144 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20147 @subsection Examples
20151 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20153 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20156 Extract back view of Equi-Angular Cubemap:
20158 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20161 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20163 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20167 @subsection Commands
20169 This filter supports subset of above options as @ref{commands}.
20171 @section vaguedenoiser
20173 Apply a wavelet based denoiser.
20175 It transforms each frame from the video input into the wavelet domain,
20176 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20177 the obtained coefficients. It does an inverse wavelet transform after.
20178 Due to wavelet properties, it should give a nice smoothed result, and
20179 reduced noise, without blurring picture features.
20181 This filter accepts the following options:
20185 The filtering strength. The higher, the more filtered the video will be.
20186 Hard thresholding can use a higher threshold than soft thresholding
20187 before the video looks overfiltered. Default value is 2.
20190 The filtering method the filter will use.
20192 It accepts the following values:
20195 All values under the threshold will be zeroed.
20198 All values under the threshold will be zeroed. All values above will be
20199 reduced by the threshold.
20202 Scales or nullifies coefficients - intermediary between (more) soft and
20203 (less) hard thresholding.
20206 Default is garrote.
20209 Number of times, the wavelet will decompose the picture. Picture can't
20210 be decomposed beyond a particular point (typically, 8 for a 640x480
20211 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20214 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20217 A list of the planes to process. By default all planes are processed.
20220 The threshold type the filter will use.
20222 It accepts the following values:
20225 Threshold used is same for all decompositions.
20228 Threshold used depends also on each decomposition coefficients.
20231 Default is universal.
20234 @section vectorscope
20236 Display 2 color component values in the two dimensional graph (which is called
20239 This filter accepts the following options:
20243 Set vectorscope mode.
20245 It accepts the following values:
20249 Gray values are displayed on graph, higher brightness means more pixels have
20250 same component color value on location in graph. This is the default mode.
20253 Gray values are displayed on graph. Surrounding pixels values which are not
20254 present in video frame are drawn in gradient of 2 color components which are
20255 set by option @code{x} and @code{y}. The 3rd color component is static.
20258 Actual color components values present in video frame are displayed on graph.
20261 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20262 on graph increases value of another color component, which is luminance by
20263 default values of @code{x} and @code{y}.
20266 Actual colors present in video frame are displayed on graph. If two different
20267 colors map to same position on graph then color with higher value of component
20268 not present in graph is picked.
20271 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20272 component picked from radial gradient.
20276 Set which color component will be represented on X-axis. Default is @code{1}.
20279 Set which color component will be represented on Y-axis. Default is @code{2}.
20282 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20283 of color component which represents frequency of (X, Y) location in graph.
20288 No envelope, this is default.
20291 Instant envelope, even darkest single pixel will be clearly highlighted.
20294 Hold maximum and minimum values presented in graph over time. This way you
20295 can still spot out of range values without constantly looking at vectorscope.
20298 Peak and instant envelope combined together.
20302 Set what kind of graticule to draw.
20311 Set graticule opacity.
20314 Set graticule flags.
20318 Draw graticule for white point.
20321 Draw graticule for black point.
20324 Draw color points short names.
20328 Set background opacity.
20330 @item lthreshold, l
20331 Set low threshold for color component not represented on X or Y axis.
20332 Values lower than this value will be ignored. Default is 0.
20333 Note this value is multiplied with actual max possible value one pixel component
20334 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20337 @item hthreshold, h
20338 Set high threshold for color component not represented on X or Y axis.
20339 Values higher than this value will be ignored. Default is 1.
20340 Note this value is multiplied with actual max possible value one pixel component
20341 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20342 is 0.9 * 255 = 230.
20344 @item colorspace, c
20345 Set what kind of colorspace to use when drawing graticule.
20355 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20356 This means no tint, and output will remain gray.
20359 @anchor{vidstabdetect}
20360 @section vidstabdetect
20362 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20363 @ref{vidstabtransform} for pass 2.
20365 This filter generates a file with relative translation and rotation
20366 transform information about subsequent frames, which is then used by
20367 the @ref{vidstabtransform} filter.
20369 To enable compilation of this filter you need to configure FFmpeg with
20370 @code{--enable-libvidstab}.
20372 This filter accepts the following options:
20376 Set the path to the file used to write the transforms information.
20377 Default value is @file{transforms.trf}.
20380 Set how shaky the video is and how quick the camera is. It accepts an
20381 integer in the range 1-10, a value of 1 means little shakiness, a
20382 value of 10 means strong shakiness. Default value is 5.
20385 Set the accuracy of the detection process. It must be a value in the
20386 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20387 accuracy. Default value is 15.
20390 Set stepsize of the search process. The region around minimum is
20391 scanned with 1 pixel resolution. Default value is 6.
20394 Set minimum contrast. Below this value a local measurement field is
20395 discarded. Must be a floating point value in the range 0-1. Default
20399 Set reference frame number for tripod mode.
20401 If enabled, the motion of the frames is compared to a reference frame
20402 in the filtered stream, identified by the specified number. The idea
20403 is to compensate all movements in a more-or-less static scene and keep
20404 the camera view absolutely still.
20406 If set to 0, it is disabled. The frames are counted starting from 1.
20409 Show fields and transforms in the resulting frames. It accepts an
20410 integer in the range 0-2. Default value is 0, which disables any
20414 @subsection Examples
20418 Use default values:
20424 Analyze strongly shaky movie and put the results in file
20425 @file{mytransforms.trf}:
20427 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20431 Visualize the result of internal transformations in the resulting
20434 vidstabdetect=show=1
20438 Analyze a video with medium shakiness using @command{ffmpeg}:
20440 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20444 @anchor{vidstabtransform}
20445 @section vidstabtransform
20447 Video stabilization/deshaking: pass 2 of 2,
20448 see @ref{vidstabdetect} for pass 1.
20450 Read a file with transform information for each frame and
20451 apply/compensate them. Together with the @ref{vidstabdetect}
20452 filter this can be used to deshake videos. See also
20453 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20454 the @ref{unsharp} filter, see below.
20456 To enable compilation of this filter you need to configure FFmpeg with
20457 @code{--enable-libvidstab}.
20459 @subsection Options
20463 Set path to the file used to read the transforms. Default value is
20464 @file{transforms.trf}.
20467 Set the number of frames (value*2 + 1) used for lowpass filtering the
20468 camera movements. Default value is 10.
20470 For example a number of 10 means that 21 frames are used (10 in the
20471 past and 10 in the future) to smoothen the motion in the video. A
20472 larger value leads to a smoother video, but limits the acceleration of
20473 the camera (pan/tilt movements). 0 is a special case where a static
20474 camera is simulated.
20477 Set the camera path optimization algorithm.
20479 Accepted values are:
20482 gaussian kernel low-pass filter on camera motion (default)
20484 averaging on transformations
20488 Set maximal number of pixels to translate frames. Default value is -1,
20492 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20493 value is -1, meaning no limit.
20496 Specify how to deal with borders that may be visible due to movement
20499 Available values are:
20502 keep image information from previous frame (default)
20504 fill the border black
20508 Invert transforms if set to 1. Default value is 0.
20511 Consider transforms as relative to previous frame if set to 1,
20512 absolute if set to 0. Default value is 0.
20515 Set percentage to zoom. A positive value will result in a zoom-in
20516 effect, a negative value in a zoom-out effect. Default value is 0 (no
20520 Set optimal zooming to avoid borders.
20522 Accepted values are:
20527 optimal static zoom value is determined (only very strong movements
20528 will lead to visible borders) (default)
20530 optimal adaptive zoom value is determined (no borders will be
20531 visible), see @option{zoomspeed}
20534 Note that the value given at zoom is added to the one calculated here.
20537 Set percent to zoom maximally each frame (enabled when
20538 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20542 Specify type of interpolation.
20544 Available values are:
20549 linear only horizontal
20551 linear in both directions (default)
20553 cubic in both directions (slow)
20557 Enable virtual tripod mode if set to 1, which is equivalent to
20558 @code{relative=0:smoothing=0}. Default value is 0.
20560 Use also @code{tripod} option of @ref{vidstabdetect}.
20563 Increase log verbosity if set to 1. Also the detected global motions
20564 are written to the temporary file @file{global_motions.trf}. Default
20568 @subsection Examples
20572 Use @command{ffmpeg} for a typical stabilization with default values:
20574 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20577 Note the use of the @ref{unsharp} filter which is always recommended.
20580 Zoom in a bit more and load transform data from a given file:
20582 vidstabtransform=zoom=5:input="mytransforms.trf"
20586 Smoothen the video even more:
20588 vidstabtransform=smoothing=30
20594 Flip the input video vertically.
20596 For example, to vertically flip a video with @command{ffmpeg}:
20598 ffmpeg -i in.avi -vf "vflip" out.avi
20603 Detect variable frame rate video.
20605 This filter tries to detect if the input is variable or constant frame rate.
20607 At end it will output number of frames detected as having variable delta pts,
20608 and ones with constant delta pts.
20609 If there was frames with variable delta, than it will also show min, max and
20610 average delta encountered.
20614 Boost or alter saturation.
20616 The filter accepts the following options:
20619 Set strength of boost if positive value or strength of alter if negative value.
20620 Default is 0. Allowed range is from -2 to 2.
20623 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20626 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20629 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20632 Set the red luma coefficient.
20635 Set the green luma coefficient.
20638 Set the blue luma coefficient.
20641 If @code{intensity} is negative and this is set to 1, colors will change,
20642 otherwise colors will be less saturated, more towards gray.
20645 @subsection Commands
20647 This filter supports the all above options as @ref{commands}.
20652 Make or reverse a natural vignetting effect.
20654 The filter accepts the following options:
20658 Set lens angle expression as a number of radians.
20660 The value is clipped in the @code{[0,PI/2]} range.
20662 Default value: @code{"PI/5"}
20666 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20670 Set forward/backward mode.
20672 Available modes are:
20675 The larger the distance from the central point, the darker the image becomes.
20678 The larger the distance from the central point, the brighter the image becomes.
20679 This can be used to reverse a vignette effect, though there is no automatic
20680 detection to extract the lens @option{angle} and other settings (yet). It can
20681 also be used to create a burning effect.
20684 Default value is @samp{forward}.
20687 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20689 It accepts the following values:
20692 Evaluate expressions only once during the filter initialization.
20695 Evaluate expressions for each incoming frame. This is way slower than the
20696 @samp{init} mode since it requires all the scalers to be re-computed, but it
20697 allows advanced dynamic expressions.
20700 Default value is @samp{init}.
20703 Set dithering to reduce the circular banding effects. Default is @code{1}
20707 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20708 Setting this value to the SAR of the input will make a rectangular vignetting
20709 following the dimensions of the video.
20711 Default is @code{1/1}.
20714 @subsection Expressions
20716 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20717 following parameters.
20722 input width and height
20725 the number of input frame, starting from 0
20728 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20729 @var{TB} units, NAN if undefined
20732 frame rate of the input video, NAN if the input frame rate is unknown
20735 the PTS (Presentation TimeStamp) of the filtered video frame,
20736 expressed in seconds, NAN if undefined
20739 time base of the input video
20743 @subsection Examples
20747 Apply simple strong vignetting effect:
20753 Make a flickering vignetting:
20755 vignette='PI/4+random(1)*PI/50':eval=frame
20760 @section vmafmotion
20762 Obtain the average VMAF motion score of a video.
20763 It is one of the component metrics of VMAF.
20765 The obtained average motion score is printed through the logging system.
20767 The filter accepts the following options:
20771 If specified, the filter will use the named file to save the motion score of
20772 each frame with respect to the previous frame.
20773 When filename equals "-" the data is sent to standard output.
20778 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20782 Stack input videos vertically.
20784 All streams must be of same pixel format and of same width.
20786 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20787 to create same output.
20789 The filter accepts the following options:
20793 Set number of input streams. Default is 2.
20796 If set to 1, force the output to terminate when the shortest input
20797 terminates. Default value is 0.
20802 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20803 Deinterlacing Filter").
20805 Based on the process described by Martin Weston for BBC R&D, and
20806 implemented based on the de-interlace algorithm written by Jim
20807 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20808 uses filter coefficients calculated by BBC R&D.
20810 This filter uses field-dominance information in frame to decide which
20811 of each pair of fields to place first in the output.
20812 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20814 There are two sets of filter coefficients, so called "simple"
20815 and "complex". Which set of filter coefficients is used can
20816 be set by passing an optional parameter:
20820 Set the interlacing filter coefficients. Accepts one of the following values:
20824 Simple filter coefficient set.
20826 More-complex filter coefficient set.
20828 Default value is @samp{complex}.
20831 Specify which frames to deinterlace. Accepts one of the following values:
20835 Deinterlace all frames,
20837 Only deinterlace frames marked as interlaced.
20840 Default value is @samp{all}.
20844 Video waveform monitor.
20846 The waveform monitor plots color component intensity. By default luminance
20847 only. Each column of the waveform corresponds to a column of pixels in the
20850 It accepts the following options:
20854 Can be either @code{row}, or @code{column}. Default is @code{column}.
20855 In row mode, the graph on the left side represents color component value 0 and
20856 the right side represents value = 255. In column mode, the top side represents
20857 color component value = 0 and bottom side represents value = 255.
20860 Set intensity. Smaller values are useful to find out how many values of the same
20861 luminance are distributed across input rows/columns.
20862 Default value is @code{0.04}. Allowed range is [0, 1].
20865 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20866 In mirrored mode, higher values will be represented on the left
20867 side for @code{row} mode and at the top for @code{column} mode. Default is
20868 @code{1} (mirrored).
20872 It accepts the following values:
20875 Presents information identical to that in the @code{parade}, except
20876 that the graphs representing color components are superimposed directly
20879 This display mode makes it easier to spot relative differences or similarities
20880 in overlapping areas of the color components that are supposed to be identical,
20881 such as neutral whites, grays, or blacks.
20884 Display separate graph for the color components side by side in
20885 @code{row} mode or one below the other in @code{column} mode.
20888 Display separate graph for the color components side by side in
20889 @code{column} mode or one below the other in @code{row} mode.
20891 Using this display mode makes it easy to spot color casts in the highlights
20892 and shadows of an image, by comparing the contours of the top and the bottom
20893 graphs of each waveform. Since whites, grays, and blacks are characterized
20894 by exactly equal amounts of red, green, and blue, neutral areas of the picture
20895 should display three waveforms of roughly equal width/height. If not, the
20896 correction is easy to perform by making level adjustments the three waveforms.
20898 Default is @code{stack}.
20900 @item components, c
20901 Set which color components to display. Default is 1, which means only luminance
20902 or red color component if input is in RGB colorspace. If is set for example to
20903 7 it will display all 3 (if) available color components.
20908 No envelope, this is default.
20911 Instant envelope, minimum and maximum values presented in graph will be easily
20912 visible even with small @code{step} value.
20915 Hold minimum and maximum values presented in graph across time. This way you
20916 can still spot out of range values without constantly looking at waveforms.
20919 Peak and instant envelope combined together.
20925 No filtering, this is default.
20928 Luma and chroma combined together.
20931 Similar as above, but shows difference between blue and red chroma.
20934 Similar as above, but use different colors.
20937 Similar as above, but again with different colors.
20940 Displays only chroma.
20943 Displays actual color value on waveform.
20946 Similar as above, but with luma showing frequency of chroma values.
20950 Set which graticule to display.
20954 Do not display graticule.
20957 Display green graticule showing legal broadcast ranges.
20960 Display orange graticule showing legal broadcast ranges.
20963 Display invert graticule showing legal broadcast ranges.
20967 Set graticule opacity.
20970 Set graticule flags.
20974 Draw numbers above lines. By default enabled.
20977 Draw dots instead of lines.
20981 Set scale used for displaying graticule.
20988 Default is digital.
20991 Set background opacity.
20995 Set tint for output.
20996 Only used with lowpass filter and when display is not overlay and input
20997 pixel formats are not RGB.
21000 @section weave, doubleweave
21002 The @code{weave} takes a field-based video input and join
21003 each two sequential fields into single frame, producing a new double
21004 height clip with half the frame rate and half the frame count.
21006 The @code{doubleweave} works same as @code{weave} but without
21007 halving frame rate and frame count.
21009 It accepts the following option:
21013 Set first field. Available values are:
21017 Set the frame as top-field-first.
21020 Set the frame as bottom-field-first.
21024 @subsection Examples
21028 Interlace video using @ref{select} and @ref{separatefields} filter:
21030 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21035 Apply the xBR high-quality magnification filter which is designed for pixel
21036 art. It follows a set of edge-detection rules, see
21037 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21039 It accepts the following option:
21043 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21044 @code{3xBR} and @code{4} for @code{4xBR}.
21045 Default is @code{3}.
21050 Apply cross fade from one input video stream to another input video stream.
21051 The cross fade is applied for specified duration.
21053 The filter accepts the following options:
21057 Set one of available transition effects:
21105 Default transition effect is fade.
21108 Set cross fade duration in seconds.
21109 Default duration is 1 second.
21112 Set cross fade start relative to first input stream in seconds.
21113 Default offset is 0.
21116 Set expression for custom transition effect.
21118 The expressions can use the following variables and functions:
21123 The coordinates of the current sample.
21127 The width and height of the image.
21130 Progress of transition effect.
21133 Currently processed plane.
21136 Return value of first input at current location and plane.
21139 Return value of second input at current location and plane.
21145 Return the value of the pixel at location (@var{x},@var{y}) of the
21146 first/second/third/fourth component of first input.
21152 Return the value of the pixel at location (@var{x},@var{y}) of the
21153 first/second/third/fourth component of second input.
21157 @subsection Examples
21161 Cross fade from one input video to another input video, with fade transition and duration of transition
21162 of 2 seconds starting at offset of 5 seconds:
21164 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21169 Pick median pixels from several input videos.
21171 The filter accepts the following options:
21175 Set number of inputs.
21176 Default is 3. Allowed range is from 3 to 255.
21177 If number of inputs is even number, than result will be mean value between two median values.
21180 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21183 Set median percentile. Default value is @code{0.5}.
21184 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21185 minimum values, and @code{1} maximum values.
21189 Stack video inputs into custom layout.
21191 All streams must be of same pixel format.
21193 The filter accepts the following options:
21197 Set number of input streams. Default is 2.
21200 Specify layout of inputs.
21201 This option requires the desired layout configuration to be explicitly set by the user.
21202 This sets position of each video input in output. Each input
21203 is separated by '|'.
21204 The first number represents the column, and the second number represents the row.
21205 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21206 where X is video input from which to take width or height.
21207 Multiple values can be used when separated by '+'. In such
21208 case values are summed together.
21210 Note that if inputs are of different sizes gaps may appear, as not all of
21211 the output video frame will be filled. Similarly, videos can overlap each
21212 other if their position doesn't leave enough space for the full frame of
21215 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21216 a layout must be set by the user.
21219 If set to 1, force the output to terminate when the shortest input
21220 terminates. Default value is 0.
21223 If set to valid color, all unused pixels will be filled with that color.
21224 By default fill is set to none, so it is disabled.
21227 @subsection Examples
21231 Display 4 inputs into 2x2 grid.
21235 input1(0, 0) | input3(w0, 0)
21236 input2(0, h0) | input4(w0, h0)
21240 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21243 Note that if inputs are of different sizes, gaps or overlaps may occur.
21246 Display 4 inputs into 1x4 grid.
21253 input4(0, h0+h1+h2)
21257 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21260 Note that if inputs are of different widths, unused space will appear.
21263 Display 9 inputs into 3x3 grid.
21267 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21268 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21269 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21273 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
21276 Note that if inputs are of different sizes, gaps or overlaps may occur.
21279 Display 16 inputs into 4x4 grid.
21283 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21284 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21285 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21286 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21290 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|
21291 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
21294 Note that if inputs are of different sizes, gaps or overlaps may occur.
21301 Deinterlace the input video ("yadif" means "yet another deinterlacing
21304 It accepts the following parameters:
21310 The interlacing mode to adopt. It accepts one of the following values:
21313 @item 0, send_frame
21314 Output one frame for each frame.
21315 @item 1, send_field
21316 Output one frame for each field.
21317 @item 2, send_frame_nospatial
21318 Like @code{send_frame}, but it skips the spatial interlacing check.
21319 @item 3, send_field_nospatial
21320 Like @code{send_field}, but it skips the spatial interlacing check.
21323 The default value is @code{send_frame}.
21326 The picture field parity assumed for the input interlaced video. It accepts one
21327 of the following values:
21331 Assume the top field is first.
21333 Assume the bottom field is first.
21335 Enable automatic detection of field parity.
21338 The default value is @code{auto}.
21339 If the interlacing is unknown or the decoder does not export this information,
21340 top field first will be assumed.
21343 Specify which frames to deinterlace. Accepts one of the following
21348 Deinterlace all frames.
21349 @item 1, interlaced
21350 Only deinterlace frames marked as interlaced.
21353 The default value is @code{all}.
21356 @section yadif_cuda
21358 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21359 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21362 It accepts the following parameters:
21368 The interlacing mode to adopt. It accepts one of the following values:
21371 @item 0, send_frame
21372 Output one frame for each frame.
21373 @item 1, send_field
21374 Output one frame for each field.
21375 @item 2, send_frame_nospatial
21376 Like @code{send_frame}, but it skips the spatial interlacing check.
21377 @item 3, send_field_nospatial
21378 Like @code{send_field}, but it skips the spatial interlacing check.
21381 The default value is @code{send_frame}.
21384 The picture field parity assumed for the input interlaced video. It accepts one
21385 of the following values:
21389 Assume the top field is first.
21391 Assume the bottom field is first.
21393 Enable automatic detection of field parity.
21396 The default value is @code{auto}.
21397 If the interlacing is unknown or the decoder does not export this information,
21398 top field first will be assumed.
21401 Specify which frames to deinterlace. Accepts one of the following
21406 Deinterlace all frames.
21407 @item 1, interlaced
21408 Only deinterlace frames marked as interlaced.
21411 The default value is @code{all}.
21416 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21417 The algorithm is described in
21418 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21420 It accepts the following parameters:
21424 Set the window radius. Default value is 3.
21427 Set which planes to filter. Default is only the first plane.
21430 Set blur strength. Default value is 128.
21433 @subsection Commands
21434 This filter supports same @ref{commands} as options.
21438 Apply Zoom & Pan effect.
21440 This filter accepts the following options:
21444 Set the zoom expression. Range is 1-10. Default is 1.
21448 Set the x and y expression. Default is 0.
21451 Set the duration expression in number of frames.
21452 This sets for how many number of frames effect will last for
21453 single input image.
21456 Set the output image size, default is 'hd720'.
21459 Set the output frame rate, default is '25'.
21462 Each expression can contain the following constants:
21481 Output frame count.
21484 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21486 @item out_time, time, ot
21487 The output timestamp expressed in seconds.
21491 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21492 for current input frame.
21496 'x' and 'y' of last output frame of previous input frame or 0 when there was
21497 not yet such frame (first input frame).
21500 Last calculated zoom from 'z' expression for current input frame.
21503 Last calculated zoom of last output frame of previous input frame.
21506 Number of output frames for current input frame. Calculated from 'd' expression
21507 for each input frame.
21510 number of output frames created for previous input frame
21513 Rational number: input width / input height
21516 sample aspect ratio
21519 display aspect ratio
21523 @subsection Examples
21527 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21529 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
21533 Zoom in up to 1.5x and pan always at center of picture:
21535 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21539 Same as above but without pausing:
21541 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21545 Zoom in 2x into center of picture only for the first second of the input video:
21547 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21554 Scale (resize) the input video, using the z.lib library:
21555 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21556 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21558 The zscale filter forces the output display aspect ratio to be the same
21559 as the input, by changing the output sample aspect ratio.
21561 If the input image format is different from the format requested by
21562 the next filter, the zscale filter will convert the input to the
21565 @subsection Options
21566 The filter accepts the following options.
21571 Set the output video dimension expression. Default value is the input
21574 If the @var{width} or @var{w} value is 0, the input width is used for
21575 the output. If the @var{height} or @var{h} value is 0, the input height
21576 is used for the output.
21578 If one and only one of the values is -n with n >= 1, the zscale filter
21579 will use a value that maintains the aspect ratio of the input image,
21580 calculated from the other specified dimension. After that it will,
21581 however, make sure that the calculated dimension is divisible by n and
21582 adjust the value if necessary.
21584 If both values are -n with n >= 1, the behavior will be identical to
21585 both values being set to 0 as previously detailed.
21587 See below for the list of accepted constants for use in the dimension
21591 Set the video size. For the syntax of this option, check the
21592 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21595 Set the dither type.
21597 Possible values are:
21602 @item error_diffusion
21608 Set the resize filter type.
21610 Possible values are:
21620 Default is bilinear.
21623 Set the color range.
21625 Possible values are:
21632 Default is same as input.
21635 Set the color primaries.
21637 Possible values are:
21647 Default is same as input.
21650 Set the transfer characteristics.
21652 Possible values are:
21666 Default is same as input.
21669 Set the colorspace matrix.
21671 Possible value are:
21682 Default is same as input.
21685 Set the input color range.
21687 Possible values are:
21694 Default is same as input.
21696 @item primariesin, pin
21697 Set the input color primaries.
21699 Possible values are:
21709 Default is same as input.
21711 @item transferin, tin
21712 Set the input transfer characteristics.
21714 Possible values are:
21725 Default is same as input.
21727 @item matrixin, min
21728 Set the input colorspace matrix.
21730 Possible value are:
21742 Set the output chroma location.
21744 Possible values are:
21755 @item chromalin, cin
21756 Set the input chroma location.
21758 Possible values are:
21770 Set the nominal peak luminance.
21773 The values of the @option{w} and @option{h} options are expressions
21774 containing the following constants:
21779 The input width and height
21783 These are the same as @var{in_w} and @var{in_h}.
21787 The output (scaled) width and height
21791 These are the same as @var{out_w} and @var{out_h}
21794 The same as @var{iw} / @var{ih}
21797 input sample aspect ratio
21800 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21804 horizontal and vertical input chroma subsample values. For example for the
21805 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21809 horizontal and vertical output chroma subsample values. For example for the
21810 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21813 @subsection Commands
21815 This filter supports the following commands:
21819 Set the output video dimension expression.
21820 The command accepts the same syntax of the corresponding option.
21822 If the specified expression is not valid, it is kept at its current
21826 @c man end VIDEO FILTERS
21828 @chapter OpenCL Video Filters
21829 @c man begin OPENCL VIDEO FILTERS
21831 Below is a description of the currently available OpenCL video filters.
21833 To enable compilation of these filters you need to configure FFmpeg with
21834 @code{--enable-opencl}.
21836 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21839 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21840 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21841 given device parameters.
21843 @item -filter_hw_device @var{name}
21844 Pass the hardware device called @var{name} to all filters in any filter graph.
21848 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21852 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21854 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21858 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.
21860 @section avgblur_opencl
21862 Apply average blur filter.
21864 The filter accepts the following options:
21868 Set horizontal radius size.
21869 Range is @code{[1, 1024]} and default value is @code{1}.
21872 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21875 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21878 @subsection Example
21882 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.
21884 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21888 @section boxblur_opencl
21890 Apply a boxblur algorithm to the input video.
21892 It accepts the following parameters:
21896 @item luma_radius, lr
21897 @item luma_power, lp
21898 @item chroma_radius, cr
21899 @item chroma_power, cp
21900 @item alpha_radius, ar
21901 @item alpha_power, ap
21905 A description of the accepted options follows.
21908 @item luma_radius, lr
21909 @item chroma_radius, cr
21910 @item alpha_radius, ar
21911 Set an expression for the box radius in pixels used for blurring the
21912 corresponding input plane.
21914 The radius value must be a non-negative number, and must not be
21915 greater than the value of the expression @code{min(w,h)/2} for the
21916 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
21919 Default value for @option{luma_radius} is "2". If not specified,
21920 @option{chroma_radius} and @option{alpha_radius} default to the
21921 corresponding value set for @option{luma_radius}.
21923 The expressions can contain the following constants:
21927 The input width and height in pixels.
21931 The input chroma image width and height in pixels.
21935 The horizontal and vertical chroma subsample values. For example, for the
21936 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
21939 @item luma_power, lp
21940 @item chroma_power, cp
21941 @item alpha_power, ap
21942 Specify how many times the boxblur filter is applied to the
21943 corresponding plane.
21945 Default value for @option{luma_power} is 2. If not specified,
21946 @option{chroma_power} and @option{alpha_power} default to the
21947 corresponding value set for @option{luma_power}.
21949 A value of 0 will disable the effect.
21952 @subsection Examples
21954 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.
21958 Apply a boxblur filter with the luma, chroma, and alpha radius
21959 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.
21961 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
21962 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
21966 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.
21968 For the luma plane, a 2x2 box radius will be run once.
21970 For the chroma plane, a 4x4 box radius will be run 5 times.
21972 For the alpha plane, a 3x3 box radius will be run 7 times.
21974 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
21978 @section colorkey_opencl
21979 RGB colorspace color keying.
21981 The filter accepts the following options:
21985 The color which will be replaced with transparency.
21988 Similarity percentage with the key color.
21990 0.01 matches only the exact key color, while 1.0 matches everything.
21995 0.0 makes pixels either fully transparent, or not transparent at all.
21997 Higher values result in semi-transparent pixels, with a higher transparency
21998 the more similar the pixels color is to the key color.
22001 @subsection Examples
22005 Make every semi-green pixel in the input transparent with some slight blending:
22007 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22011 @section convolution_opencl
22013 Apply convolution of 3x3, 5x5, 7x7 matrix.
22015 The filter accepts the following options:
22022 Set matrix for each plane.
22023 Matrix is sequence of 9, 25 or 49 signed numbers.
22024 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22030 Set multiplier for calculated value for each plane.
22031 If unset or 0, it will be sum of all matrix elements.
22032 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22038 Set bias for each plane. This value is added to the result of the multiplication.
22039 Useful for making the overall image brighter or darker.
22040 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22044 @subsection Examples
22050 -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
22056 -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
22060 Apply edge enhance:
22062 -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
22068 -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
22072 Apply laplacian edge detector which includes diagonals:
22074 -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
22080 -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
22084 @section erosion_opencl
22086 Apply erosion effect to the video.
22088 This filter replaces the pixel by the local(3x3) minimum.
22090 It accepts the following options:
22097 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22098 If @code{0}, plane will remain unchanged.
22101 Flag which specifies the pixel to refer to.
22102 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22104 Flags to local 3x3 coordinates region centered on @code{x}:
22113 @subsection Example
22117 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.
22119 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22123 @section deshake_opencl
22124 Feature-point based video stabilization filter.
22126 The filter accepts the following options:
22130 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22133 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22135 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22137 Viewing point matches in the output video is only supported for RGB input.
22139 Defaults to @code{0}.
22141 @item adaptive_crop
22142 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22144 Defaults to @code{1}.
22146 @item refine_features
22147 Whether or not feature points should be refined at a sub-pixel level.
22149 This can be turned off for a slight performance gain at the cost of precision.
22151 Defaults to @code{1}.
22153 @item smooth_strength
22154 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22156 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22158 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22160 Defaults to @code{0.0}.
22162 @item smooth_window_multiplier
22163 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22165 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22167 Acceptable values range from @code{0.1} to @code{10.0}.
22169 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22170 potentially improving smoothness, but also increase latency and memory usage.
22172 Defaults to @code{2.0}.
22176 @subsection Examples
22180 Stabilize a video with a fixed, medium smoothing strength:
22182 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22186 Stabilize a video with debugging (both in console and in rendered video):
22188 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22192 @section dilation_opencl
22194 Apply dilation effect to the video.
22196 This filter replaces the pixel by the local(3x3) maximum.
22198 It accepts the following options:
22205 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22206 If @code{0}, plane will remain unchanged.
22209 Flag which specifies the pixel to refer to.
22210 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22212 Flags to local 3x3 coordinates region centered on @code{x}:
22221 @subsection Example
22225 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.
22227 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22231 @section nlmeans_opencl
22233 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22235 @section overlay_opencl
22237 Overlay one video on top of another.
22239 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22240 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22242 The filter accepts the following options:
22247 Set the x coordinate of the overlaid video on the main video.
22248 Default value is @code{0}.
22251 Set the y coordinate of the overlaid video on the main video.
22252 Default value is @code{0}.
22256 @subsection Examples
22260 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22262 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22265 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22267 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22272 @section pad_opencl
22274 Add paddings to the input image, and place the original input at the
22275 provided @var{x}, @var{y} coordinates.
22277 It accepts the following options:
22282 Specify an expression for the size of the output image with the
22283 paddings added. If the value for @var{width} or @var{height} is 0, the
22284 corresponding input size is used for the output.
22286 The @var{width} expression can reference the value set by the
22287 @var{height} expression, and vice versa.
22289 The default value of @var{width} and @var{height} is 0.
22293 Specify the offsets to place the input image at within the padded area,
22294 with respect to the top/left border of the output image.
22296 The @var{x} expression can reference the value set by the @var{y}
22297 expression, and vice versa.
22299 The default value of @var{x} and @var{y} is 0.
22301 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22302 so the input image is centered on the padded area.
22305 Specify the color of the padded area. For the syntax of this option,
22306 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22307 manual,ffmpeg-utils}.
22310 Pad to an aspect instead to a resolution.
22313 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22314 options are expressions containing the following constants:
22319 The input video width and height.
22323 These are the same as @var{in_w} and @var{in_h}.
22327 The output width and height (the size of the padded area), as
22328 specified by the @var{width} and @var{height} expressions.
22332 These are the same as @var{out_w} and @var{out_h}.
22336 The x and y offsets as specified by the @var{x} and @var{y}
22337 expressions, or NAN if not yet specified.
22340 same as @var{iw} / @var{ih}
22343 input sample aspect ratio
22346 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22349 @section prewitt_opencl
22351 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22353 The filter accepts the following option:
22357 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22360 Set value which will be multiplied with filtered result.
22361 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22364 Set value which will be added to filtered result.
22365 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22368 @subsection Example
22372 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22374 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22378 @anchor{program_opencl}
22379 @section program_opencl
22381 Filter video using an OpenCL program.
22386 OpenCL program source file.
22389 Kernel name in program.
22392 Number of inputs to the filter. Defaults to 1.
22395 Size of output frames. Defaults to the same as the first input.
22399 The @code{program_opencl} filter also supports the @ref{framesync} options.
22401 The program source file must contain a kernel function with the given name,
22402 which will be run once for each plane of the output. Each run on a plane
22403 gets enqueued as a separate 2D global NDRange with one work-item for each
22404 pixel to be generated. The global ID offset for each work-item is therefore
22405 the coordinates of a pixel in the destination image.
22407 The kernel function needs to take the following arguments:
22410 Destination image, @var{__write_only image2d_t}.
22412 This image will become the output; the kernel should write all of it.
22414 Frame index, @var{unsigned int}.
22416 This is a counter starting from zero and increasing by one for each frame.
22418 Source images, @var{__read_only image2d_t}.
22420 These are the most recent images on each input. The kernel may read from
22421 them to generate the output, but they can't be written to.
22428 Copy the input to the output (output must be the same size as the input).
22430 __kernel void copy(__write_only image2d_t destination,
22431 unsigned int index,
22432 __read_only image2d_t source)
22434 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22436 int2 location = (int2)(get_global_id(0), get_global_id(1));
22438 float4 value = read_imagef(source, sampler, location);
22440 write_imagef(destination, location, value);
22445 Apply a simple transformation, rotating the input by an amount increasing
22446 with the index counter. Pixel values are linearly interpolated by the
22447 sampler, and the output need not have the same dimensions as the input.
22449 __kernel void rotate_image(__write_only image2d_t dst,
22450 unsigned int index,
22451 __read_only image2d_t src)
22453 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22454 CLK_FILTER_LINEAR);
22456 float angle = (float)index / 100.0f;
22458 float2 dst_dim = convert_float2(get_image_dim(dst));
22459 float2 src_dim = convert_float2(get_image_dim(src));
22461 float2 dst_cen = dst_dim / 2.0f;
22462 float2 src_cen = src_dim / 2.0f;
22464 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22466 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22468 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22469 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22471 src_pos = src_pos * src_dim / dst_dim;
22473 float2 src_loc = src_pos + src_cen;
22475 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22476 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22477 write_imagef(dst, dst_loc, 0.5f);
22479 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22484 Blend two inputs together, with the amount of each input used varying
22485 with the index counter.
22487 __kernel void blend_images(__write_only image2d_t dst,
22488 unsigned int index,
22489 __read_only image2d_t src1,
22490 __read_only image2d_t src2)
22492 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22493 CLK_FILTER_LINEAR);
22495 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22497 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22498 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22499 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22501 float4 val1 = read_imagef(src1, sampler, src1_loc);
22502 float4 val2 = read_imagef(src2, sampler, src2_loc);
22504 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22510 @section roberts_opencl
22511 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22513 The filter accepts the following option:
22517 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22520 Set value which will be multiplied with filtered result.
22521 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22524 Set value which will be added to filtered result.
22525 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22528 @subsection Example
22532 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22534 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22538 @section sobel_opencl
22540 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22542 The filter accepts the following option:
22546 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22549 Set value which will be multiplied with filtered result.
22550 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22553 Set value which will be added to filtered result.
22554 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22557 @subsection Example
22561 Apply sobel operator with scale set to 2 and delta set to 10
22563 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22567 @section tonemap_opencl
22569 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22571 It accepts the following parameters:
22575 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22578 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22581 Apply desaturation for highlights that exceed this level of brightness. The
22582 higher the parameter, the more color information will be preserved. This
22583 setting helps prevent unnaturally blown-out colors for super-highlights, by
22584 (smoothly) turning into white instead. This makes images feel more natural,
22585 at the cost of reducing information about out-of-range colors.
22587 The default value is 0.5, and the algorithm here is a little different from
22588 the cpu version tonemap currently. A setting of 0.0 disables this option.
22591 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22592 is used to detect whether the scene has changed or not. If the distance between
22593 the current frame average brightness and the current running average exceeds
22594 a threshold value, we would re-calculate scene average and peak brightness.
22595 The default value is 0.2.
22598 Specify the output pixel format.
22600 Currently supported formats are:
22607 Set the output color range.
22609 Possible values are:
22615 Default is same as input.
22618 Set the output color primaries.
22620 Possible values are:
22626 Default is same as input.
22629 Set the output transfer characteristics.
22631 Possible values are:
22640 Set the output colorspace matrix.
22642 Possible value are:
22648 Default is same as input.
22652 @subsection Example
22656 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22658 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22662 @section unsharp_opencl
22664 Sharpen or blur the input video.
22666 It accepts the following parameters:
22669 @item luma_msize_x, lx
22670 Set the luma matrix horizontal size.
22671 Range is @code{[1, 23]} and default value is @code{5}.
22673 @item luma_msize_y, ly
22674 Set the luma matrix vertical size.
22675 Range is @code{[1, 23]} and default value is @code{5}.
22677 @item luma_amount, la
22678 Set the luma effect strength.
22679 Range is @code{[-10, 10]} and default value is @code{1.0}.
22681 Negative values will blur the input video, while positive values will
22682 sharpen it, a value of zero will disable the effect.
22684 @item chroma_msize_x, cx
22685 Set the chroma matrix horizontal size.
22686 Range is @code{[1, 23]} and default value is @code{5}.
22688 @item chroma_msize_y, cy
22689 Set the chroma matrix vertical size.
22690 Range is @code{[1, 23]} and default value is @code{5}.
22692 @item chroma_amount, ca
22693 Set the chroma effect strength.
22694 Range is @code{[-10, 10]} and default value is @code{0.0}.
22696 Negative values will blur the input video, while positive values will
22697 sharpen it, a value of zero will disable the effect.
22701 All parameters are optional and default to the equivalent of the
22702 string '5:5:1.0:5:5:0.0'.
22704 @subsection Examples
22708 Apply strong luma sharpen effect:
22710 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22714 Apply a strong blur of both luma and chroma parameters:
22716 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22720 @section xfade_opencl
22722 Cross fade two videos with custom transition effect by using OpenCL.
22724 It accepts the following options:
22728 Set one of possible transition effects.
22732 Select custom transition effect, the actual transition description
22733 will be picked from source and kernel options.
22745 Default transition is fade.
22749 OpenCL program source file for custom transition.
22752 Set name of kernel to use for custom transition from program source file.
22755 Set duration of video transition.
22758 Set time of start of transition relative to first video.
22761 The program source file must contain a kernel function with the given name,
22762 which will be run once for each plane of the output. Each run on a plane
22763 gets enqueued as a separate 2D global NDRange with one work-item for each
22764 pixel to be generated. The global ID offset for each work-item is therefore
22765 the coordinates of a pixel in the destination image.
22767 The kernel function needs to take the following arguments:
22770 Destination image, @var{__write_only image2d_t}.
22772 This image will become the output; the kernel should write all of it.
22775 First Source image, @var{__read_only image2d_t}.
22776 Second Source image, @var{__read_only image2d_t}.
22778 These are the most recent images on each input. The kernel may read from
22779 them to generate the output, but they can't be written to.
22782 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22789 Apply dots curtain transition effect:
22791 __kernel void blend_images(__write_only image2d_t dst,
22792 __read_only image2d_t src1,
22793 __read_only image2d_t src2,
22796 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22797 CLK_FILTER_LINEAR);
22798 int2 p = (int2)(get_global_id(0), get_global_id(1));
22799 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22800 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22803 float2 dots = (float2)(20.0, 20.0);
22804 float2 center = (float2)(0,0);
22807 float4 val1 = read_imagef(src1, sampler, p);
22808 float4 val2 = read_imagef(src2, sampler, p);
22809 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22811 write_imagef(dst, p, next ? val1 : val2);
22817 @c man end OPENCL VIDEO FILTERS
22819 @chapter VAAPI Video Filters
22820 @c man begin VAAPI VIDEO FILTERS
22822 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22824 To enable compilation of these filters you need to configure FFmpeg with
22825 @code{--enable-vaapi}.
22827 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}
22829 @section tonemap_vaapi
22831 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22832 It maps the dynamic range of HDR10 content to the SDR content.
22833 It currently only accepts HDR10 as input.
22835 It accepts the following parameters:
22839 Specify the output pixel format.
22841 Currently supported formats are:
22850 Set the output color primaries.
22852 Default is same as input.
22855 Set the output transfer characteristics.
22860 Set the output colorspace matrix.
22862 Default is same as input.
22866 @subsection Example
22870 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22872 tonemap_vaapi=format=p010:t=bt2020-10
22876 @c man end VAAPI VIDEO FILTERS
22878 @chapter Video Sources
22879 @c man begin VIDEO SOURCES
22881 Below is a description of the currently available video sources.
22885 Buffer video frames, and make them available to the filter chain.
22887 This source is mainly intended for a programmatic use, in particular
22888 through the interface defined in @file{libavfilter/buffersrc.h}.
22890 It accepts the following parameters:
22895 Specify the size (width and height) of the buffered video frames. For the
22896 syntax of this option, check the
22897 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22900 The input video width.
22903 The input video height.
22906 A string representing the pixel format of the buffered video frames.
22907 It may be a number corresponding to a pixel format, or a pixel format
22911 Specify the timebase assumed by the timestamps of the buffered frames.
22914 Specify the frame rate expected for the video stream.
22916 @item pixel_aspect, sar
22917 The sample (pixel) aspect ratio of the input video.
22920 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
22921 to the filtergraph description to specify swscale flags for automatically
22922 inserted scalers. See @ref{Filtergraph syntax}.
22924 @item hw_frames_ctx
22925 When using a hardware pixel format, this should be a reference to an
22926 AVHWFramesContext describing input frames.
22931 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
22934 will instruct the source to accept video frames with size 320x240 and
22935 with format "yuv410p", assuming 1/24 as the timestamps timebase and
22936 square pixels (1:1 sample aspect ratio).
22937 Since the pixel format with name "yuv410p" corresponds to the number 6
22938 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
22939 this example corresponds to:
22941 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
22944 Alternatively, the options can be specified as a flat string, but this
22945 syntax is deprecated:
22947 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
22951 Create a pattern generated by an elementary cellular automaton.
22953 The initial state of the cellular automaton can be defined through the
22954 @option{filename} and @option{pattern} options. If such options are
22955 not specified an initial state is created randomly.
22957 At each new frame a new row in the video is filled with the result of
22958 the cellular automaton next generation. The behavior when the whole
22959 frame is filled is defined by the @option{scroll} option.
22961 This source accepts the following options:
22965 Read the initial cellular automaton state, i.e. the starting row, from
22966 the specified file.
22967 In the file, each non-whitespace character is considered an alive
22968 cell, a newline will terminate the row, and further characters in the
22969 file will be ignored.
22972 Read the initial cellular automaton state, i.e. the starting row, from
22973 the specified string.
22975 Each non-whitespace character in the string is considered an alive
22976 cell, a newline will terminate the row, and further characters in the
22977 string will be ignored.
22980 Set the video rate, that is the number of frames generated per second.
22983 @item random_fill_ratio, ratio
22984 Set the random fill ratio for the initial cellular automaton row. It
22985 is a floating point number value ranging from 0 to 1, defaults to
22988 This option is ignored when a file or a pattern is specified.
22990 @item random_seed, seed
22991 Set the seed for filling randomly the initial row, must be an integer
22992 included between 0 and UINT32_MAX. If not specified, or if explicitly
22993 set to -1, the filter will try to use a good random seed on a best
22997 Set the cellular automaton rule, it is a number ranging from 0 to 255.
22998 Default value is 110.
23001 Set the size of the output video. For the syntax of this option, check the
23002 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23004 If @option{filename} or @option{pattern} is specified, the size is set
23005 by default to the width of the specified initial state row, and the
23006 height is set to @var{width} * PHI.
23008 If @option{size} is set, it must contain the width of the specified
23009 pattern string, and the specified pattern will be centered in the
23012 If a filename or a pattern string is not specified, the size value
23013 defaults to "320x518" (used for a randomly generated initial state).
23016 If set to 1, scroll the output upward when all the rows in the output
23017 have been already filled. If set to 0, the new generated row will be
23018 written over the top row just after the bottom row is filled.
23021 @item start_full, full
23022 If set to 1, completely fill the output with generated rows before
23023 outputting the first frame.
23024 This is the default behavior, for disabling set the value to 0.
23027 If set to 1, stitch the left and right row edges together.
23028 This is the default behavior, for disabling set the value to 0.
23031 @subsection Examples
23035 Read the initial state from @file{pattern}, and specify an output of
23038 cellauto=f=pattern:s=200x400
23042 Generate a random initial row with a width of 200 cells, with a fill
23045 cellauto=ratio=2/3:s=200x200
23049 Create a pattern generated by rule 18 starting by a single alive cell
23050 centered on an initial row with width 100:
23052 cellauto=p=@@:s=100x400:full=0:rule=18
23056 Specify a more elaborated initial pattern:
23058 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23063 @anchor{coreimagesrc}
23064 @section coreimagesrc
23065 Video source generated on GPU using Apple's CoreImage API on OSX.
23067 This video source is a specialized version of the @ref{coreimage} video filter.
23068 Use a core image generator at the beginning of the applied filterchain to
23069 generate the content.
23071 The coreimagesrc video source accepts the following options:
23073 @item list_generators
23074 List all available generators along with all their respective options as well as
23075 possible minimum and maximum values along with the default values.
23077 list_generators=true
23081 Specify the size of the sourced video. For the syntax of this option, check the
23082 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23083 The default value is @code{320x240}.
23086 Specify the frame rate of the sourced video, as the number of frames
23087 generated per second. It has to be a string in the format
23088 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23089 number or a valid video frame rate abbreviation. The default value is
23093 Set the sample aspect ratio of the sourced video.
23096 Set the duration of the sourced video. See
23097 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23098 for the accepted syntax.
23100 If not specified, or the expressed duration is negative, the video is
23101 supposed to be generated forever.
23104 Additionally, all options of the @ref{coreimage} video filter are accepted.
23105 A complete filterchain can be used for further processing of the
23106 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23107 and examples for details.
23109 @subsection Examples
23114 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23115 given as complete and escaped command-line for Apple's standard bash shell:
23117 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23119 This example is equivalent to the QRCode example of @ref{coreimage} without the
23120 need for a nullsrc video source.
23125 Generate several gradients.
23129 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23130 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23133 Set frame rate, expressed as number of frames per second. Default
23136 @item c0, c1, c2, c3, c4, c5, c6, c7
23137 Set 8 colors. Default values for colors is to pick random one.
23139 @item x0, y0, y0, y1
23140 Set gradient line source and destination points. If negative or out of range, random ones
23144 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23147 Set seed for picking gradient line points.
23150 Set the duration of the sourced video. See
23151 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23152 for the accepted syntax.
23154 If not specified, or the expressed duration is negative, the video is
23155 supposed to be generated forever.
23158 Set speed of gradients rotation.
23162 @section mandelbrot
23164 Generate a Mandelbrot set fractal, and progressively zoom towards the
23165 point specified with @var{start_x} and @var{start_y}.
23167 This source accepts the following options:
23172 Set the terminal pts value. Default value is 400.
23175 Set the terminal scale value.
23176 Must be a floating point value. Default value is 0.3.
23179 Set the inner coloring mode, that is the algorithm used to draw the
23180 Mandelbrot fractal internal region.
23182 It shall assume one of the following values:
23187 Show time until convergence.
23189 Set color based on point closest to the origin of the iterations.
23194 Default value is @var{mincol}.
23197 Set the bailout value. Default value is 10.0.
23200 Set the maximum of iterations performed by the rendering
23201 algorithm. Default value is 7189.
23204 Set outer coloring mode.
23205 It shall assume one of following values:
23207 @item iteration_count
23208 Set iteration count mode.
23209 @item normalized_iteration_count
23210 set normalized iteration count mode.
23212 Default value is @var{normalized_iteration_count}.
23215 Set frame rate, expressed as number of frames per second. Default
23219 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23220 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23223 Set the initial scale value. Default value is 3.0.
23226 Set the initial x position. Must be a floating point value between
23227 -100 and 100. Default value is -0.743643887037158704752191506114774.
23230 Set the initial y position. Must be a floating point value between
23231 -100 and 100. Default value is -0.131825904205311970493132056385139.
23236 Generate various test patterns, as generated by the MPlayer test filter.
23238 The size of the generated video is fixed, and is 256x256.
23239 This source is useful in particular for testing encoding features.
23241 This source accepts the following options:
23246 Specify the frame rate of the sourced video, as the number of frames
23247 generated per second. It has to be a string in the format
23248 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23249 number or a valid video frame rate abbreviation. The default value is
23253 Set the duration of the sourced video. See
23254 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23255 for the accepted syntax.
23257 If not specified, or the expressed duration is negative, the video is
23258 supposed to be generated forever.
23262 Set the number or the name of the test to perform. Supported tests are:
23276 @item max_frames, m
23277 Set the maximum number of frames generated for each test, default value is 30.
23281 Default value is "all", which will cycle through the list of all tests.
23286 mptestsrc=t=dc_luma
23289 will generate a "dc_luma" test pattern.
23291 @section frei0r_src
23293 Provide a frei0r source.
23295 To enable compilation of this filter you need to install the frei0r
23296 header and configure FFmpeg with @code{--enable-frei0r}.
23298 This source accepts the following parameters:
23303 The size of the video to generate. For the syntax of this option, check the
23304 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23307 The framerate of the generated video. It may be a string of the form
23308 @var{num}/@var{den} or a frame rate abbreviation.
23311 The name to the frei0r source to load. For more information regarding frei0r and
23312 how to set the parameters, read the @ref{frei0r} section in the video filters
23315 @item filter_params
23316 A '|'-separated list of parameters to pass to the frei0r source.
23320 For example, to generate a frei0r partik0l source with size 200x200
23321 and frame rate 10 which is overlaid on the overlay filter main input:
23323 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23328 Generate a life pattern.
23330 This source is based on a generalization of John Conway's life game.
23332 The sourced input represents a life grid, each pixel represents a cell
23333 which can be in one of two possible states, alive or dead. Every cell
23334 interacts with its eight neighbours, which are the cells that are
23335 horizontally, vertically, or diagonally adjacent.
23337 At each interaction the grid evolves according to the adopted rule,
23338 which specifies the number of neighbor alive cells which will make a
23339 cell stay alive or born. The @option{rule} option allows one to specify
23342 This source accepts the following options:
23346 Set the file from which to read the initial grid state. In the file,
23347 each non-whitespace character is considered an alive cell, and newline
23348 is used to delimit the end of each row.
23350 If this option is not specified, the initial grid is generated
23354 Set the video rate, that is the number of frames generated per second.
23357 @item random_fill_ratio, ratio
23358 Set the random fill ratio for the initial random grid. It is a
23359 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23360 It is ignored when a file is specified.
23362 @item random_seed, seed
23363 Set the seed for filling the initial random grid, must be an integer
23364 included between 0 and UINT32_MAX. If not specified, or if explicitly
23365 set to -1, the filter will try to use a good random seed on a best
23371 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23372 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23373 @var{NS} specifies the number of alive neighbor cells which make a
23374 live cell stay alive, and @var{NB} the number of alive neighbor cells
23375 which make a dead cell to become alive (i.e. to "born").
23376 "s" and "b" can be used in place of "S" and "B", respectively.
23378 Alternatively a rule can be specified by an 18-bits integer. The 9
23379 high order bits are used to encode the next cell state if it is alive
23380 for each number of neighbor alive cells, the low order bits specify
23381 the rule for "borning" new cells. Higher order bits encode for an
23382 higher number of neighbor cells.
23383 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23384 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23386 Default value is "S23/B3", which is the original Conway's game of life
23387 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23388 cells, and will born a new cell if there are three alive cells around
23392 Set the size of the output video. For the syntax of this option, check the
23393 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23395 If @option{filename} is specified, the size is set by default to the
23396 same size of the input file. If @option{size} is set, it must contain
23397 the size specified in the input file, and the initial grid defined in
23398 that file is centered in the larger resulting area.
23400 If a filename is not specified, the size value defaults to "320x240"
23401 (used for a randomly generated initial grid).
23404 If set to 1, stitch the left and right grid edges together, and the
23405 top and bottom edges also. Defaults to 1.
23408 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23409 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23410 value from 0 to 255.
23413 Set the color of living (or new born) cells.
23416 Set the color of dead cells. If @option{mold} is set, this is the first color
23417 used to represent a dead cell.
23420 Set mold color, for definitely dead and moldy cells.
23422 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23423 ffmpeg-utils manual,ffmpeg-utils}.
23426 @subsection Examples
23430 Read a grid from @file{pattern}, and center it on a grid of size
23433 life=f=pattern:s=300x300
23437 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23439 life=ratio=2/3:s=200x200
23443 Specify a custom rule for evolving a randomly generated grid:
23449 Full example with slow death effect (mold) using @command{ffplay}:
23451 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23458 @anchor{haldclutsrc}
23461 @anchor{pal100bars}
23462 @anchor{rgbtestsrc}
23464 @anchor{smptehdbars}
23467 @anchor{yuvtestsrc}
23468 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23470 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23472 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23474 The @code{color} source provides an uniformly colored input.
23476 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23477 @ref{haldclut} filter.
23479 The @code{nullsrc} source returns unprocessed video frames. It is
23480 mainly useful to be employed in analysis / debugging tools, or as the
23481 source for filters which ignore the input data.
23483 The @code{pal75bars} source generates a color bars pattern, based on
23484 EBU PAL recommendations with 75% color levels.
23486 The @code{pal100bars} source generates a color bars pattern, based on
23487 EBU PAL recommendations with 100% color levels.
23489 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23490 detecting RGB vs BGR issues. You should see a red, green and blue
23491 stripe from top to bottom.
23493 The @code{smptebars} source generates a color bars pattern, based on
23494 the SMPTE Engineering Guideline EG 1-1990.
23496 The @code{smptehdbars} source generates a color bars pattern, based on
23497 the SMPTE RP 219-2002.
23499 The @code{testsrc} source generates a test video pattern, showing a
23500 color pattern, a scrolling gradient and a timestamp. This is mainly
23501 intended for testing purposes.
23503 The @code{testsrc2} source is similar to testsrc, but supports more
23504 pixel formats instead of just @code{rgb24}. This allows using it as an
23505 input for other tests without requiring a format conversion.
23507 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23508 see a y, cb and cr stripe from top to bottom.
23510 The sources accept the following parameters:
23515 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23516 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23517 pixels to be used as identity matrix for 3D lookup tables. Each component is
23518 coded on a @code{1/(N*N)} scale.
23521 Specify the color of the source, only available in the @code{color}
23522 source. For the syntax of this option, check the
23523 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23526 Specify the size of the sourced video. For the syntax of this option, check the
23527 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23528 The default value is @code{320x240}.
23530 This option is not available with the @code{allrgb}, @code{allyuv}, and
23531 @code{haldclutsrc} filters.
23534 Specify the frame rate of the sourced video, as the number of frames
23535 generated per second. It has to be a string in the format
23536 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23537 number or a valid video frame rate abbreviation. The default value is
23541 Set the duration of the sourced video. See
23542 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23543 for the accepted syntax.
23545 If not specified, or the expressed duration is negative, the video is
23546 supposed to be generated forever.
23548 Since the frame rate is used as time base, all frames including the last one
23549 will have their full duration. If the specified duration is not a multiple
23550 of the frame duration, it will be rounded up.
23553 Set the sample aspect ratio of the sourced video.
23556 Specify the alpha (opacity) of the background, only available in the
23557 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23558 255 (fully opaque, the default).
23561 Set the number of decimals to show in the timestamp, only available in the
23562 @code{testsrc} source.
23564 The displayed timestamp value will correspond to the original
23565 timestamp value multiplied by the power of 10 of the specified
23566 value. Default value is 0.
23569 @subsection Examples
23573 Generate a video with a duration of 5.3 seconds, with size
23574 176x144 and a frame rate of 10 frames per second:
23576 testsrc=duration=5.3:size=qcif:rate=10
23580 The following graph description will generate a red source
23581 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23584 color=c=red@@0.2:s=qcif:r=10
23588 If the input content is to be ignored, @code{nullsrc} can be used. The
23589 following command generates noise in the luminance plane by employing
23590 the @code{geq} filter:
23592 nullsrc=s=256x256, geq=random(1)*255:128:128
23596 @subsection Commands
23598 The @code{color} source supports the following commands:
23602 Set the color of the created image. Accepts the same syntax of the
23603 corresponding @option{color} option.
23608 Generate video using an OpenCL program.
23613 OpenCL program source file.
23616 Kernel name in program.
23619 Size of frames to generate. This must be set.
23622 Pixel format to use for the generated frames. This must be set.
23625 Number of frames generated every second. Default value is '25'.
23629 For details of how the program loading works, see the @ref{program_opencl}
23636 Generate a colour ramp by setting pixel values from the position of the pixel
23637 in the output image. (Note that this will work with all pixel formats, but
23638 the generated output will not be the same.)
23640 __kernel void ramp(__write_only image2d_t dst,
23641 unsigned int index)
23643 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23646 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23648 write_imagef(dst, loc, val);
23653 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23655 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23656 unsigned int index)
23658 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23660 float4 value = 0.0f;
23661 int x = loc.x + index;
23662 int y = loc.y + index;
23663 while (x > 0 || y > 0) {
23664 if (x % 3 == 1 && y % 3 == 1) {
23672 write_imagef(dst, loc, value);
23678 @section sierpinski
23680 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23682 This source accepts the following options:
23686 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23687 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23690 Set frame rate, expressed as number of frames per second. Default
23694 Set seed which is used for random panning.
23697 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23700 Set fractal type, can be default @code{carpet} or @code{triangle}.
23703 @c man end VIDEO SOURCES
23705 @chapter Video Sinks
23706 @c man begin VIDEO SINKS
23708 Below is a description of the currently available video sinks.
23710 @section buffersink
23712 Buffer video frames, and make them available to the end of the filter
23715 This sink is mainly intended for programmatic use, in particular
23716 through the interface defined in @file{libavfilter/buffersink.h}
23717 or the options system.
23719 It accepts a pointer to an AVBufferSinkContext structure, which
23720 defines the incoming buffers' formats, to be passed as the opaque
23721 parameter to @code{avfilter_init_filter} for initialization.
23725 Null video sink: do absolutely nothing with the input video. It is
23726 mainly useful as a template and for use in analysis / debugging
23729 @c man end VIDEO SINKS
23731 @chapter Multimedia Filters
23732 @c man begin MULTIMEDIA FILTERS
23734 Below is a description of the currently available multimedia filters.
23738 Convert input audio to a video output, displaying the audio bit scope.
23740 The filter accepts the following options:
23744 Set frame rate, expressed as number of frames per second. Default
23748 Specify the video size for the output. For the syntax of this option, check the
23749 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23750 Default value is @code{1024x256}.
23753 Specify list of colors separated by space or by '|' which will be used to
23754 draw channels. Unrecognized or missing colors will be replaced
23758 @section adrawgraph
23759 Draw a graph using input audio metadata.
23761 See @ref{drawgraph}
23763 @section agraphmonitor
23765 See @ref{graphmonitor}.
23767 @section ahistogram
23769 Convert input audio to a video output, displaying the volume histogram.
23771 The filter accepts the following options:
23775 Specify how histogram is calculated.
23777 It accepts the following values:
23780 Use single histogram for all channels.
23782 Use separate histogram for each channel.
23784 Default is @code{single}.
23787 Set frame rate, expressed as number of frames per second. Default
23791 Specify the video size for the output. For the syntax of this option, check the
23792 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23793 Default value is @code{hd720}.
23798 It accepts the following values:
23809 reverse logarithmic
23811 Default is @code{log}.
23814 Set amplitude scale.
23816 It accepts the following values:
23823 Default is @code{log}.
23826 Set how much frames to accumulate in histogram.
23827 Default is 1. Setting this to -1 accumulates all frames.
23830 Set histogram ratio of window height.
23833 Set sonogram sliding.
23835 It accepts the following values:
23838 replace old rows with new ones.
23840 scroll from top to bottom.
23842 Default is @code{replace}.
23845 @section aphasemeter
23847 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23848 representing mean phase of current audio frame. A video output can also be produced and is
23849 enabled by default. The audio is passed through as first output.
23851 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23852 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23853 and @code{1} means channels are in phase.
23855 The filter accepts the following options, all related to its video output:
23859 Set the output frame rate. Default value is @code{25}.
23862 Set the video size for the output. For the syntax of this option, check the
23863 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23864 Default value is @code{800x400}.
23869 Specify the red, green, blue contrast. Default values are @code{2},
23870 @code{7} and @code{1}.
23871 Allowed range is @code{[0, 255]}.
23874 Set color which will be used for drawing median phase. If color is
23875 @code{none} which is default, no median phase value will be drawn.
23878 Enable video output. Default is enabled.
23881 @subsection phasing detection
23883 The filter also detects out of phase and mono sequences in stereo streams.
23884 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
23886 The filter accepts the following options for this detection:
23890 Enable mono and out of phase detection. Default is disabled.
23893 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
23894 Allowed range is @code{[0, 1]}.
23897 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
23898 Allowed range is @code{[90, 180]}.
23901 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
23904 @subsection Examples
23908 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
23910 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
23914 @section avectorscope
23916 Convert input audio to a video output, representing the audio vector
23919 The filter is used to measure the difference between channels of stereo
23920 audio stream. A monaural signal, consisting of identical left and right
23921 signal, results in straight vertical line. Any stereo separation is visible
23922 as a deviation from this line, creating a Lissajous figure.
23923 If the straight (or deviation from it) but horizontal line appears this
23924 indicates that the left and right channels are out of phase.
23926 The filter accepts the following options:
23930 Set the vectorscope mode.
23932 Available values are:
23935 Lissajous rotated by 45 degrees.
23938 Same as above but not rotated.
23941 Shape resembling half of circle.
23944 Default value is @samp{lissajous}.
23947 Set the video size for the output. For the syntax of this option, check the
23948 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23949 Default value is @code{400x400}.
23952 Set the output frame rate. Default value is @code{25}.
23958 Specify the red, green, blue and alpha contrast. Default values are @code{40},
23959 @code{160}, @code{80} and @code{255}.
23960 Allowed range is @code{[0, 255]}.
23966 Specify the red, green, blue and alpha fade. Default values are @code{15},
23967 @code{10}, @code{5} and @code{5}.
23968 Allowed range is @code{[0, 255]}.
23971 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
23972 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
23975 Set the vectorscope drawing mode.
23977 Available values are:
23980 Draw dot for each sample.
23983 Draw line between previous and current sample.
23986 Default value is @samp{dot}.
23989 Specify amplitude scale of audio samples.
23991 Available values are:
24007 Swap left channel axis with right channel axis.
24017 Mirror only x axis.
24020 Mirror only y axis.
24028 @subsection Examples
24032 Complete example using @command{ffplay}:
24034 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24035 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24039 @section bench, abench
24041 Benchmark part of a filtergraph.
24043 The filter accepts the following options:
24047 Start or stop a timer.
24049 Available values are:
24052 Get the current time, set it as frame metadata (using the key
24053 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24056 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24057 the input frame metadata to get the time difference. Time difference, average,
24058 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24059 @code{min}) are then printed. The timestamps are expressed in seconds.
24063 @subsection Examples
24067 Benchmark @ref{selectivecolor} filter:
24069 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24075 Concatenate audio and video streams, joining them together one after the
24078 The filter works on segments of synchronized video and audio streams. All
24079 segments must have the same number of streams of each type, and that will
24080 also be the number of streams at output.
24082 The filter accepts the following options:
24087 Set the number of segments. Default is 2.
24090 Set the number of output video streams, that is also the number of video
24091 streams in each segment. Default is 1.
24094 Set the number of output audio streams, that is also the number of audio
24095 streams in each segment. Default is 0.
24098 Activate unsafe mode: do not fail if segments have a different format.
24102 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24103 @var{a} audio outputs.
24105 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24106 segment, in the same order as the outputs, then the inputs for the second
24109 Related streams do not always have exactly the same duration, for various
24110 reasons including codec frame size or sloppy authoring. For that reason,
24111 related synchronized streams (e.g. a video and its audio track) should be
24112 concatenated at once. The concat filter will use the duration of the longest
24113 stream in each segment (except the last one), and if necessary pad shorter
24114 audio streams with silence.
24116 For this filter to work correctly, all segments must start at timestamp 0.
24118 All corresponding streams must have the same parameters in all segments; the
24119 filtering system will automatically select a common pixel format for video
24120 streams, and a common sample format, sample rate and channel layout for
24121 audio streams, but other settings, such as resolution, must be converted
24122 explicitly by the user.
24124 Different frame rates are acceptable but will result in variable frame rate
24125 at output; be sure to configure the output file to handle it.
24127 @subsection Examples
24131 Concatenate an opening, an episode and an ending, all in bilingual version
24132 (video in stream 0, audio in streams 1 and 2):
24134 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24135 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24136 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24137 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24141 Concatenate two parts, handling audio and video separately, using the
24142 (a)movie sources, and adjusting the resolution:
24144 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24145 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24146 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24148 Note that a desync will happen at the stitch if the audio and video streams
24149 do not have exactly the same duration in the first file.
24153 @subsection Commands
24155 This filter supports the following commands:
24158 Close the current segment and step to the next one
24164 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24165 level. By default, it logs a message at a frequency of 10Hz with the
24166 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24167 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24169 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24170 sample format is double-precision floating point. The input stream will be converted to
24171 this specification, if needed. Users may need to insert aformat and/or aresample filters
24172 after this filter to obtain the original parameters.
24174 The filter also has a video output (see the @var{video} option) with a real
24175 time graph to observe the loudness evolution. The graphic contains the logged
24176 message mentioned above, so it is not printed anymore when this option is set,
24177 unless the verbose logging is set. The main graphing area contains the
24178 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24179 the momentary loudness (400 milliseconds), but can optionally be configured
24180 to instead display short-term loudness (see @var{gauge}).
24182 The green area marks a +/- 1LU target range around the target loudness
24183 (-23LUFS by default, unless modified through @var{target}).
24185 More information about the Loudness Recommendation EBU R128 on
24186 @url{http://tech.ebu.ch/loudness}.
24188 The filter accepts the following options:
24193 Activate the video output. The audio stream is passed unchanged whether this
24194 option is set or no. The video stream will be the first output stream if
24195 activated. Default is @code{0}.
24198 Set the video size. This option is for video only. For the syntax of this
24200 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24201 Default and minimum resolution is @code{640x480}.
24204 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24205 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24206 other integer value between this range is allowed.
24209 Set metadata injection. If set to @code{1}, the audio input will be segmented
24210 into 100ms output frames, each of them containing various loudness information
24211 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24213 Default is @code{0}.
24216 Force the frame logging level.
24218 Available values are:
24221 information logging level
24223 verbose logging level
24226 By default, the logging level is set to @var{info}. If the @option{video} or
24227 the @option{metadata} options are set, it switches to @var{verbose}.
24232 Available modes can be cumulated (the option is a @code{flag} type). Possible
24236 Disable any peak mode (default).
24238 Enable sample-peak mode.
24240 Simple peak mode looking for the higher sample value. It logs a message
24241 for sample-peak (identified by @code{SPK}).
24243 Enable true-peak mode.
24245 If enabled, the peak lookup is done on an over-sampled version of the input
24246 stream for better peak accuracy. It logs a message for true-peak.
24247 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24248 This mode requires a build with @code{libswresample}.
24252 Treat mono input files as "dual mono". If a mono file is intended for playback
24253 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24254 If set to @code{true}, this option will compensate for this effect.
24255 Multi-channel input files are not affected by this option.
24258 Set a specific pan law to be used for the measurement of dual mono files.
24259 This parameter is optional, and has a default value of -3.01dB.
24262 Set a specific target level (in LUFS) used as relative zero in the visualization.
24263 This parameter is optional and has a default value of -23LUFS as specified
24264 by EBU R128. However, material published online may prefer a level of -16LUFS
24265 (e.g. for use with podcasts or video platforms).
24268 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24269 @code{shortterm}. By default the momentary value will be used, but in certain
24270 scenarios it may be more useful to observe the short term value instead (e.g.
24274 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24275 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24276 video output, not the summary or continuous log output.
24279 @subsection Examples
24283 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24285 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24289 Run an analysis with @command{ffmpeg}:
24291 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24295 @section interleave, ainterleave
24297 Temporally interleave frames from several inputs.
24299 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24301 These filters read frames from several inputs and send the oldest
24302 queued frame to the output.
24304 Input streams must have well defined, monotonically increasing frame
24307 In order to submit one frame to output, these filters need to enqueue
24308 at least one frame for each input, so they cannot work in case one
24309 input is not yet terminated and will not receive incoming frames.
24311 For example consider the case when one input is a @code{select} filter
24312 which always drops input frames. The @code{interleave} filter will keep
24313 reading from that input, but it will never be able to send new frames
24314 to output until the input sends an end-of-stream signal.
24316 Also, depending on inputs synchronization, the filters will drop
24317 frames in case one input receives more frames than the other ones, and
24318 the queue is already filled.
24320 These filters accept the following options:
24324 Set the number of different inputs, it is 2 by default.
24327 How to determine the end-of-stream.
24331 The duration of the longest input. (default)
24334 The duration of the shortest input.
24337 The duration of the first input.
24342 @subsection Examples
24346 Interleave frames belonging to different streams using @command{ffmpeg}:
24348 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24352 Add flickering blur effect:
24354 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24358 @section metadata, ametadata
24360 Manipulate frame metadata.
24362 This filter accepts the following options:
24366 Set mode of operation of the filter.
24368 Can be one of the following:
24372 If both @code{value} and @code{key} is set, select frames
24373 which have such metadata. If only @code{key} is set, select
24374 every frame that has such key in metadata.
24377 Add new metadata @code{key} and @code{value}. If key is already available
24381 Modify value of already present key.
24384 If @code{value} is set, delete only keys that have such value.
24385 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24389 Print key and its value if metadata was found. If @code{key} is not set print all
24390 metadata values available in frame.
24394 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24397 Set metadata value which will be used. This option is mandatory for
24398 @code{modify} and @code{add} mode.
24401 Which function to use when comparing metadata value and @code{value}.
24403 Can be one of following:
24407 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24410 Values are interpreted as strings, returns true if metadata value starts with
24411 the @code{value} option string.
24414 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24417 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24420 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24423 Values are interpreted as floats, returns true if expression from option @code{expr}
24427 Values are interpreted as strings, returns true if metadata value ends with
24428 the @code{value} option string.
24432 Set expression which is used when @code{function} is set to @code{expr}.
24433 The expression is evaluated through the eval API and can contain the following
24438 Float representation of @code{value} from metadata key.
24441 Float representation of @code{value} as supplied by user in @code{value} option.
24445 If specified in @code{print} mode, output is written to the named file. Instead of
24446 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24447 for standard output. If @code{file} option is not set, output is written to the log
24448 with AV_LOG_INFO loglevel.
24451 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24455 @subsection Examples
24459 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24462 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24465 Print silencedetect output to file @file{metadata.txt}.
24467 silencedetect,ametadata=mode=print:file=metadata.txt
24470 Direct all metadata to a pipe with file descriptor 4.
24472 metadata=mode=print:file='pipe\:4'
24476 @section perms, aperms
24478 Set read/write permissions for the output frames.
24480 These filters are mainly aimed at developers to test direct path in the
24481 following filter in the filtergraph.
24483 The filters accept the following options:
24487 Select the permissions mode.
24489 It accepts the following values:
24492 Do nothing. This is the default.
24494 Set all the output frames read-only.
24496 Set all the output frames directly writable.
24498 Make the frame read-only if writable, and writable if read-only.
24500 Set each output frame read-only or writable randomly.
24504 Set the seed for the @var{random} mode, must be an integer included between
24505 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24506 @code{-1}, the filter will try to use a good random seed on a best effort
24510 Note: in case of auto-inserted filter between the permission filter and the
24511 following one, the permission might not be received as expected in that
24512 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24513 perms/aperms filter can avoid this problem.
24515 @section realtime, arealtime
24517 Slow down filtering to match real time approximately.
24519 These filters will pause the filtering for a variable amount of time to
24520 match the output rate with the input timestamps.
24521 They are similar to the @option{re} option to @code{ffmpeg}.
24523 They accept the following options:
24527 Time limit for the pauses. Any pause longer than that will be considered
24528 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24530 Speed factor for processing. The value must be a float larger than zero.
24531 Values larger than 1.0 will result in faster than realtime processing,
24532 smaller will slow processing down. The @var{limit} is automatically adapted
24533 accordingly. Default is 1.0.
24535 A processing speed faster than what is possible without these filters cannot
24540 @section select, aselect
24542 Select frames to pass in output.
24544 This filter accepts the following options:
24549 Set expression, which is evaluated for each input frame.
24551 If the expression is evaluated to zero, the frame is discarded.
24553 If the evaluation result is negative or NaN, the frame is sent to the
24554 first output; otherwise it is sent to the output with index
24555 @code{ceil(val)-1}, assuming that the input index starts from 0.
24557 For example a value of @code{1.2} corresponds to the output with index
24558 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24561 Set the number of outputs. The output to which to send the selected
24562 frame is based on the result of the evaluation. Default value is 1.
24565 The expression can contain the following constants:
24569 The (sequential) number of the filtered frame, starting from 0.
24572 The (sequential) number of the selected frame, starting from 0.
24574 @item prev_selected_n
24575 The sequential number of the last selected frame. It's NAN if undefined.
24578 The timebase of the input timestamps.
24581 The PTS (Presentation TimeStamp) of the filtered video frame,
24582 expressed in @var{TB} units. It's NAN if undefined.
24585 The PTS of the filtered video frame,
24586 expressed in seconds. It's NAN if undefined.
24589 The PTS of the previously filtered video frame. It's NAN if undefined.
24591 @item prev_selected_pts
24592 The PTS of the last previously filtered video frame. It's NAN if undefined.
24594 @item prev_selected_t
24595 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24598 The PTS of the first video frame in the video. It's NAN if undefined.
24601 The time of the first video frame in the video. It's NAN if undefined.
24603 @item pict_type @emph{(video only)}
24604 The type of the filtered frame. It can assume one of the following
24616 @item interlace_type @emph{(video only)}
24617 The frame interlace type. It can assume one of the following values:
24620 The frame is progressive (not interlaced).
24622 The frame is top-field-first.
24624 The frame is bottom-field-first.
24627 @item consumed_sample_n @emph{(audio only)}
24628 the number of selected samples before the current frame
24630 @item samples_n @emph{(audio only)}
24631 the number of samples in the current frame
24633 @item sample_rate @emph{(audio only)}
24634 the input sample rate
24637 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24640 the position in the file of the filtered frame, -1 if the information
24641 is not available (e.g. for synthetic video)
24643 @item scene @emph{(video only)}
24644 value between 0 and 1 to indicate a new scene; a low value reflects a low
24645 probability for the current frame to introduce a new scene, while a higher
24646 value means the current frame is more likely to be one (see the example below)
24648 @item concatdec_select
24649 The concat demuxer can select only part of a concat input file by setting an
24650 inpoint and an outpoint, but the output packets may not be entirely contained
24651 in the selected interval. By using this variable, it is possible to skip frames
24652 generated by the concat demuxer which are not exactly contained in the selected
24655 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24656 and the @var{lavf.concat.duration} packet metadata values which are also
24657 present in the decoded frames.
24659 The @var{concatdec_select} variable is -1 if the frame pts is at least
24660 start_time and either the duration metadata is missing or the frame pts is less
24661 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24664 That basically means that an input frame is selected if its pts is within the
24665 interval set by the concat demuxer.
24669 The default value of the select expression is "1".
24671 @subsection Examples
24675 Select all frames in input:
24680 The example above is the same as:
24692 Select only I-frames:
24694 select='eq(pict_type\,I)'
24698 Select one frame every 100:
24700 select='not(mod(n\,100))'
24704 Select only frames contained in the 10-20 time interval:
24706 select=between(t\,10\,20)
24710 Select only I-frames contained in the 10-20 time interval:
24712 select=between(t\,10\,20)*eq(pict_type\,I)
24716 Select frames with a minimum distance of 10 seconds:
24718 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24722 Use aselect to select only audio frames with samples number > 100:
24724 aselect='gt(samples_n\,100)'
24728 Create a mosaic of the first scenes:
24730 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24733 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24737 Send even and odd frames to separate outputs, and compose them:
24739 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24743 Select useful frames from an ffconcat file which is using inpoints and
24744 outpoints but where the source files are not intra frame only.
24746 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24750 @section sendcmd, asendcmd
24752 Send commands to filters in the filtergraph.
24754 These filters read commands to be sent to other filters in the
24757 @code{sendcmd} must be inserted between two video filters,
24758 @code{asendcmd} must be inserted between two audio filters, but apart
24759 from that they act the same way.
24761 The specification of commands can be provided in the filter arguments
24762 with the @var{commands} option, or in a file specified by the
24763 @var{filename} option.
24765 These filters accept the following options:
24768 Set the commands to be read and sent to the other filters.
24770 Set the filename of the commands to be read and sent to the other
24774 @subsection Commands syntax
24776 A commands description consists of a sequence of interval
24777 specifications, comprising a list of commands to be executed when a
24778 particular event related to that interval occurs. The occurring event
24779 is typically the current frame time entering or leaving a given time
24782 An interval is specified by the following syntax:
24784 @var{START}[-@var{END}] @var{COMMANDS};
24787 The time interval is specified by the @var{START} and @var{END} times.
24788 @var{END} is optional and defaults to the maximum time.
24790 The current frame time is considered within the specified interval if
24791 it is included in the interval [@var{START}, @var{END}), that is when
24792 the time is greater or equal to @var{START} and is lesser than
24795 @var{COMMANDS} consists of a sequence of one or more command
24796 specifications, separated by ",", relating to that interval. The
24797 syntax of a command specification is given by:
24799 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24802 @var{FLAGS} is optional and specifies the type of events relating to
24803 the time interval which enable sending the specified command, and must
24804 be a non-null sequence of identifier flags separated by "+" or "|" and
24805 enclosed between "[" and "]".
24807 The following flags are recognized:
24810 The command is sent when the current frame timestamp enters the
24811 specified interval. In other words, the command is sent when the
24812 previous frame timestamp was not in the given interval, and the
24816 The command is sent when the current frame timestamp leaves the
24817 specified interval. In other words, the command is sent when the
24818 previous frame timestamp was in the given interval, and the
24822 The command @var{ARG} is interpreted as expression and result of
24823 expression is passed as @var{ARG}.
24825 The expression is evaluated through the eval API and can contain the following
24830 Original position in the file of the frame, or undefined if undefined
24831 for the current frame.
24834 The presentation timestamp in input.
24837 The count of the input frame for video or audio, starting from 0.
24840 The time in seconds of the current frame.
24843 The start time in seconds of the current command interval.
24846 The end time in seconds of the current command interval.
24849 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24854 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24857 @var{TARGET} specifies the target of the command, usually the name of
24858 the filter class or a specific filter instance name.
24860 @var{COMMAND} specifies the name of the command for the target filter.
24862 @var{ARG} is optional and specifies the optional list of argument for
24863 the given @var{COMMAND}.
24865 Between one interval specification and another, whitespaces, or
24866 sequences of characters starting with @code{#} until the end of line,
24867 are ignored and can be used to annotate comments.
24869 A simplified BNF description of the commands specification syntax
24872 @var{COMMAND_FLAG} ::= "enter" | "leave"
24873 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24874 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24875 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24876 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24877 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24880 @subsection Examples
24884 Specify audio tempo change at second 4:
24886 asendcmd=c='4.0 atempo tempo 1.5',atempo
24890 Target a specific filter instance:
24892 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
24896 Specify a list of drawtext and hue commands in a file.
24898 # show text in the interval 5-10
24899 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
24900 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
24902 # desaturate the image in the interval 15-20
24903 15.0-20.0 [enter] hue s 0,
24904 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
24906 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
24908 # apply an exponential saturation fade-out effect, starting from time 25
24909 25 [enter] hue s exp(25-t)
24912 A filtergraph allowing to read and process the above command list
24913 stored in a file @file{test.cmd}, can be specified with:
24915 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
24920 @section setpts, asetpts
24922 Change the PTS (presentation timestamp) of the input frames.
24924 @code{setpts} works on video frames, @code{asetpts} on audio frames.
24926 This filter accepts the following options:
24931 The expression which is evaluated for each frame to construct its timestamp.
24935 The expression is evaluated through the eval API and can contain the following
24939 @item FRAME_RATE, FR
24940 frame rate, only defined for constant frame-rate video
24943 The presentation timestamp in input
24946 The count of the input frame for video or the number of consumed samples,
24947 not including the current frame for audio, starting from 0.
24949 @item NB_CONSUMED_SAMPLES
24950 The number of consumed samples, not including the current frame (only
24953 @item NB_SAMPLES, S
24954 The number of samples in the current frame (only audio)
24956 @item SAMPLE_RATE, SR
24957 The audio sample rate.
24960 The PTS of the first frame.
24963 the time in seconds of the first frame
24966 State whether the current frame is interlaced.
24969 the time in seconds of the current frame
24972 original position in the file of the frame, or undefined if undefined
24973 for the current frame
24976 The previous input PTS.
24979 previous input time in seconds
24982 The previous output PTS.
24985 previous output time in seconds
24988 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
24992 The wallclock (RTC) time at the start of the movie in microseconds.
24995 The timebase of the input timestamps.
24999 @subsection Examples
25003 Start counting PTS from zero
25005 setpts=PTS-STARTPTS
25009 Apply fast motion effect:
25015 Apply slow motion effect:
25021 Set fixed rate of 25 frames per second:
25027 Set fixed rate 25 fps with some jitter:
25029 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25033 Apply an offset of 10 seconds to the input PTS:
25039 Generate timestamps from a "live source" and rebase onto the current timebase:
25041 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25045 Generate timestamps by counting samples:
25054 Force color range for the output video frame.
25056 The @code{setrange} filter marks the color range property for the
25057 output frames. It does not change the input frame, but only sets the
25058 corresponding property, which affects how the frame is treated by
25061 The filter accepts the following options:
25066 Available values are:
25070 Keep the same color range property.
25072 @item unspecified, unknown
25073 Set the color range as unspecified.
25075 @item limited, tv, mpeg
25076 Set the color range as limited.
25078 @item full, pc, jpeg
25079 Set the color range as full.
25083 @section settb, asettb
25085 Set the timebase to use for the output frames timestamps.
25086 It is mainly useful for testing timebase configuration.
25088 It accepts the following parameters:
25093 The expression which is evaluated into the output timebase.
25097 The value for @option{tb} is an arithmetic expression representing a
25098 rational. The expression can contain the constants "AVTB" (the default
25099 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25100 audio only). Default value is "intb".
25102 @subsection Examples
25106 Set the timebase to 1/25:
25112 Set the timebase to 1/10:
25118 Set the timebase to 1001/1000:
25124 Set the timebase to 2*intb:
25130 Set the default timebase value:
25137 Convert input audio to a video output representing frequency spectrum
25138 logarithmically using Brown-Puckette constant Q transform algorithm with
25139 direct frequency domain coefficient calculation (but the transform itself
25140 is not really constant Q, instead the Q factor is actually variable/clamped),
25141 with musical tone scale, from E0 to D#10.
25143 The filter accepts the following options:
25147 Specify the video size for the output. It must be even. For the syntax of this option,
25148 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25149 Default value is @code{1920x1080}.
25152 Set the output frame rate. Default value is @code{25}.
25155 Set the bargraph height. It must be even. Default value is @code{-1} which
25156 computes the bargraph height automatically.
25159 Set the axis height. It must be even. Default value is @code{-1} which computes
25160 the axis height automatically.
25163 Set the sonogram height. It must be even. Default value is @code{-1} which
25164 computes the sonogram height automatically.
25167 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25168 instead. Default value is @code{1}.
25170 @item sono_v, volume
25171 Specify the sonogram volume expression. It can contain variables:
25174 the @var{bar_v} evaluated expression
25175 @item frequency, freq, f
25176 the frequency where it is evaluated
25177 @item timeclamp, tc
25178 the value of @var{timeclamp} option
25182 @item a_weighting(f)
25183 A-weighting of equal loudness
25184 @item b_weighting(f)
25185 B-weighting of equal loudness
25186 @item c_weighting(f)
25187 C-weighting of equal loudness.
25189 Default value is @code{16}.
25191 @item bar_v, volume2
25192 Specify the bargraph volume expression. It can contain variables:
25195 the @var{sono_v} evaluated expression
25196 @item frequency, freq, f
25197 the frequency where it is evaluated
25198 @item timeclamp, tc
25199 the value of @var{timeclamp} option
25203 @item a_weighting(f)
25204 A-weighting of equal loudness
25205 @item b_weighting(f)
25206 B-weighting of equal loudness
25207 @item c_weighting(f)
25208 C-weighting of equal loudness.
25210 Default value is @code{sono_v}.
25212 @item sono_g, gamma
25213 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25214 higher gamma makes the spectrum having more range. Default value is @code{3}.
25215 Acceptable range is @code{[1, 7]}.
25217 @item bar_g, gamma2
25218 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25222 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25223 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25225 @item timeclamp, tc
25226 Specify the transform timeclamp. At low frequency, there is trade-off between
25227 accuracy in time domain and frequency domain. If timeclamp is lower,
25228 event in time domain is represented more accurately (such as fast bass drum),
25229 otherwise event in frequency domain is represented more accurately
25230 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25233 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25234 limits future samples by applying asymmetric windowing in time domain, useful
25235 when low latency is required. Accepted range is @code{[0, 1]}.
25238 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25239 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25242 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25243 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25246 This option is deprecated and ignored.
25249 Specify the transform length in time domain. Use this option to control accuracy
25250 trade-off between time domain and frequency domain at every frequency sample.
25251 It can contain variables:
25253 @item frequency, freq, f
25254 the frequency where it is evaluated
25255 @item timeclamp, tc
25256 the value of @var{timeclamp} option.
25258 Default value is @code{384*tc/(384+tc*f)}.
25261 Specify the transform count for every video frame. Default value is @code{6}.
25262 Acceptable range is @code{[1, 30]}.
25265 Specify the transform count for every single pixel. Default value is @code{0},
25266 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25269 Specify font file for use with freetype to draw the axis. If not specified,
25270 use embedded font. Note that drawing with font file or embedded font is not
25271 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25275 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25276 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25280 Specify font color expression. This is arithmetic expression that should return
25281 integer value 0xRRGGBB. It can contain variables:
25283 @item frequency, freq, f
25284 the frequency where it is evaluated
25285 @item timeclamp, tc
25286 the value of @var{timeclamp} option
25291 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25292 @item r(x), g(x), b(x)
25293 red, green, and blue value of intensity x.
25295 Default value is @code{st(0, (midi(f)-59.5)/12);
25296 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25297 r(1-ld(1)) + b(ld(1))}.
25300 Specify image file to draw the axis. This option override @var{fontfile} and
25301 @var{fontcolor} option.
25304 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25305 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25306 Default value is @code{1}.
25309 Set colorspace. The accepted values are:
25312 Unspecified (default)
25321 BT.470BG or BT.601-6 625
25324 SMPTE-170M or BT.601-6 525
25330 BT.2020 with non-constant luminance
25335 Set spectrogram color scheme. This is list of floating point values with format
25336 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25337 The default is @code{1|0.5|0|0|0.5|1}.
25341 @subsection Examples
25345 Playing audio while showing the spectrum:
25347 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25351 Same as above, but with frame rate 30 fps:
25353 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25357 Playing at 1280x720:
25359 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25363 Disable sonogram display:
25369 A1 and its harmonics: A1, A2, (near)E3, A3:
25371 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),
25372 asplit[a][out1]; [a] showcqt [out0]'
25376 Same as above, but with more accuracy in frequency domain:
25378 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),
25379 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25385 bar_v=10:sono_v=bar_v*a_weighting(f)
25389 Custom gamma, now spectrum is linear to the amplitude.
25395 Custom tlength equation:
25397 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)))'
25401 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25403 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25407 Custom font using fontconfig:
25409 font='Courier New,Monospace,mono|bold'
25413 Custom frequency range with custom axis using image file:
25415 axisfile=myaxis.png:basefreq=40:endfreq=10000
25421 Convert input audio to video output representing the audio power spectrum.
25422 Audio amplitude is on Y-axis while frequency is on X-axis.
25424 The filter accepts the following options:
25428 Specify size of video. For the syntax of this option, check the
25429 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25430 Default is @code{1024x512}.
25434 This set how each frequency bin will be represented.
25436 It accepts the following values:
25442 Default is @code{bar}.
25445 Set amplitude scale.
25447 It accepts the following values:
25461 Default is @code{log}.
25464 Set frequency scale.
25466 It accepts the following values:
25475 Reverse logarithmic scale.
25477 Default is @code{lin}.
25480 Set window size. Allowed range is from 16 to 65536.
25482 Default is @code{2048}
25485 Set windowing function.
25487 It accepts the following values:
25510 Default is @code{hanning}.
25513 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25514 which means optimal overlap for selected window function will be picked.
25517 Set time averaging. Setting this to 0 will display current maximal peaks.
25518 Default is @code{1}, which means time averaging is disabled.
25521 Specify list of colors separated by space or by '|' which will be used to
25522 draw channel frequencies. Unrecognized or missing colors will be replaced
25526 Set channel display mode.
25528 It accepts the following values:
25533 Default is @code{combined}.
25536 Set minimum amplitude used in @code{log} amplitude scaler.
25539 Set data display mode.
25541 It accepts the following values:
25547 Default is @code{magnitude}.
25550 @section showspatial
25552 Convert stereo input audio to a video output, representing the spatial relationship
25553 between two channels.
25555 The filter accepts the following options:
25559 Specify the video size for the output. For the syntax of this option, check the
25560 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25561 Default value is @code{512x512}.
25564 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25567 Set window function.
25569 It accepts the following values:
25594 Default value is @code{hann}.
25597 Set ratio of overlap window. Default value is @code{0.5}.
25598 When value is @code{1} overlap is set to recommended size for specific
25599 window function currently used.
25602 @anchor{showspectrum}
25603 @section showspectrum
25605 Convert input audio to a video output, representing the audio frequency
25608 The filter accepts the following options:
25612 Specify the video size for the output. For the syntax of this option, check the
25613 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25614 Default value is @code{640x512}.
25617 Specify how the spectrum should slide along the window.
25619 It accepts the following values:
25622 the samples start again on the left when they reach the right
25624 the samples scroll from right to left
25626 frames are only produced when the samples reach the right
25628 the samples scroll from left to right
25631 Default value is @code{replace}.
25634 Specify display mode.
25636 It accepts the following values:
25639 all channels are displayed in the same row
25641 all channels are displayed in separate rows
25644 Default value is @samp{combined}.
25647 Specify display color mode.
25649 It accepts the following values:
25652 each channel is displayed in a separate color
25654 each channel is displayed using the same color scheme
25656 each channel is displayed using the rainbow color scheme
25658 each channel is displayed using the moreland color scheme
25660 each channel is displayed using the nebulae color scheme
25662 each channel is displayed using the fire color scheme
25664 each channel is displayed using the fiery color scheme
25666 each channel is displayed using the fruit color scheme
25668 each channel is displayed using the cool color scheme
25670 each channel is displayed using the magma color scheme
25672 each channel is displayed using the green color scheme
25674 each channel is displayed using the viridis color scheme
25676 each channel is displayed using the plasma color scheme
25678 each channel is displayed using the cividis color scheme
25680 each channel is displayed using the terrain color scheme
25683 Default value is @samp{channel}.
25686 Specify scale used for calculating intensity color values.
25688 It accepts the following values:
25693 square root, default
25704 Default value is @samp{sqrt}.
25707 Specify frequency scale.
25709 It accepts the following values:
25717 Default value is @samp{lin}.
25720 Set saturation modifier for displayed colors. Negative values provide
25721 alternative color scheme. @code{0} is no saturation at all.
25722 Saturation must be in [-10.0, 10.0] range.
25723 Default value is @code{1}.
25726 Set window function.
25728 It accepts the following values:
25753 Default value is @code{hann}.
25756 Set orientation of time vs frequency axis. Can be @code{vertical} or
25757 @code{horizontal}. Default is @code{vertical}.
25760 Set ratio of overlap window. Default value is @code{0}.
25761 When value is @code{1} overlap is set to recommended size for specific
25762 window function currently used.
25765 Set scale gain for calculating intensity color values.
25766 Default value is @code{1}.
25769 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25772 Set color rotation, must be in [-1.0, 1.0] range.
25773 Default value is @code{0}.
25776 Set start frequency from which to display spectrogram. Default is @code{0}.
25779 Set stop frequency to which to display spectrogram. Default is @code{0}.
25782 Set upper frame rate limit. Default is @code{auto}, unlimited.
25785 Draw time and frequency axes and legends. Default is disabled.
25788 The usage is very similar to the showwaves filter; see the examples in that
25791 @subsection Examples
25795 Large window with logarithmic color scaling:
25797 showspectrum=s=1280x480:scale=log
25801 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25803 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25804 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25808 @section showspectrumpic
25810 Convert input audio to a single video frame, representing the audio frequency
25813 The filter accepts the following options:
25817 Specify the video size for the output. For the syntax of this option, check the
25818 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25819 Default value is @code{4096x2048}.
25822 Specify display mode.
25824 It accepts the following values:
25827 all channels are displayed in the same row
25829 all channels are displayed in separate rows
25831 Default value is @samp{combined}.
25834 Specify display color mode.
25836 It accepts the following values:
25839 each channel is displayed in a separate color
25841 each channel is displayed using the same color scheme
25843 each channel is displayed using the rainbow color scheme
25845 each channel is displayed using the moreland color scheme
25847 each channel is displayed using the nebulae color scheme
25849 each channel is displayed using the fire color scheme
25851 each channel is displayed using the fiery color scheme
25853 each channel is displayed using the fruit color scheme
25855 each channel is displayed using the cool color scheme
25857 each channel is displayed using the magma color scheme
25859 each channel is displayed using the green color scheme
25861 each channel is displayed using the viridis color scheme
25863 each channel is displayed using the plasma color scheme
25865 each channel is displayed using the cividis color scheme
25867 each channel is displayed using the terrain color scheme
25869 Default value is @samp{intensity}.
25872 Specify scale used for calculating intensity color values.
25874 It accepts the following values:
25879 square root, default
25889 Default value is @samp{log}.
25892 Specify frequency scale.
25894 It accepts the following values:
25902 Default value is @samp{lin}.
25905 Set saturation modifier for displayed colors. Negative values provide
25906 alternative color scheme. @code{0} is no saturation at all.
25907 Saturation must be in [-10.0, 10.0] range.
25908 Default value is @code{1}.
25911 Set window function.
25913 It accepts the following values:
25937 Default value is @code{hann}.
25940 Set orientation of time vs frequency axis. Can be @code{vertical} or
25941 @code{horizontal}. Default is @code{vertical}.
25944 Set scale gain for calculating intensity color values.
25945 Default value is @code{1}.
25948 Draw time and frequency axes and legends. Default is enabled.
25951 Set color rotation, must be in [-1.0, 1.0] range.
25952 Default value is @code{0}.
25955 Set start frequency from which to display spectrogram. Default is @code{0}.
25958 Set stop frequency to which to display spectrogram. Default is @code{0}.
25961 @subsection Examples
25965 Extract an audio spectrogram of a whole audio track
25966 in a 1024x1024 picture using @command{ffmpeg}:
25968 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
25972 @section showvolume
25974 Convert input audio volume to a video output.
25976 The filter accepts the following options:
25983 Set border width, allowed range is [0, 5]. Default is 1.
25986 Set channel width, allowed range is [80, 8192]. Default is 400.
25989 Set channel height, allowed range is [1, 900]. Default is 20.
25992 Set fade, allowed range is [0, 1]. Default is 0.95.
25995 Set volume color expression.
25997 The expression can use the following variables:
26001 Current max volume of channel in dB.
26007 Current channel number, starting from 0.
26011 If set, displays channel names. Default is enabled.
26014 If set, displays volume values. Default is enabled.
26017 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26018 default is @code{h}.
26021 Set step size, allowed range is [0, 5]. Default is 0, which means
26025 Set background opacity, allowed range is [0, 1]. Default is 0.
26028 Set metering mode, can be peak: @code{p} or rms: @code{r},
26029 default is @code{p}.
26032 Set display scale, can be linear: @code{lin} or log: @code{log},
26033 default is @code{lin}.
26037 If set to > 0., display a line for the max level
26038 in the previous seconds.
26039 default is disabled: @code{0.}
26042 The color of the max line. Use when @code{dm} option is set to > 0.
26043 default is: @code{orange}
26048 Convert input audio to a video output, representing the samples waves.
26050 The filter accepts the following options:
26054 Specify the video size for the output. For the syntax of this option, check the
26055 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26056 Default value is @code{600x240}.
26061 Available values are:
26064 Draw a point for each sample.
26067 Draw a vertical line for each sample.
26070 Draw a point for each sample and a line between them.
26073 Draw a centered vertical line for each sample.
26076 Default value is @code{point}.
26079 Set the number of samples which are printed on the same column. A
26080 larger value will decrease the frame rate. Must be a positive
26081 integer. This option can be set only if the value for @var{rate}
26082 is not explicitly specified.
26085 Set the (approximate) output frame rate. This is done by setting the
26086 option @var{n}. Default value is "25".
26088 @item split_channels
26089 Set if channels should be drawn separately or overlap. Default value is 0.
26092 Set colors separated by '|' which are going to be used for drawing of each channel.
26095 Set amplitude scale.
26097 Available values are:
26115 Set the draw mode. This is mostly useful to set for high @var{n}.
26117 Available values are:
26120 Scale pixel values for each drawn sample.
26123 Draw every sample directly.
26126 Default value is @code{scale}.
26129 @subsection Examples
26133 Output the input file audio and the corresponding video representation
26136 amovie=a.mp3,asplit[out0],showwaves[out1]
26140 Create a synthetic signal and show it with showwaves, forcing a
26141 frame rate of 30 frames per second:
26143 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26147 @section showwavespic
26149 Convert input audio to a single video frame, representing the samples waves.
26151 The filter accepts the following options:
26155 Specify the video size for the output. For the syntax of this option, check the
26156 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26157 Default value is @code{600x240}.
26159 @item split_channels
26160 Set if channels should be drawn separately or overlap. Default value is 0.
26163 Set colors separated by '|' which are going to be used for drawing of each channel.
26166 Set amplitude scale.
26168 Available values are:
26188 Available values are:
26191 Scale pixel values for each drawn sample.
26194 Draw every sample directly.
26197 Default value is @code{scale}.
26200 Set the filter mode.
26202 Available values are:
26205 Use average samples values for each drawn sample.
26208 Use peak samples values for each drawn sample.
26211 Default value is @code{average}.
26214 @subsection Examples
26218 Extract a channel split representation of the wave form of a whole audio track
26219 in a 1024x800 picture using @command{ffmpeg}:
26221 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26225 @section sidedata, asidedata
26227 Delete frame side data, or select frames based on it.
26229 This filter accepts the following options:
26233 Set mode of operation of the filter.
26235 Can be one of the following:
26239 Select every frame with side data of @code{type}.
26242 Delete side data of @code{type}. If @code{type} is not set, delete all side
26248 Set side data type used with all modes. Must be set for @code{select} mode. For
26249 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26250 in @file{libavutil/frame.h}. For example, to choose
26251 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26255 @section spectrumsynth
26257 Synthesize audio from 2 input video spectrums, first input stream represents
26258 magnitude across time and second represents phase across time.
26259 The filter will transform from frequency domain as displayed in videos back
26260 to time domain as presented in audio output.
26262 This filter is primarily created for reversing processed @ref{showspectrum}
26263 filter outputs, but can synthesize sound from other spectrograms too.
26264 But in such case results are going to be poor if the phase data is not
26265 available, because in such cases phase data need to be recreated, usually
26266 it's just recreated from random noise.
26267 For best results use gray only output (@code{channel} color mode in
26268 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26269 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26270 @code{data} option. Inputs videos should generally use @code{fullframe}
26271 slide mode as that saves resources needed for decoding video.
26273 The filter accepts the following options:
26277 Specify sample rate of output audio, the sample rate of audio from which
26278 spectrum was generated may differ.
26281 Set number of channels represented in input video spectrums.
26284 Set scale which was used when generating magnitude input spectrum.
26285 Can be @code{lin} or @code{log}. Default is @code{log}.
26288 Set slide which was used when generating inputs spectrums.
26289 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26290 Default is @code{fullframe}.
26293 Set window function used for resynthesis.
26296 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26297 which means optimal overlap for selected window function will be picked.
26300 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26301 Default is @code{vertical}.
26304 @subsection Examples
26308 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26309 then resynthesize videos back to audio with spectrumsynth:
26311 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
26312 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
26313 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26317 @section split, asplit
26319 Split input into several identical outputs.
26321 @code{asplit} works with audio input, @code{split} with video.
26323 The filter accepts a single parameter which specifies the number of outputs. If
26324 unspecified, it defaults to 2.
26326 @subsection Examples
26330 Create two separate outputs from the same input:
26332 [in] split [out0][out1]
26336 To create 3 or more outputs, you need to specify the number of
26339 [in] asplit=3 [out0][out1][out2]
26343 Create two separate outputs from the same input, one cropped and
26346 [in] split [splitout1][splitout2];
26347 [splitout1] crop=100:100:0:0 [cropout];
26348 [splitout2] pad=200:200:100:100 [padout];
26352 Create 5 copies of the input audio with @command{ffmpeg}:
26354 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26360 Receive commands sent through a libzmq client, and forward them to
26361 filters in the filtergraph.
26363 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26364 must be inserted between two video filters, @code{azmq} between two
26365 audio filters. Both are capable to send messages to any filter type.
26367 To enable these filters you need to install the libzmq library and
26368 headers and configure FFmpeg with @code{--enable-libzmq}.
26370 For more information about libzmq see:
26371 @url{http://www.zeromq.org/}
26373 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26374 receives messages sent through a network interface defined by the
26375 @option{bind_address} (or the abbreviation "@option{b}") option.
26376 Default value of this option is @file{tcp://localhost:5555}. You may
26377 want to alter this value to your needs, but do not forget to escape any
26378 ':' signs (see @ref{filtergraph escaping}).
26380 The received message must be in the form:
26382 @var{TARGET} @var{COMMAND} [@var{ARG}]
26385 @var{TARGET} specifies the target of the command, usually the name of
26386 the filter class or a specific filter instance name. The default
26387 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26388 but you can override this by using the @samp{filter_name@@id} syntax
26389 (see @ref{Filtergraph syntax}).
26391 @var{COMMAND} specifies the name of the command for the target filter.
26393 @var{ARG} is optional and specifies the optional argument list for the
26394 given @var{COMMAND}.
26396 Upon reception, the message is processed and the corresponding command
26397 is injected into the filtergraph. Depending on the result, the filter
26398 will send a reply to the client, adopting the format:
26400 @var{ERROR_CODE} @var{ERROR_REASON}
26404 @var{MESSAGE} is optional.
26406 @subsection Examples
26408 Look at @file{tools/zmqsend} for an example of a zmq client which can
26409 be used to send commands processed by these filters.
26411 Consider the following filtergraph generated by @command{ffplay}.
26412 In this example the last overlay filter has an instance name. All other
26413 filters will have default instance names.
26416 ffplay -dumpgraph 1 -f lavfi "
26417 color=s=100x100:c=red [l];
26418 color=s=100x100:c=blue [r];
26419 nullsrc=s=200x100, zmq [bg];
26420 [bg][l] overlay [bg+l];
26421 [bg+l][r] overlay@@my=x=100 "
26424 To change the color of the left side of the video, the following
26425 command can be used:
26427 echo Parsed_color_0 c yellow | tools/zmqsend
26430 To change the right side:
26432 echo Parsed_color_1 c pink | tools/zmqsend
26435 To change the position of the right side:
26437 echo overlay@@my x 150 | tools/zmqsend
26441 @c man end MULTIMEDIA FILTERS
26443 @chapter Multimedia Sources
26444 @c man begin MULTIMEDIA SOURCES
26446 Below is a description of the currently available multimedia sources.
26450 This is the same as @ref{movie} source, except it selects an audio
26456 Read audio and/or video stream(s) from a movie container.
26458 It accepts the following parameters:
26462 The name of the resource to read (not necessarily a file; it can also be a
26463 device or a stream accessed through some protocol).
26465 @item format_name, f
26466 Specifies the format assumed for the movie to read, and can be either
26467 the name of a container or an input device. If not specified, the
26468 format is guessed from @var{movie_name} or by probing.
26470 @item seek_point, sp
26471 Specifies the seek point in seconds. The frames will be output
26472 starting from this seek point. The parameter is evaluated with
26473 @code{av_strtod}, so the numerical value may be suffixed by an IS
26474 postfix. The default value is "0".
26477 Specifies the streams to read. Several streams can be specified,
26478 separated by "+". The source will then have as many outputs, in the
26479 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26480 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26481 respectively the default (best suited) video and audio stream. Default
26482 is "dv", or "da" if the filter is called as "amovie".
26484 @item stream_index, si
26485 Specifies the index of the video stream to read. If the value is -1,
26486 the most suitable video stream will be automatically selected. The default
26487 value is "-1". Deprecated. If the filter is called "amovie", it will select
26488 audio instead of video.
26491 Specifies how many times to read the stream in sequence.
26492 If the value is 0, the stream will be looped infinitely.
26493 Default value is "1".
26495 Note that when the movie is looped the source timestamps are not
26496 changed, so it will generate non monotonically increasing timestamps.
26498 @item discontinuity
26499 Specifies the time difference between frames above which the point is
26500 considered a timestamp discontinuity which is removed by adjusting the later
26504 It allows overlaying a second video on top of the main input of
26505 a filtergraph, as shown in this graph:
26507 input -----------> deltapts0 --> overlay --> output
26510 movie --> scale--> deltapts1 -------+
26512 @subsection Examples
26516 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26517 on top of the input labelled "in":
26519 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26520 [in] setpts=PTS-STARTPTS [main];
26521 [main][over] overlay=16:16 [out]
26525 Read from a video4linux2 device, and overlay it on top of the input
26528 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26529 [in] setpts=PTS-STARTPTS [main];
26530 [main][over] overlay=16:16 [out]
26534 Read the first video stream and the audio stream with id 0x81 from
26535 dvd.vob; the video is connected to the pad named "video" and the audio is
26536 connected to the pad named "audio":
26538 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26542 @subsection Commands
26544 Both movie and amovie support the following commands:
26547 Perform seek using "av_seek_frame".
26548 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26551 @var{stream_index}: If stream_index is -1, a default
26552 stream is selected, and @var{timestamp} is automatically converted
26553 from AV_TIME_BASE units to the stream specific time_base.
26555 @var{timestamp}: Timestamp in AVStream.time_base units
26556 or, if no stream is specified, in AV_TIME_BASE units.
26558 @var{flags}: Flags which select direction and seeking mode.
26562 Get movie duration in AV_TIME_BASE units.
26566 @c man end MULTIMEDIA SOURCES