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 Options for filters with several inputs (framesync)
317 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
319 Some filters with several inputs support a common set of options.
320 These options can only be set by name, not with the short notation.
324 The action to take when EOF is encountered on the secondary input; it accepts
325 one of the following values:
329 Repeat the last frame (the default).
333 Pass the main input through.
337 If set to 1, force the output to terminate when the shortest input
338 terminates. Default value is 0.
341 If set to 1, force the filter to extend the last frame of secondary streams
342 until the end of the primary stream. A value of 0 disables this behavior.
346 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
348 @chapter Audio Filters
349 @c man begin AUDIO FILTERS
351 When you configure your FFmpeg build, you can disable any of the
352 existing filters using @code{--disable-filters}.
353 The configure output will show the audio filters included in your
356 Below is a description of the currently available audio filters.
360 A compressor is mainly used to reduce the dynamic range of a signal.
361 Especially modern music is mostly compressed at a high ratio to
362 improve the overall loudness. It's done to get the highest attention
363 of a listener, "fatten" the sound and bring more "power" to the track.
364 If a signal is compressed too much it may sound dull or "dead"
365 afterwards or it may start to "pump" (which could be a powerful effect
366 but can also destroy a track completely).
367 The right compression is the key to reach a professional sound and is
368 the high art of mixing and mastering. Because of its complex settings
369 it may take a long time to get the right feeling for this kind of effect.
371 Compression is done by detecting the volume above a chosen level
372 @code{threshold} and dividing it by the factor set with @code{ratio}.
373 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
374 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
375 the signal would cause distortion of the waveform the reduction can be
376 levelled over the time. This is done by setting "Attack" and "Release".
377 @code{attack} determines how long the signal has to rise above the threshold
378 before any reduction will occur and @code{release} sets the time the signal
379 has to fall below the threshold to reduce the reduction again. Shorter signals
380 than the chosen attack time will be left untouched.
381 The overall reduction of the signal can be made up afterwards with the
382 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
383 raising the makeup to this level results in a signal twice as loud than the
384 source. To gain a softer entry in the compression the @code{knee} flattens the
385 hard edge at the threshold in the range of the chosen decibels.
387 The filter accepts the following options:
391 Set input gain. Default is 1. Range is between 0.015625 and 64.
394 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
395 Default is @code{downward}.
398 If a signal of stream rises above this level it will affect the gain
400 By default it is 0.125. Range is between 0.00097563 and 1.
403 Set a ratio by which the signal is reduced. 1:2 means that if the level
404 rose 4dB above the threshold, it will be only 2dB above after the reduction.
405 Default is 2. Range is between 1 and 20.
408 Amount of milliseconds the signal has to rise above the threshold before gain
409 reduction starts. Default is 20. Range is between 0.01 and 2000.
412 Amount of milliseconds the signal has to fall below the threshold before
413 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
416 Set the amount by how much signal will be amplified after processing.
417 Default is 1. Range is from 1 to 64.
420 Curve the sharp knee around the threshold to enter gain reduction more softly.
421 Default is 2.82843. Range is between 1 and 8.
424 Choose if the @code{average} level between all channels of input stream
425 or the louder(@code{maximum}) channel of input stream affects the
426 reduction. Default is @code{average}.
429 Should the exact signal be taken in case of @code{peak} or an RMS one in case
430 of @code{rms}. Default is @code{rms} which is mostly smoother.
433 How much to use compressed signal in output. Default is 1.
434 Range is between 0 and 1.
438 Simple audio dynamic range compression/expansion filter.
440 The filter accepts the following options:
444 Set contrast. Default is 33. Allowed range is between 0 and 100.
449 Copy the input audio source unchanged to the output. This is mainly useful for
454 Apply cross fade from one input audio stream to another input audio stream.
455 The cross fade is applied for specified duration near the end of first stream.
457 The filter accepts the following options:
461 Specify the number of samples for which the cross fade effect has to last.
462 At the end of the cross fade effect the first input audio will be completely
463 silent. Default is 44100.
466 Specify the duration of the cross fade effect. See
467 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
468 for the accepted syntax.
469 By default the duration is determined by @var{nb_samples}.
470 If set this option is used instead of @var{nb_samples}.
473 Should first stream end overlap with second stream start. Default is enabled.
476 Set curve for cross fade transition for first stream.
479 Set curve for cross fade transition for second stream.
481 For description of available curve types see @ref{afade} filter description.
488 Cross fade from one input to another:
490 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
494 Cross fade from one input to another but without overlapping:
496 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
501 Split audio stream into several bands.
503 This filter splits audio stream into two or more frequency ranges.
504 Summing all streams back will give flat output.
506 The filter accepts the following options:
510 Set split frequencies. Those must be positive and increasing.
513 Set filter order, can be @var{2nd}, @var{4th} or @var{8th}.
514 Default is @var{4th}.
519 Reduce audio bit resolution.
521 This filter is bit crusher with enhanced functionality. A bit crusher
522 is used to audibly reduce number of bits an audio signal is sampled
523 with. This doesn't change the bit depth at all, it just produces the
524 effect. Material reduced in bit depth sounds more harsh and "digital".
525 This filter is able to even round to continuous values instead of discrete
527 Additionally it has a D/C offset which results in different crushing of
528 the lower and the upper half of the signal.
529 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
531 Another feature of this filter is the logarithmic mode.
532 This setting switches from linear distances between bits to logarithmic ones.
533 The result is a much more "natural" sounding crusher which doesn't gate low
534 signals for example. The human ear has a logarithmic perception,
535 so this kind of crushing is much more pleasant.
536 Logarithmic crushing is also able to get anti-aliased.
538 The filter accepts the following options:
554 Can be linear: @code{lin} or logarithmic: @code{log}.
563 Set sample reduction.
566 Enable LFO. By default disabled.
577 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
581 Remove impulsive noise from input audio.
583 Samples detected as impulsive noise are replaced by interpolated samples using
584 autoregressive modelling.
588 Set window size, in milliseconds. Allowed range is from @code{10} to
589 @code{100}. Default value is @code{55} milliseconds.
590 This sets size of window which will be processed at once.
593 Set window overlap, in percentage of window size. Allowed range is from
594 @code{50} to @code{95}. Default value is @code{75} percent.
595 Setting this to a very high value increases impulsive noise removal but makes
596 whole process much slower.
599 Set autoregression order, in percentage of window size. Allowed range is from
600 @code{0} to @code{25}. Default value is @code{2} percent. This option also
601 controls quality of interpolated samples using neighbour good samples.
604 Set threshold value. Allowed range is from @code{1} to @code{100}.
605 Default value is @code{2}.
606 This controls the strength of impulsive noise which is going to be removed.
607 The lower value, the more samples will be detected as impulsive noise.
610 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
611 @code{10}. Default value is @code{2}.
612 If any two samples detected as noise are spaced less than this value then any
613 sample between those two samples will be also detected as noise.
618 It accepts the following values:
621 Select overlap-add method. Even not interpolated samples are slightly
622 changed with this method.
625 Select overlap-save method. Not interpolated samples remain unchanged.
628 Default value is @code{a}.
632 Remove clipped samples from input audio.
634 Samples detected as clipped are replaced by interpolated samples using
635 autoregressive modelling.
639 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
640 Default value is @code{55} milliseconds.
641 This sets size of window which will be processed at once.
644 Set window overlap, in percentage of window size. Allowed range is from @code{50}
645 to @code{95}. Default value is @code{75} percent.
648 Set autoregression order, in percentage of window size. Allowed range is from
649 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
650 quality of interpolated samples using neighbour good samples.
653 Set threshold value. Allowed range is from @code{1} to @code{100}.
654 Default value is @code{10}. Higher values make clip detection less aggressive.
657 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
658 Default value is @code{1000}. Higher values make clip detection less aggressive.
663 It accepts the following values:
666 Select overlap-add method. Even not interpolated samples are slightly changed
670 Select overlap-save method. Not interpolated samples remain unchanged.
673 Default value is @code{a}.
678 Delay one or more audio channels.
680 Samples in delayed channel are filled with silence.
682 The filter accepts the following option:
686 Set list of delays in milliseconds for each channel separated by '|'.
687 Unused delays will be silently ignored. If number of given delays is
688 smaller than number of channels all remaining channels will not be delayed.
689 If you want to delay exact number of samples, append 'S' to number.
690 If you want instead to delay in seconds, append 's' to number.
697 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
698 the second channel (and any other channels that may be present) unchanged.
704 Delay second channel by 500 samples, the third channel by 700 samples and leave
705 the first channel (and any other channels that may be present) unchanged.
711 @section aderivative, aintegral
713 Compute derivative/integral of audio stream.
715 Applying both filters one after another produces original audio.
719 Apply echoing to the input audio.
721 Echoes are reflected sound and can occur naturally amongst mountains
722 (and sometimes large buildings) when talking or shouting; digital echo
723 effects emulate this behaviour and are often used to help fill out the
724 sound of a single instrument or vocal. The time difference between the
725 original signal and the reflection is the @code{delay}, and the
726 loudness of the reflected signal is the @code{decay}.
727 Multiple echoes can have different delays and decays.
729 A description of the accepted parameters follows.
733 Set input gain of reflected signal. Default is @code{0.6}.
736 Set output gain of reflected signal. Default is @code{0.3}.
739 Set list of time intervals in milliseconds between original signal and reflections
740 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
741 Default is @code{1000}.
744 Set list of loudness of reflected signals separated by '|'.
745 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
746 Default is @code{0.5}.
753 Make it sound as if there are twice as many instruments as are actually playing:
755 aecho=0.8:0.88:60:0.4
759 If delay is very short, then it sounds like a (metallic) robot playing music:
765 A longer delay will sound like an open air concert in the mountains:
767 aecho=0.8:0.9:1000:0.3
771 Same as above but with one more mountain:
773 aecho=0.8:0.9:1000|1800:0.3|0.25
778 Audio emphasis filter creates or restores material directly taken from LPs or
779 emphased CDs with different filter curves. E.g. to store music on vinyl the
780 signal has to be altered by a filter first to even out the disadvantages of
781 this recording medium.
782 Once the material is played back the inverse filter has to be applied to
783 restore the distortion of the frequency response.
785 The filter accepts the following options:
795 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
796 use @code{production} mode. Default is @code{reproduction} mode.
799 Set filter type. Selects medium. Can be one of the following:
811 select Compact Disc (CD).
817 select 50µs (FM-KF).
819 select 75µs (FM-KF).
825 Modify an audio signal according to the specified expressions.
827 This filter accepts one or more expressions (one for each channel),
828 which are evaluated and used to modify a corresponding audio signal.
830 It accepts the following parameters:
834 Set the '|'-separated expressions list for each separate channel. If
835 the number of input channels is greater than the number of
836 expressions, the last specified expression is used for the remaining
839 @item channel_layout, c
840 Set output channel layout. If not specified, the channel layout is
841 specified by the number of expressions. If set to @samp{same}, it will
842 use by default the same input channel layout.
845 Each expression in @var{exprs} can contain the following constants and functions:
849 channel number of the current expression
852 number of the evaluated sample, starting from 0
858 time of the evaluated sample expressed in seconds
861 @item nb_out_channels
862 input and output number of channels
865 the value of input channel with number @var{CH}
868 Note: this filter is slow. For faster processing you should use a
877 aeval=val(ch)/2:c=same
881 Invert phase of the second channel:
890 Apply fade-in/out effect to input audio.
892 A description of the accepted parameters follows.
896 Specify the effect type, can be either @code{in} for fade-in, or
897 @code{out} for a fade-out effect. Default is @code{in}.
899 @item start_sample, ss
900 Specify the number of the start sample for starting to apply the fade
901 effect. Default is 0.
904 Specify the number of samples for which the fade effect has to last. At
905 the end of the fade-in effect the output audio will have the same
906 volume as the input audio, at the end of the fade-out transition
907 the output audio will be silence. Default is 44100.
910 Specify the start time of the fade effect. Default is 0.
911 The value must be specified as a time duration; see
912 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
913 for the accepted syntax.
914 If set this option is used instead of @var{start_sample}.
917 Specify the duration of the fade effect. See
918 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
919 for the accepted syntax.
920 At the end of the fade-in effect the output audio will have the same
921 volume as the input audio, at the end of the fade-out transition
922 the output audio will be silence.
923 By default the duration is determined by @var{nb_samples}.
924 If set this option is used instead of @var{nb_samples}.
927 Set curve for fade transition.
929 It accepts the following values:
932 select triangular, linear slope (default)
934 select quarter of sine wave
936 select half of sine wave
938 select exponential sine wave
942 select inverted parabola
956 select inverted quarter of sine wave
958 select inverted half of sine wave
960 select double-exponential seat
962 select double-exponential sigmoid
964 select logistic sigmoid
974 Fade in first 15 seconds of audio:
980 Fade out last 25 seconds of a 900 seconds audio:
982 afade=t=out:st=875:d=25
987 Denoise audio samples with FFT.
989 A description of the accepted parameters follows.
993 Set the noise reduction in dB, allowed range is 0.01 to 97.
994 Default value is 12 dB.
997 Set the noise floor in dB, allowed range is -80 to -20.
998 Default value is -50 dB.
1003 It accepts the following values:
1012 Select shellac noise.
1015 Select custom noise, defined in @code{bn} option.
1017 Default value is white noise.
1021 Set custom band noise for every one of 15 bands.
1022 Bands are separated by ' ' or '|'.
1025 Set the residual floor in dB, allowed range is -80 to -20.
1026 Default value is -38 dB.
1029 Enable noise tracking. By default is disabled.
1030 With this enabled, noise floor is automatically adjusted.
1033 Enable residual tracking. By default is disabled.
1036 Set the output mode.
1038 It accepts the following values:
1041 Pass input unchanged.
1044 Pass noise filtered out.
1049 Default value is @var{o}.
1053 @subsection Commands
1055 This filter supports the following commands:
1057 @item sample_noise, sn
1058 Start or stop measuring noise profile.
1059 Syntax for the command is : "start" or "stop" string.
1060 After measuring noise profile is stopped it will be
1061 automatically applied in filtering.
1063 @item noise_reduction, nr
1064 Change noise reduction. Argument is single float number.
1065 Syntax for the command is : "@var{noise_reduction}"
1067 @item noise_floor, nf
1068 Change noise floor. Argument is single float number.
1069 Syntax for the command is : "@var{noise_floor}"
1071 @item output_mode, om
1072 Change output mode operation.
1073 Syntax for the command is : "i", "o" or "n" string.
1077 Apply arbitrary expressions to samples in frequency domain.
1081 Set frequency domain real expression for each separate channel separated
1082 by '|'. Default is "re".
1083 If the number of input channels is greater than the number of
1084 expressions, the last specified expression is used for the remaining
1088 Set frequency domain imaginary expression for each separate channel
1089 separated by '|'. Default is "im".
1091 Each expression in @var{real} and @var{imag} can contain the following
1092 constants and functions:
1099 current frequency bin number
1102 number of available bins
1105 channel number of the current expression
1114 current real part of frequency bin of current channel
1117 current imaginary part of frequency bin of current channel
1120 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1123 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1127 Set window size. Allowed range is from 16 to 131072.
1128 Default is @code{4096}
1131 Set window function. Default is @code{hann}.
1134 Set window overlap. If set to 1, the recommended overlap for selected
1135 window function will be picked. Default is @code{0.75}.
1138 @subsection Examples
1142 Leave almost only low frequencies in audio:
1144 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1151 Apply an arbitrary Frequency Impulse Response filter.
1153 This filter is designed for applying long FIR filters,
1154 up to 60 seconds long.
1156 It can be used as component for digital crossover filters,
1157 room equalization, cross talk cancellation, wavefield synthesis,
1158 auralization, ambiophonics, ambisonics and spatialization.
1160 This filter uses the second stream as FIR coefficients.
1161 If the second stream holds a single channel, it will be used
1162 for all input channels in the first stream, otherwise
1163 the number of channels in the second stream must be same as
1164 the number of channels in the first stream.
1166 It accepts the following parameters:
1170 Set dry gain. This sets input gain.
1173 Set wet gain. This sets final output gain.
1176 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1179 Enable applying gain measured from power of IR.
1181 Set which approach to use for auto gain measurement.
1185 Do not apply any gain.
1188 select peak gain, very conservative approach. This is default value.
1191 select DC gain, limited application.
1194 select gain to noise approach, this is most popular one.
1198 Set gain to be applied to IR coefficients before filtering.
1199 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1202 Set format of IR stream. Can be @code{mono} or @code{input}.
1203 Default is @code{input}.
1206 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1207 Allowed range is 0.1 to 60 seconds.
1210 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1211 By default it is disabled.
1214 Set for which IR channel to display frequency response. By default is first channel
1215 displayed. This option is used only when @var{response} is enabled.
1218 Set video stream size. This option is used only when @var{response} is enabled.
1221 Set video stream frame rate. This option is used only when @var{response} is enabled.
1224 Set minimal partition size used for convolution. Default is @var{8192}.
1225 Allowed range is from @var{8} to @var{32768}.
1226 Lower values decreases latency at cost of higher CPU usage.
1229 Set maximal partition size used for convolution. Default is @var{8192}.
1230 Allowed range is from @var{8} to @var{32768}.
1231 Lower values may increase CPU usage.
1234 @subsection Examples
1238 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1240 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1247 Set output format constraints for the input audio. The framework will
1248 negotiate the most appropriate format to minimize conversions.
1250 It accepts the following parameters:
1254 A '|'-separated list of requested sample formats.
1257 A '|'-separated list of requested sample rates.
1259 @item channel_layouts
1260 A '|'-separated list of requested channel layouts.
1262 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1263 for the required syntax.
1266 If a parameter is omitted, all values are allowed.
1268 Force the output to either unsigned 8-bit or signed 16-bit stereo
1270 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1275 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1276 processing reduces disturbing noise between useful signals.
1278 Gating is done by detecting the volume below a chosen level @var{threshold}
1279 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1280 floor is set via @var{range}. Because an exact manipulation of the signal
1281 would cause distortion of the waveform the reduction can be levelled over
1282 time. This is done by setting @var{attack} and @var{release}.
1284 @var{attack} determines how long the signal has to fall below the threshold
1285 before any reduction will occur and @var{release} sets the time the signal
1286 has to rise above the threshold to reduce the reduction again.
1287 Shorter signals than the chosen attack time will be left untouched.
1291 Set input level before filtering.
1292 Default is 1. Allowed range is from 0.015625 to 64.
1295 Set the mode of operation. Can be @code{upward} or @code{downward}.
1296 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1297 will be amplified, expanding dynamic range in upward direction.
1298 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1301 Set the level of gain reduction when the signal is below the threshold.
1302 Default is 0.06125. Allowed range is from 0 to 1.
1303 Setting this to 0 disables reduction and then filter behaves like expander.
1306 If a signal rises above this level the gain reduction is released.
1307 Default is 0.125. Allowed range is from 0 to 1.
1310 Set a ratio by which the signal is reduced.
1311 Default is 2. Allowed range is from 1 to 9000.
1314 Amount of milliseconds the signal has to rise above the threshold before gain
1316 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1319 Amount of milliseconds the signal has to fall below the threshold before the
1320 reduction is increased again. Default is 250 milliseconds.
1321 Allowed range is from 0.01 to 9000.
1324 Set amount of amplification of signal after processing.
1325 Default is 1. Allowed range is from 1 to 64.
1328 Curve the sharp knee around the threshold to enter gain reduction more softly.
1329 Default is 2.828427125. Allowed range is from 1 to 8.
1332 Choose if exact signal should be taken for detection or an RMS like one.
1333 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1336 Choose if the average level between all channels or the louder channel affects
1338 Default is @code{average}. Can be @code{average} or @code{maximum}.
1343 Apply an arbitrary Infinite Impulse Response filter.
1345 It accepts the following parameters:
1349 Set numerator/zeros coefficients.
1352 Set denominator/poles coefficients.
1364 Set coefficients format.
1370 Z-plane zeros/poles, cartesian (default)
1372 Z-plane zeros/poles, polar radians
1374 Z-plane zeros/poles, polar degrees
1378 Set kind of processing.
1379 Can be @code{d} - direct or @code{s} - serial cascading. Default is @code{s}.
1382 Set filtering precision.
1386 double-precision floating-point (default)
1388 single-precision floating-point
1396 How much to use filtered signal in output. Default is 1.
1397 Range is between 0 and 1.
1400 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1401 By default it is disabled.
1404 Set for which IR channel to display frequency response. By default is first channel
1405 displayed. This option is used only when @var{response} is enabled.
1408 Set video stream size. This option is used only when @var{response} is enabled.
1411 Coefficients in @code{tf} format are separated by spaces and are in ascending
1414 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1415 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1418 Different coefficients and gains can be provided for every channel, in such case
1419 use '|' to separate coefficients or gains. Last provided coefficients will be
1420 used for all remaining channels.
1422 @subsection Examples
1426 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1428 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
1432 Same as above but in @code{zp} format:
1434 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
1440 The limiter prevents an input signal from rising over a desired threshold.
1441 This limiter uses lookahead technology to prevent your signal from distorting.
1442 It means that there is a small delay after the signal is processed. Keep in mind
1443 that the delay it produces is the attack time you set.
1445 The filter accepts the following options:
1449 Set input gain. Default is 1.
1452 Set output gain. Default is 1.
1455 Don't let signals above this level pass the limiter. Default is 1.
1458 The limiter will reach its attenuation level in this amount of time in
1459 milliseconds. Default is 5 milliseconds.
1462 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1463 Default is 50 milliseconds.
1466 When gain reduction is always needed ASC takes care of releasing to an
1467 average reduction level rather than reaching a reduction of 0 in the release
1471 Select how much the release time is affected by ASC, 0 means nearly no changes
1472 in release time while 1 produces higher release times.
1475 Auto level output signal. Default is enabled.
1476 This normalizes audio back to 0dB if enabled.
1479 Depending on picked setting it is recommended to upsample input 2x or 4x times
1480 with @ref{aresample} before applying this filter.
1484 Apply a two-pole all-pass filter with central frequency (in Hz)
1485 @var{frequency}, and filter-width @var{width}.
1486 An all-pass filter changes the audio's frequency to phase relationship
1487 without changing its frequency to amplitude relationship.
1489 The filter accepts the following options:
1493 Set frequency in Hz.
1496 Set method to specify band-width of filter.
1511 Specify the band-width of a filter in width_type units.
1514 How much to use filtered signal in output. Default is 1.
1515 Range is between 0 and 1.
1518 Specify which channels to filter, by default all available are filtered.
1521 @subsection Commands
1523 This filter supports the following commands:
1526 Change allpass frequency.
1527 Syntax for the command is : "@var{frequency}"
1530 Change allpass width_type.
1531 Syntax for the command is : "@var{width_type}"
1534 Change allpass width.
1535 Syntax for the command is : "@var{width}"
1539 Syntax for the command is : "@var{mix}"
1546 The filter accepts the following options:
1550 Set the number of loops. Setting this value to -1 will result in infinite loops.
1554 Set maximal number of samples. Default is 0.
1557 Set first sample of loop. Default is 0.
1563 Merge two or more audio streams into a single multi-channel stream.
1565 The filter accepts the following options:
1570 Set the number of inputs. Default is 2.
1574 If the channel layouts of the inputs are disjoint, and therefore compatible,
1575 the channel layout of the output will be set accordingly and the channels
1576 will be reordered as necessary. If the channel layouts of the inputs are not
1577 disjoint, the output will have all the channels of the first input then all
1578 the channels of the second input, in that order, and the channel layout of
1579 the output will be the default value corresponding to the total number of
1582 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1583 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1584 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1585 first input, b1 is the first channel of the second input).
1587 On the other hand, if both input are in stereo, the output channels will be
1588 in the default order: a1, a2, b1, b2, and the channel layout will be
1589 arbitrarily set to 4.0, which may or may not be the expected value.
1591 All inputs must have the same sample rate, and format.
1593 If inputs do not have the same duration, the output will stop with the
1596 @subsection Examples
1600 Merge two mono files into a stereo stream:
1602 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1606 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1608 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
1614 Mixes multiple audio inputs into a single output.
1616 Note that this filter only supports float samples (the @var{amerge}
1617 and @var{pan} audio filters support many formats). If the @var{amix}
1618 input has integer samples then @ref{aresample} will be automatically
1619 inserted to perform the conversion to float samples.
1623 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1625 will mix 3 input audio streams to a single output with the same duration as the
1626 first input and a dropout transition time of 3 seconds.
1628 It accepts the following parameters:
1632 The number of inputs. If unspecified, it defaults to 2.
1635 How to determine the end-of-stream.
1639 The duration of the longest input. (default)
1642 The duration of the shortest input.
1645 The duration of the first input.
1649 @item dropout_transition
1650 The transition time, in seconds, for volume renormalization when an input
1651 stream ends. The default value is 2 seconds.
1654 Specify weight of each input audio stream as sequence.
1655 Each weight is separated by space. By default all inputs have same weight.
1660 Multiply first audio stream with second audio stream and store result
1661 in output audio stream. Multiplication is done by multiplying each
1662 sample from first stream with sample at same position from second stream.
1664 With this element-wise multiplication one can create amplitude fades and
1665 amplitude modulations.
1667 @section anequalizer
1669 High-order parametric multiband equalizer for each channel.
1671 It accepts the following parameters:
1675 This option string is in format:
1676 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1677 Each equalizer band is separated by '|'.
1681 Set channel number to which equalization will be applied.
1682 If input doesn't have that channel the entry is ignored.
1685 Set central frequency for band.
1686 If input doesn't have that frequency the entry is ignored.
1689 Set band width in hertz.
1692 Set band gain in dB.
1695 Set filter type for band, optional, can be:
1699 Butterworth, this is default.
1710 With this option activated frequency response of anequalizer is displayed
1714 Set video stream size. Only useful if curves option is activated.
1717 Set max gain that will be displayed. Only useful if curves option is activated.
1718 Setting this to a reasonable value makes it possible to display gain which is derived from
1719 neighbour bands which are too close to each other and thus produce higher gain
1720 when both are activated.
1723 Set frequency scale used to draw frequency response in video output.
1724 Can be linear or logarithmic. Default is logarithmic.
1727 Set color for each channel curve which is going to be displayed in video stream.
1728 This is list of color names separated by space or by '|'.
1729 Unrecognised or missing colors will be replaced by white color.
1732 @subsection Examples
1736 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1737 for first 2 channels using Chebyshev type 1 filter:
1739 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Alter existing filter parameters.
1749 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1751 @var{fN} is existing filter number, starting from 0, if no such filter is available
1753 @var{freq} set new frequency parameter.
1754 @var{width} set new width parameter in herz.
1755 @var{gain} set new gain parameter in dB.
1757 Full filter invocation with asendcmd may look like this:
1758 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1763 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1765 Each sample is adjusted by looking for other samples with similar contexts. This
1766 context similarity is defined by comparing their surrounding patches of size
1767 @option{p}. Patches are searched in an area of @option{r} around the sample.
1769 The filter accepts the following options:
1773 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1776 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1777 Default value is 2 milliseconds.
1780 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1781 Default value is 6 milliseconds.
1784 Set the output mode.
1786 It accepts the following values:
1789 Pass input unchanged.
1792 Pass noise filtered out.
1797 Default value is @var{o}.
1801 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
1804 @subsection Commands
1806 This filter supports the following commands:
1809 Change denoise strength. Argument is single float number.
1810 Syntax for the command is : "@var{s}"
1814 Syntax for the command is : "i", "o" or "n" string.
1819 Pass the audio source unchanged to the output.
1823 Pad the end of an audio stream with silence.
1825 This can be used together with @command{ffmpeg} @option{-shortest} to
1826 extend audio streams to the same length as the video stream.
1828 A description of the accepted options follows.
1832 Set silence packet size. Default value is 4096.
1835 Set the number of samples of silence to add to the end. After the
1836 value is reached, the stream is terminated. This option is mutually
1837 exclusive with @option{whole_len}.
1840 Set the minimum total number of samples in the output audio stream. If
1841 the value is longer than the input audio length, silence is added to
1842 the end, until the value is reached. This option is mutually exclusive
1843 with @option{pad_len}.
1846 Specify the duration of samples of silence to add. See
1847 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1848 for the accepted syntax. Used only if set to non-zero value.
1851 Specify the minimum total duration in the output audio stream. See
1852 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1853 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
1854 the input audio length, silence is added to the end, until the value is reached.
1855 This option is mutually exclusive with @option{pad_dur}
1858 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
1859 nor @option{whole_dur} option is set, the filter will add silence to the end of
1860 the input stream indefinitely.
1862 @subsection Examples
1866 Add 1024 samples of silence to the end of the input:
1872 Make sure the audio output will contain at least 10000 samples, pad
1873 the input with silence if required:
1875 apad=whole_len=10000
1879 Use @command{ffmpeg} to pad the audio input with silence, so that the
1880 video stream will always result the shortest and will be converted
1881 until the end in the output file when using the @option{shortest}
1884 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
1889 Add a phasing effect to the input audio.
1891 A phaser filter creates series of peaks and troughs in the frequency spectrum.
1892 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
1894 A description of the accepted parameters follows.
1898 Set input gain. Default is 0.4.
1901 Set output gain. Default is 0.74
1904 Set delay in milliseconds. Default is 3.0.
1907 Set decay. Default is 0.4.
1910 Set modulation speed in Hz. Default is 0.5.
1913 Set modulation type. Default is triangular.
1915 It accepts the following values:
1924 Audio pulsator is something between an autopanner and a tremolo.
1925 But it can produce funny stereo effects as well. Pulsator changes the volume
1926 of the left and right channel based on a LFO (low frequency oscillator) with
1927 different waveforms and shifted phases.
1928 This filter have the ability to define an offset between left and right
1929 channel. An offset of 0 means that both LFO shapes match each other.
1930 The left and right channel are altered equally - a conventional tremolo.
1931 An offset of 50% means that the shape of the right channel is exactly shifted
1932 in phase (or moved backwards about half of the frequency) - pulsator acts as
1933 an autopanner. At 1 both curves match again. Every setting in between moves the
1934 phase shift gapless between all stages and produces some "bypassing" sounds with
1935 sine and triangle waveforms. The more you set the offset near 1 (starting from
1936 the 0.5) the faster the signal passes from the left to the right speaker.
1938 The filter accepts the following options:
1942 Set input gain. By default it is 1. Range is [0.015625 - 64].
1945 Set output gain. By default it is 1. Range is [0.015625 - 64].
1948 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
1949 sawup or sawdown. Default is sine.
1952 Set modulation. Define how much of original signal is affected by the LFO.
1955 Set left channel offset. Default is 0. Allowed range is [0 - 1].
1958 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
1961 Set pulse width. Default is 1. Allowed range is [0 - 2].
1964 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
1967 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
1971 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
1975 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
1976 if timing is set to hz.
1982 Resample the input audio to the specified parameters, using the
1983 libswresample library. If none are specified then the filter will
1984 automatically convert between its input and output.
1986 This filter is also able to stretch/squeeze the audio data to make it match
1987 the timestamps or to inject silence / cut out audio to make it match the
1988 timestamps, do a combination of both or do neither.
1990 The filter accepts the syntax
1991 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
1992 expresses a sample rate and @var{resampler_options} is a list of
1993 @var{key}=@var{value} pairs, separated by ":". See the
1994 @ref{Resampler Options,,"Resampler Options" section in the
1995 ffmpeg-resampler(1) manual,ffmpeg-resampler}
1996 for the complete list of supported options.
1998 @subsection Examples
2002 Resample the input audio to 44100Hz:
2008 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2009 samples per second compensation:
2011 aresample=async=1000
2017 Reverse an audio clip.
2019 Warning: This filter requires memory to buffer the entire clip, so trimming
2022 @subsection Examples
2026 Take the first 5 seconds of a clip, and reverse it.
2028 atrim=end=5,areverse
2032 @section asetnsamples
2034 Set the number of samples per each output audio frame.
2036 The last output packet may contain a different number of samples, as
2037 the filter will flush all the remaining samples when the input audio
2040 The filter accepts the following options:
2044 @item nb_out_samples, n
2045 Set the number of frames per each output audio frame. The number is
2046 intended as the number of samples @emph{per each channel}.
2047 Default value is 1024.
2050 If set to 1, the filter will pad the last audio frame with zeroes, so
2051 that the last frame will contain the same number of samples as the
2052 previous ones. Default value is 1.
2055 For example, to set the number of per-frame samples to 1234 and
2056 disable padding for the last frame, use:
2058 asetnsamples=n=1234:p=0
2063 Set the sample rate without altering the PCM data.
2064 This will result in a change of speed and pitch.
2066 The filter accepts the following options:
2069 @item sample_rate, r
2070 Set the output sample rate. Default is 44100 Hz.
2075 Show a line containing various information for each input audio frame.
2076 The input audio is not modified.
2078 The shown line contains a sequence of key/value pairs of the form
2079 @var{key}:@var{value}.
2081 The following values are shown in the output:
2085 The (sequential) number of the input frame, starting from 0.
2088 The presentation timestamp of the input frame, in time base units; the time base
2089 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2092 The presentation timestamp of the input frame in seconds.
2095 position of the frame in the input stream, -1 if this information in
2096 unavailable and/or meaningless (for example in case of synthetic audio)
2105 The sample rate for the audio frame.
2108 The number of samples (per channel) in the frame.
2111 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2112 audio, the data is treated as if all the planes were concatenated.
2114 @item plane_checksums
2115 A list of Adler-32 checksums for each data plane.
2119 Apply audio soft clipping.
2121 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2122 along a smooth curve, rather than the abrupt shape of hard-clipping.
2124 This filter accepts the following options:
2128 Set type of soft-clipping.
2130 It accepts the following values:
2142 Set additional parameter which controls sigmoid function.
2146 Automatic Speech Recognition
2148 This filter uses PocketSphinx for speech recognition. To enable
2149 compilation of this filter, you need to configure FFmpeg with
2150 @code{--enable-pocketsphinx}.
2152 It accepts the following options:
2156 Set sampling rate of input audio. Defaults is @code{16000}.
2157 This need to match speech models, otherwise one will get poor results.
2160 Set dictionary containing acoustic model files.
2163 Set pronunciation dictionary.
2166 Set language model file.
2169 Set language model set.
2172 Set which language model to use.
2175 Set output for log messages.
2178 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2183 Display time domain statistical information about the audio channels.
2184 Statistics are calculated and displayed for each audio channel and,
2185 where applicable, an overall figure is also given.
2187 It accepts the following option:
2190 Short window length in seconds, used for peak and trough RMS measurement.
2191 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2195 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2196 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2199 Available keys for each channel are:
2241 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2242 this @code{lavfi.astats.Overall.Peak_count}.
2244 For description what each key means read below.
2247 Set number of frame after which stats are going to be recalculated.
2248 Default is disabled.
2250 @item measure_perchannel
2251 Select the entries which need to be measured per channel. The metadata keys can
2252 be used as flags, default is @option{all} which measures everything.
2253 @option{none} disables all per channel measurement.
2255 @item measure_overall
2256 Select the entries which need to be measured overall. The metadata keys can
2257 be used as flags, default is @option{all} which measures everything.
2258 @option{none} disables all overall measurement.
2262 A description of each shown parameter follows:
2266 Mean amplitude displacement from zero.
2269 Minimal sample level.
2272 Maximal sample level.
2274 @item Min difference
2275 Minimal difference between two consecutive samples.
2277 @item Max difference
2278 Maximal difference between two consecutive samples.
2280 @item Mean difference
2281 Mean difference between two consecutive samples.
2282 The average of each difference between two consecutive samples.
2284 @item RMS difference
2285 Root Mean Square difference between two consecutive samples.
2289 Standard peak and RMS level measured in dBFS.
2293 Peak and trough values for RMS level measured over a short window.
2296 Standard ratio of peak to RMS level (note: not in dB).
2299 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2300 (i.e. either @var{Min level} or @var{Max level}).
2303 Number of occasions (not the number of samples) that the signal attained either
2304 @var{Min level} or @var{Max level}.
2307 Overall bit depth of audio. Number of bits used for each sample.
2310 Measured dynamic range of audio in dB.
2312 @item Zero crossings
2313 Number of points where the waveform crosses the zero level axis.
2315 @item Zero crossings rate
2316 Rate of Zero crossings and number of audio samples.
2323 The filter accepts exactly one parameter, the audio tempo. If not
2324 specified then the filter will assume nominal 1.0 tempo. Tempo must
2325 be in the [0.5, 100.0] range.
2327 Note that tempo greater than 2 will skip some samples rather than
2328 blend them in. If for any reason this is a concern it is always
2329 possible to daisy-chain several instances of atempo to achieve the
2330 desired product tempo.
2332 @subsection Examples
2336 Slow down audio to 80% tempo:
2342 To speed up audio to 300% tempo:
2348 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2350 atempo=sqrt(3),atempo=sqrt(3)
2356 Trim the input so that the output contains one continuous subpart of the input.
2358 It accepts the following parameters:
2361 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2362 sample with the timestamp @var{start} will be the first sample in the output.
2365 Specify time of the first audio sample that will be dropped, i.e. the
2366 audio sample immediately preceding the one with the timestamp @var{end} will be
2367 the last sample in the output.
2370 Same as @var{start}, except this option sets the start timestamp in samples
2374 Same as @var{end}, except this option sets the end timestamp in samples instead
2378 The maximum duration of the output in seconds.
2381 The number of the first sample that should be output.
2384 The number of the first sample that should be dropped.
2387 @option{start}, @option{end}, and @option{duration} are expressed as time
2388 duration specifications; see
2389 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2391 Note that the first two sets of the start/end options and the @option{duration}
2392 option look at the frame timestamp, while the _sample options simply count the
2393 samples that pass through the filter. So start/end_pts and start/end_sample will
2394 give different results when the timestamps are wrong, inexact or do not start at
2395 zero. Also note that this filter does not modify the timestamps. If you wish
2396 to have the output timestamps start at zero, insert the asetpts filter after the
2399 If multiple start or end options are set, this filter tries to be greedy and
2400 keep all samples that match at least one of the specified constraints. To keep
2401 only the part that matches all the constraints at once, chain multiple atrim
2404 The defaults are such that all the input is kept. So it is possible to set e.g.
2405 just the end values to keep everything before the specified time.
2410 Drop everything except the second minute of input:
2412 ffmpeg -i INPUT -af atrim=60:120
2416 Keep only the first 1000 samples:
2418 ffmpeg -i INPUT -af atrim=end_sample=1000
2425 Apply a two-pole Butterworth band-pass filter with central
2426 frequency @var{frequency}, and (3dB-point) band-width width.
2427 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2428 instead of the default: constant 0dB peak gain.
2429 The filter roll off at 6dB per octave (20dB per decade).
2431 The filter accepts the following options:
2435 Set the filter's central frequency. Default is @code{3000}.
2438 Constant skirt gain if set to 1. Defaults to 0.
2441 Set method to specify band-width of filter.
2456 Specify the band-width of a filter in width_type units.
2459 How much to use filtered signal in output. Default is 1.
2460 Range is between 0 and 1.
2463 Specify which channels to filter, by default all available are filtered.
2466 @subsection Commands
2468 This filter supports the following commands:
2471 Change bandpass frequency.
2472 Syntax for the command is : "@var{frequency}"
2475 Change bandpass width_type.
2476 Syntax for the command is : "@var{width_type}"
2479 Change bandpass width.
2480 Syntax for the command is : "@var{width}"
2483 Change bandpass mix.
2484 Syntax for the command is : "@var{mix}"
2489 Apply a two-pole Butterworth band-reject filter with central
2490 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2491 The filter roll off at 6dB per octave (20dB per decade).
2493 The filter accepts the following options:
2497 Set the filter's central frequency. Default is @code{3000}.
2500 Set method to specify band-width of filter.
2515 Specify the band-width of a filter in width_type units.
2518 How much to use filtered signal in output. Default is 1.
2519 Range is between 0 and 1.
2522 Specify which channels to filter, by default all available are filtered.
2525 @subsection Commands
2527 This filter supports the following commands:
2530 Change bandreject frequency.
2531 Syntax for the command is : "@var{frequency}"
2534 Change bandreject width_type.
2535 Syntax for the command is : "@var{width_type}"
2538 Change bandreject width.
2539 Syntax for the command is : "@var{width}"
2542 Change bandreject mix.
2543 Syntax for the command is : "@var{mix}"
2546 @section bass, lowshelf
2548 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2549 shelving filter with a response similar to that of a standard
2550 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2552 The filter accepts the following options:
2556 Give the gain at 0 Hz. Its useful range is about -20
2557 (for a large cut) to +20 (for a large boost).
2558 Beware of clipping when using a positive gain.
2561 Set the filter's central frequency and so can be used
2562 to extend or reduce the frequency range to be boosted or cut.
2563 The default value is @code{100} Hz.
2566 Set method to specify band-width of filter.
2581 Determine how steep is the filter's shelf transition.
2584 How much to use filtered signal in output. Default is 1.
2585 Range is between 0 and 1.
2588 Specify which channels to filter, by default all available are filtered.
2591 @subsection Commands
2593 This filter supports the following commands:
2596 Change bass frequency.
2597 Syntax for the command is : "@var{frequency}"
2600 Change bass width_type.
2601 Syntax for the command is : "@var{width_type}"
2605 Syntax for the command is : "@var{width}"
2609 Syntax for the command is : "@var{gain}"
2613 Syntax for the command is : "@var{mix}"
2618 Apply a biquad IIR filter with the given coefficients.
2619 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2620 are the numerator and denominator coefficients respectively.
2621 and @var{channels}, @var{c} specify which channels to filter, by default all
2622 available are filtered.
2624 @subsection Commands
2626 This filter supports the following commands:
2634 Change biquad parameter.
2635 Syntax for the command is : "@var{value}"
2638 How much to use filtered signal in output. Default is 1.
2639 Range is between 0 and 1.
2643 Bauer stereo to binaural transformation, which improves headphone listening of
2644 stereo audio records.
2646 To enable compilation of this filter you need to configure FFmpeg with
2647 @code{--enable-libbs2b}.
2649 It accepts the following parameters:
2653 Pre-defined crossfeed level.
2657 Default level (fcut=700, feed=50).
2660 Chu Moy circuit (fcut=700, feed=60).
2663 Jan Meier circuit (fcut=650, feed=95).
2668 Cut frequency (in Hz).
2677 Remap input channels to new locations.
2679 It accepts the following parameters:
2682 Map channels from input to output. The argument is a '|'-separated list of
2683 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2684 @var{in_channel} form. @var{in_channel} can be either the name of the input
2685 channel (e.g. FL for front left) or its index in the input channel layout.
2686 @var{out_channel} is the name of the output channel or its index in the output
2687 channel layout. If @var{out_channel} is not given then it is implicitly an
2688 index, starting with zero and increasing by one for each mapping.
2690 @item channel_layout
2691 The channel layout of the output stream.
2694 If no mapping is present, the filter will implicitly map input channels to
2695 output channels, preserving indices.
2697 @subsection Examples
2701 For example, assuming a 5.1+downmix input MOV file,
2703 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2705 will create an output WAV file tagged as stereo from the downmix channels of
2709 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2711 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2715 @section channelsplit
2717 Split each channel from an input audio stream into a separate output stream.
2719 It accepts the following parameters:
2721 @item channel_layout
2722 The channel layout of the input stream. The default is "stereo".
2724 A channel layout describing the channels to be extracted as separate output streams
2725 or "all" to extract each input channel as a separate stream. The default is "all".
2727 Choosing channels not present in channel layout in the input will result in an error.
2730 @subsection Examples
2734 For example, assuming a stereo input MP3 file,
2736 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2738 will create an output Matroska file with two audio streams, one containing only
2739 the left channel and the other the right channel.
2742 Split a 5.1 WAV file into per-channel files:
2744 ffmpeg -i in.wav -filter_complex
2745 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2746 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2747 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2752 Extract only LFE from a 5.1 WAV file:
2754 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
2755 -map '[LFE]' lfe.wav
2760 Add a chorus effect to the audio.
2762 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2764 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2765 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2766 The modulation depth defines the range the modulated delay is played before or after
2767 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2768 sound tuned around the original one, like in a chorus where some vocals are slightly
2771 It accepts the following parameters:
2774 Set input gain. Default is 0.4.
2777 Set output gain. Default is 0.4.
2780 Set delays. A typical delay is around 40ms to 60ms.
2792 @subsection Examples
2798 chorus=0.7:0.9:55:0.4:0.25:2
2804 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2808 Fuller sounding chorus with three delays:
2810 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
2815 Compress or expand the audio's dynamic range.
2817 It accepts the following parameters:
2823 A list of times in seconds for each channel over which the instantaneous level
2824 of the input signal is averaged to determine its volume. @var{attacks} refers to
2825 increase of volume and @var{decays} refers to decrease of volume. For most
2826 situations, the attack time (response to the audio getting louder) should be
2827 shorter than the decay time, because the human ear is more sensitive to sudden
2828 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
2829 a typical value for decay is 0.8 seconds.
2830 If specified number of attacks & decays is lower than number of channels, the last
2831 set attack/decay will be used for all remaining channels.
2834 A list of points for the transfer function, specified in dB relative to the
2835 maximum possible signal amplitude. Each key points list must be defined using
2836 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
2837 @code{x0/y0 x1/y1 x2/y2 ....}
2839 The input values must be in strictly increasing order but the transfer function
2840 does not have to be monotonically rising. The point @code{0/0} is assumed but
2841 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
2842 function are @code{-70/-70|-60/-20|1/0}.
2845 Set the curve radius in dB for all joints. It defaults to 0.01.
2848 Set the additional gain in dB to be applied at all points on the transfer
2849 function. This allows for easy adjustment of the overall gain.
2853 Set an initial volume, in dB, to be assumed for each channel when filtering
2854 starts. This permits the user to supply a nominal level initially, so that, for
2855 example, a very large gain is not applied to initial signal levels before the
2856 companding has begun to operate. A typical value for audio which is initially
2857 quiet is -90 dB. It defaults to 0.
2860 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
2861 delayed before being fed to the volume adjuster. Specifying a delay
2862 approximately equal to the attack/decay times allows the filter to effectively
2863 operate in predictive rather than reactive mode. It defaults to 0.
2867 @subsection Examples
2871 Make music with both quiet and loud passages suitable for listening to in a
2874 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
2877 Another example for audio with whisper and explosion parts:
2879 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
2883 A noise gate for when the noise is at a lower level than the signal:
2885 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
2889 Here is another noise gate, this time for when the noise is at a higher level
2890 than the signal (making it, in some ways, similar to squelch):
2892 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
2896 2:1 compression starting at -6dB:
2898 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
2902 2:1 compression starting at -9dB:
2904 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
2908 2:1 compression starting at -12dB:
2910 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
2914 2:1 compression starting at -18dB:
2916 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
2920 3:1 compression starting at -15dB:
2922 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
2928 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
2934 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
2938 Hard limiter at -6dB:
2940 compand=attacks=0:points=-80/-80|-6/-6|20/-6
2944 Hard limiter at -12dB:
2946 compand=attacks=0:points=-80/-80|-12/-12|20/-12
2950 Hard noise gate at -35 dB:
2952 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
2958 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
2962 @section compensationdelay
2964 Compensation Delay Line is a metric based delay to compensate differing
2965 positions of microphones or speakers.
2967 For example, you have recorded guitar with two microphones placed in
2968 different locations. Because the front of sound wave has fixed speed in
2969 normal conditions, the phasing of microphones can vary and depends on
2970 their location and interposition. The best sound mix can be achieved when
2971 these microphones are in phase (synchronized). Note that a distance of
2972 ~30 cm between microphones makes one microphone capture the signal in
2973 antiphase to the other microphone. That makes the final mix sound moody.
2974 This filter helps to solve phasing problems by adding different delays
2975 to each microphone track and make them synchronized.
2977 The best result can be reached when you take one track as base and
2978 synchronize other tracks one by one with it.
2979 Remember that synchronization/delay tolerance depends on sample rate, too.
2980 Higher sample rates will give more tolerance.
2982 The filter accepts the following parameters:
2986 Set millimeters distance. This is compensation distance for fine tuning.
2990 Set cm distance. This is compensation distance for tightening distance setup.
2994 Set meters distance. This is compensation distance for hard distance setup.
2998 Set dry amount. Amount of unprocessed (dry) signal.
3002 Set wet amount. Amount of processed (wet) signal.
3006 Set temperature in degrees Celsius. This is the temperature of the environment.
3011 Apply headphone crossfeed filter.
3013 Crossfeed is the process of blending the left and right channels of stereo
3015 It is mainly used to reduce extreme stereo separation of low frequencies.
3017 The intent is to produce more speaker like sound to the listener.
3019 The filter accepts the following options:
3023 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3024 This sets gain of low shelf filter for side part of stereo image.
3025 Default is -6dB. Max allowed is -30db when strength is set to 1.
3028 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3029 This sets cut off frequency of low shelf filter. Default is cut off near
3030 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3033 Set input gain. Default is 0.9.
3036 Set output gain. Default is 1.
3039 @section crystalizer
3040 Simple algorithm to expand audio dynamic range.
3042 The filter accepts the following options:
3046 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3047 (unchanged sound) to 10.0 (maximum effect).
3050 Enable clipping. By default is enabled.
3054 Apply a DC shift to the audio.
3056 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3057 in the recording chain) from the audio. The effect of a DC offset is reduced
3058 headroom and hence volume. The @ref{astats} filter can be used to determine if
3059 a signal has a DC offset.
3063 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3067 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3068 used to prevent clipping.
3073 Apply de-essing to the audio samples.
3077 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3081 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3085 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3089 Set the output mode.
3091 It accepts the following values:
3094 Pass input unchanged.
3097 Pass ess filtered out.
3102 Default value is @var{o}.
3108 Measure audio dynamic range.
3110 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3111 is found in transition material. And anything less that 8 have very poor dynamics
3112 and is very compressed.
3114 The filter accepts the following options:
3118 Set window length in seconds used to split audio into segments of equal length.
3119 Default is 3 seconds.
3123 Dynamic Audio Normalizer.
3125 This filter applies a certain amount of gain to the input audio in order
3126 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3127 contrast to more "simple" normalization algorithms, the Dynamic Audio
3128 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3129 This allows for applying extra gain to the "quiet" sections of the audio
3130 while avoiding distortions or clipping the "loud" sections. In other words:
3131 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3132 sections, in the sense that the volume of each section is brought to the
3133 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3134 this goal *without* applying "dynamic range compressing". It will retain 100%
3135 of the dynamic range *within* each section of the audio file.
3139 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3140 Default is 500 milliseconds.
3141 The Dynamic Audio Normalizer processes the input audio in small chunks,
3142 referred to as frames. This is required, because a peak magnitude has no
3143 meaning for just a single sample value. Instead, we need to determine the
3144 peak magnitude for a contiguous sequence of sample values. While a "standard"
3145 normalizer would simply use the peak magnitude of the complete file, the
3146 Dynamic Audio Normalizer determines the peak magnitude individually for each
3147 frame. The length of a frame is specified in milliseconds. By default, the
3148 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3149 been found to give good results with most files.
3150 Note that the exact frame length, in number of samples, will be determined
3151 automatically, based on the sampling rate of the individual input audio file.
3154 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3155 number. Default is 31.
3156 Probably the most important parameter of the Dynamic Audio Normalizer is the
3157 @code{window size} of the Gaussian smoothing filter. The filter's window size
3158 is specified in frames, centered around the current frame. For the sake of
3159 simplicity, this must be an odd number. Consequently, the default value of 31
3160 takes into account the current frame, as well as the 15 preceding frames and
3161 the 15 subsequent frames. Using a larger window results in a stronger
3162 smoothing effect and thus in less gain variation, i.e. slower gain
3163 adaptation. Conversely, using a smaller window results in a weaker smoothing
3164 effect and thus in more gain variation, i.e. faster gain adaptation.
3165 In other words, the more you increase this value, the more the Dynamic Audio
3166 Normalizer will behave like a "traditional" normalization filter. On the
3167 contrary, the more you decrease this value, the more the Dynamic Audio
3168 Normalizer will behave like a dynamic range compressor.
3171 Set the target peak value. This specifies the highest permissible magnitude
3172 level for the normalized audio input. This filter will try to approach the
3173 target peak magnitude as closely as possible, but at the same time it also
3174 makes sure that the normalized signal will never exceed the peak magnitude.
3175 A frame's maximum local gain factor is imposed directly by the target peak
3176 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3177 It is not recommended to go above this value.
3180 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3181 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3182 factor for each input frame, i.e. the maximum gain factor that does not
3183 result in clipping or distortion. The maximum gain factor is determined by
3184 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3185 additionally bounds the frame's maximum gain factor by a predetermined
3186 (global) maximum gain factor. This is done in order to avoid excessive gain
3187 factors in "silent" or almost silent frames. By default, the maximum gain
3188 factor is 10.0, For most inputs the default value should be sufficient and
3189 it usually is not recommended to increase this value. Though, for input
3190 with an extremely low overall volume level, it may be necessary to allow even
3191 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3192 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3193 Instead, a "sigmoid" threshold function will be applied. This way, the
3194 gain factors will smoothly approach the threshold value, but never exceed that
3198 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3199 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3200 This means that the maximum local gain factor for each frame is defined
3201 (only) by the frame's highest magnitude sample. This way, the samples can
3202 be amplified as much as possible without exceeding the maximum signal
3203 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3204 Normalizer can also take into account the frame's root mean square,
3205 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3206 determine the power of a time-varying signal. It is therefore considered
3207 that the RMS is a better approximation of the "perceived loudness" than
3208 just looking at the signal's peak magnitude. Consequently, by adjusting all
3209 frames to a constant RMS value, a uniform "perceived loudness" can be
3210 established. If a target RMS value has been specified, a frame's local gain
3211 factor is defined as the factor that would result in exactly that RMS value.
3212 Note, however, that the maximum local gain factor is still restricted by the
3213 frame's highest magnitude sample, in order to prevent clipping.
3216 Enable channels coupling. By default is enabled.
3217 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3218 amount. This means the same gain factor will be applied to all channels, i.e.
3219 the maximum possible gain factor is determined by the "loudest" channel.
3220 However, in some recordings, it may happen that the volume of the different
3221 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3222 In this case, this option can be used to disable the channel coupling. This way,
3223 the gain factor will be determined independently for each channel, depending
3224 only on the individual channel's highest magnitude sample. This allows for
3225 harmonizing the volume of the different channels.
3228 Enable DC bias correction. By default is disabled.
3229 An audio signal (in the time domain) is a sequence of sample values.
3230 In the Dynamic Audio Normalizer these sample values are represented in the
3231 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3232 audio signal, or "waveform", should be centered around the zero point.
3233 That means if we calculate the mean value of all samples in a file, or in a
3234 single frame, then the result should be 0.0 or at least very close to that
3235 value. If, however, there is a significant deviation of the mean value from
3236 0.0, in either positive or negative direction, this is referred to as a
3237 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3238 Audio Normalizer provides optional DC bias correction.
3239 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3240 the mean value, or "DC correction" offset, of each input frame and subtract
3241 that value from all of the frame's sample values which ensures those samples
3242 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3243 boundaries, the DC correction offset values will be interpolated smoothly
3244 between neighbouring frames.
3246 @item altboundary, b
3247 Enable alternative boundary mode. By default is disabled.
3248 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3249 around each frame. This includes the preceding frames as well as the
3250 subsequent frames. However, for the "boundary" frames, located at the very
3251 beginning and at the very end of the audio file, not all neighbouring
3252 frames are available. In particular, for the first few frames in the audio
3253 file, the preceding frames are not known. And, similarly, for the last few
3254 frames in the audio file, the subsequent frames are not known. Thus, the
3255 question arises which gain factors should be assumed for the missing frames
3256 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3257 to deal with this situation. The default boundary mode assumes a gain factor
3258 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3259 "fade out" at the beginning and at the end of the input, respectively.
3262 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3263 By default, the Dynamic Audio Normalizer does not apply "traditional"
3264 compression. This means that signal peaks will not be pruned and thus the
3265 full dynamic range will be retained within each local neighbourhood. However,
3266 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3267 normalization algorithm with a more "traditional" compression.
3268 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3269 (thresholding) function. If (and only if) the compression feature is enabled,
3270 all input frames will be processed by a soft knee thresholding function prior
3271 to the actual normalization process. Put simply, the thresholding function is
3272 going to prune all samples whose magnitude exceeds a certain threshold value.
3273 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3274 value. Instead, the threshold value will be adjusted for each individual
3276 In general, smaller parameters result in stronger compression, and vice versa.
3277 Values below 3.0 are not recommended, because audible distortion may appear.
3282 Make audio easier to listen to on headphones.
3284 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3285 so that when listened to on headphones the stereo image is moved from
3286 inside your head (standard for headphones) to outside and in front of
3287 the listener (standard for speakers).
3293 Apply a two-pole peaking equalisation (EQ) filter. With this
3294 filter, the signal-level at and around a selected frequency can
3295 be increased or decreased, whilst (unlike bandpass and bandreject
3296 filters) that at all other frequencies is unchanged.
3298 In order to produce complex equalisation curves, this filter can
3299 be given several times, each with a different central frequency.
3301 The filter accepts the following options:
3305 Set the filter's central frequency in Hz.
3308 Set method to specify band-width of filter.
3323 Specify the band-width of a filter in width_type units.
3326 Set the required gain or attenuation in dB.
3327 Beware of clipping when using a positive gain.
3330 How much to use filtered signal in output. Default is 1.
3331 Range is between 0 and 1.
3334 Specify which channels to filter, by default all available are filtered.
3337 @subsection Examples
3340 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3342 equalizer=f=1000:t=h:width=200:g=-10
3346 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3348 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3352 @subsection Commands
3354 This filter supports the following commands:
3357 Change equalizer frequency.
3358 Syntax for the command is : "@var{frequency}"
3361 Change equalizer width_type.
3362 Syntax for the command is : "@var{width_type}"
3365 Change equalizer width.
3366 Syntax for the command is : "@var{width}"
3369 Change equalizer gain.
3370 Syntax for the command is : "@var{gain}"
3373 Change equalizer mix.
3374 Syntax for the command is : "@var{mix}"
3377 @section extrastereo
3379 Linearly increases the difference between left and right channels which
3380 adds some sort of "live" effect to playback.
3382 The filter accepts the following options:
3386 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3387 (average of both channels), with 1.0 sound will be unchanged, with
3388 -1.0 left and right channels will be swapped.
3391 Enable clipping. By default is enabled.
3394 @section firequalizer
3395 Apply FIR Equalization using arbitrary frequency response.
3397 The filter accepts the following option:
3401 Set gain curve equation (in dB). The expression can contain variables:
3404 the evaluated frequency
3408 channel number, set to 0 when multichannels evaluation is disabled
3410 channel id, see libavutil/channel_layout.h, set to the first channel id when
3411 multichannels evaluation is disabled
3415 channel_layout, see libavutil/channel_layout.h
3420 @item gain_interpolate(f)
3421 interpolate gain on frequency f based on gain_entry
3422 @item cubic_interpolate(f)
3423 same as gain_interpolate, but smoother
3425 This option is also available as command. Default is @code{gain_interpolate(f)}.
3428 Set gain entry for gain_interpolate function. The expression can
3432 store gain entry at frequency f with value g
3434 This option is also available as command.
3437 Set filter delay in seconds. Higher value means more accurate.
3438 Default is @code{0.01}.
3441 Set filter accuracy in Hz. Lower value means more accurate.
3442 Default is @code{5}.
3445 Set window function. Acceptable values are:
3448 rectangular window, useful when gain curve is already smooth
3450 hann window (default)
3456 3-terms continuous 1st derivative nuttall window
3458 minimum 3-terms discontinuous nuttall window
3460 4-terms continuous 1st derivative nuttall window
3462 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3464 blackman-harris window
3470 If enabled, use fixed number of audio samples. This improves speed when
3471 filtering with large delay. Default is disabled.
3474 Enable multichannels evaluation on gain. Default is disabled.
3477 Enable zero phase mode by subtracting timestamp to compensate delay.
3478 Default is disabled.
3481 Set scale used by gain. Acceptable values are:
3484 linear frequency, linear gain
3486 linear frequency, logarithmic (in dB) gain (default)
3488 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3490 logarithmic frequency, logarithmic gain
3494 Set file for dumping, suitable for gnuplot.
3497 Set scale for dumpfile. Acceptable values are same with scale option.
3501 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3502 Default is disabled.
3505 Enable minimum phase impulse response. Default is disabled.
3508 @subsection Examples
3513 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3516 lowpass at 1000 Hz with gain_entry:
3518 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3521 custom equalization:
3523 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3526 higher delay with zero phase to compensate delay:
3528 firequalizer=delay=0.1:fixed=on:zero_phase=on
3531 lowpass on left channel, highpass on right channel:
3533 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3534 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3539 Apply a flanging effect to the audio.
3541 The filter accepts the following options:
3545 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3548 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3551 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3555 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3556 Default value is 71.
3559 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3562 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3563 Default value is @var{sinusoidal}.
3566 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3567 Default value is 25.
3570 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3571 Default is @var{linear}.
3575 Apply Haas effect to audio.
3577 Note that this makes most sense to apply on mono signals.
3578 With this filter applied to mono signals it give some directionality and
3579 stretches its stereo image.
3581 The filter accepts the following options:
3585 Set input level. By default is @var{1}, or 0dB
3588 Set output level. By default is @var{1}, or 0dB.
3591 Set gain applied to side part of signal. By default is @var{1}.
3594 Set kind of middle source. Can be one of the following:
3604 Pick middle part signal of stereo image.
3607 Pick side part signal of stereo image.
3611 Change middle phase. By default is disabled.
3614 Set left channel delay. By default is @var{2.05} milliseconds.
3617 Set left channel balance. By default is @var{-1}.
3620 Set left channel gain. By default is @var{1}.
3623 Change left phase. By default is disabled.
3626 Set right channel delay. By defaults is @var{2.12} milliseconds.
3629 Set right channel balance. By default is @var{1}.
3632 Set right channel gain. By default is @var{1}.
3635 Change right phase. By default is enabled.
3640 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3641 embedded HDCD codes is expanded into a 20-bit PCM stream.
3643 The filter supports the Peak Extend and Low-level Gain Adjustment features
3644 of HDCD, and detects the Transient Filter flag.
3647 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3650 When using the filter with wav, note the default encoding for wav is 16-bit,
3651 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3652 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3654 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3655 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3658 The filter accepts the following options:
3661 @item disable_autoconvert
3662 Disable any automatic format conversion or resampling in the filter graph.
3664 @item process_stereo
3665 Process the stereo channels together. If target_gain does not match between
3666 channels, consider it invalid and use the last valid target_gain.
3669 Set the code detect timer period in ms.
3672 Always extend peaks above -3dBFS even if PE isn't signaled.
3675 Replace audio with a solid tone and adjust the amplitude to signal some
3676 specific aspect of the decoding process. The output file can be loaded in
3677 an audio editor alongside the original to aid analysis.
3679 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3686 Gain adjustment level at each sample
3688 Samples where peak extend occurs
3690 Samples where the code detect timer is active
3692 Samples where the target gain does not match between channels
3698 Apply head-related transfer functions (HRTFs) to create virtual
3699 loudspeakers around the user for binaural listening via headphones.
3700 The HRIRs are provided via additional streams, for each channel
3701 one stereo input stream is needed.
3703 The filter accepts the following options:
3707 Set mapping of input streams for convolution.
3708 The argument is a '|'-separated list of channel names in order as they
3709 are given as additional stream inputs for filter.
3710 This also specify number of input streams. Number of input streams
3711 must be not less than number of channels in first stream plus one.
3714 Set gain applied to audio. Value is in dB. Default is 0.
3717 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3718 processing audio in time domain which is slow.
3719 @var{freq} is processing audio in frequency domain which is fast.
3720 Default is @var{freq}.
3723 Set custom gain for LFE channels. Value is in dB. Default is 0.
3726 Set size of frame in number of samples which will be processed at once.
3727 Default value is @var{1024}. Allowed range is from 1024 to 96000.
3730 Set format of hrir stream.
3731 Default value is @var{stereo}. Alternative value is @var{multich}.
3732 If value is set to @var{stereo}, number of additional streams should
3733 be greater or equal to number of input channels in first input stream.
3734 Also each additional stream should have stereo number of channels.
3735 If value is set to @var{multich}, number of additional streams should
3736 be exactly one. Also number of input channels of additional stream
3737 should be equal or greater than twice number of channels of first input
3741 @subsection Examples
3745 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3746 each amovie filter use stereo file with IR coefficients as input.
3747 The files give coefficients for each position of virtual loudspeaker:
3750 -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"
3755 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3756 but now in @var{multich} @var{hrir} format.
3758 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"
3765 Apply a high-pass filter with 3dB point frequency.
3766 The filter can be either single-pole, or double-pole (the default).
3767 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3769 The filter accepts the following options:
3773 Set frequency in Hz. Default is 3000.
3776 Set number of poles. Default is 2.
3779 Set method to specify band-width of filter.
3794 Specify the band-width of a filter in width_type units.
3795 Applies only to double-pole filter.
3796 The default is 0.707q and gives a Butterworth response.
3799 How much to use filtered signal in output. Default is 1.
3800 Range is between 0 and 1.
3803 Specify which channels to filter, by default all available are filtered.
3806 @subsection Commands
3808 This filter supports the following commands:
3811 Change highpass frequency.
3812 Syntax for the command is : "@var{frequency}"
3815 Change highpass width_type.
3816 Syntax for the command is : "@var{width_type}"
3819 Change highpass width.
3820 Syntax for the command is : "@var{width}"
3823 Change highpass mix.
3824 Syntax for the command is : "@var{mix}"
3829 Join multiple input streams into one multi-channel stream.
3831 It accepts the following parameters:
3835 The number of input streams. It defaults to 2.
3837 @item channel_layout
3838 The desired output channel layout. It defaults to stereo.
3841 Map channels from inputs to output. The argument is a '|'-separated list of
3842 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
3843 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
3844 can be either the name of the input channel (e.g. FL for front left) or its
3845 index in the specified input stream. @var{out_channel} is the name of the output
3849 The filter will attempt to guess the mappings when they are not specified
3850 explicitly. It does so by first trying to find an unused matching input channel
3851 and if that fails it picks the first unused input channel.
3853 Join 3 inputs (with properly set channel layouts):
3855 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
3858 Build a 5.1 output from 6 single-channel streams:
3860 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
3861 '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'
3867 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
3869 To enable compilation of this filter you need to configure FFmpeg with
3870 @code{--enable-ladspa}.
3874 Specifies the name of LADSPA plugin library to load. If the environment
3875 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
3876 each one of the directories specified by the colon separated list in
3877 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
3878 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
3879 @file{/usr/lib/ladspa/}.
3882 Specifies the plugin within the library. Some libraries contain only
3883 one plugin, but others contain many of them. If this is not set filter
3884 will list all available plugins within the specified library.
3887 Set the '|' separated list of controls which are zero or more floating point
3888 values that determine the behavior of the loaded plugin (for example delay,
3890 Controls need to be defined using the following syntax:
3891 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
3892 @var{valuei} is the value set on the @var{i}-th control.
3893 Alternatively they can be also defined using the following syntax:
3894 @var{value0}|@var{value1}|@var{value2}|..., where
3895 @var{valuei} is the value set on the @var{i}-th control.
3896 If @option{controls} is set to @code{help}, all available controls and
3897 their valid ranges are printed.
3899 @item sample_rate, s
3900 Specify the sample rate, default to 44100. Only used if plugin have
3904 Set the number of samples per channel per each output frame, default
3905 is 1024. Only used if plugin have zero inputs.
3908 Set the minimum duration of the sourced audio. See
3909 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
3910 for the accepted syntax.
3911 Note that the resulting duration may be greater than the specified duration,
3912 as the generated audio is always cut at the end of a complete frame.
3913 If not specified, or the expressed duration is negative, the audio is
3914 supposed to be generated forever.
3915 Only used if plugin have zero inputs.
3919 @subsection Examples
3923 List all available plugins within amp (LADSPA example plugin) library:
3929 List all available controls and their valid ranges for @code{vcf_notch}
3930 plugin from @code{VCF} library:
3932 ladspa=f=vcf:p=vcf_notch:c=help
3936 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
3939 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
3943 Add reverberation to the audio using TAP-plugins
3944 (Tom's Audio Processing plugins):
3946 ladspa=file=tap_reverb:tap_reverb
3950 Generate white noise, with 0.2 amplitude:
3952 ladspa=file=cmt:noise_source_white:c=c0=.2
3956 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
3957 @code{C* Audio Plugin Suite} (CAPS) library:
3959 ladspa=file=caps:Click:c=c1=20'
3963 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
3965 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
3969 Increase volume by 20dB using fast lookahead limiter from Steve Harris
3970 @code{SWH Plugins} collection:
3972 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
3976 Attenuate low frequencies using Multiband EQ from Steve Harris
3977 @code{SWH Plugins} collection:
3979 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
3983 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
3986 ladspa=caps:Narrower
3990 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
3992 ladspa=caps:White:.2
3996 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
3998 ladspa=caps:Fractal:c=c1=1
4002 Dynamic volume normalization using @code{VLevel} plugin:
4004 ladspa=vlevel-ladspa:vlevel_mono
4008 @subsection Commands
4010 This filter supports the following commands:
4013 Modify the @var{N}-th control value.
4015 If the specified value is not valid, it is ignored and prior one is kept.
4020 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4021 Support for both single pass (livestreams, files) and double pass (files) modes.
4022 This algorithm can target IL, LRA, and maximum true peak. To accurately detect true peaks,
4023 the audio stream will be upsampled to 192 kHz unless the normalization mode is linear.
4024 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4026 The filter accepts the following options:
4030 Set integrated loudness target.
4031 Range is -70.0 - -5.0. Default value is -24.0.
4034 Set loudness range target.
4035 Range is 1.0 - 20.0. Default value is 7.0.
4038 Set maximum true peak.
4039 Range is -9.0 - +0.0. Default value is -2.0.
4041 @item measured_I, measured_i
4042 Measured IL of input file.
4043 Range is -99.0 - +0.0.
4045 @item measured_LRA, measured_lra
4046 Measured LRA of input file.
4047 Range is 0.0 - 99.0.
4049 @item measured_TP, measured_tp
4050 Measured true peak of input file.
4051 Range is -99.0 - +99.0.
4053 @item measured_thresh
4054 Measured threshold of input file.
4055 Range is -99.0 - +0.0.
4058 Set offset gain. Gain is applied before the true-peak limiter.
4059 Range is -99.0 - +99.0. Default is +0.0.
4062 Normalize linearly if possible.
4063 measured_I, measured_LRA, measured_TP, and measured_thresh must also
4064 to be specified in order to use this mode.
4065 Options are true or false. Default is true.
4068 Treat mono input files as "dual-mono". If a mono file is intended for playback
4069 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4070 If set to @code{true}, this option will compensate for this effect.
4071 Multi-channel input files are not affected by this option.
4072 Options are true or false. Default is false.
4075 Set print format for stats. Options are summary, json, or none.
4076 Default value is none.
4081 Apply a low-pass filter with 3dB point frequency.
4082 The filter can be either single-pole or double-pole (the default).
4083 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4085 The filter accepts the following options:
4089 Set frequency in Hz. Default is 500.
4092 Set number of poles. Default is 2.
4095 Set method to specify band-width of filter.
4110 Specify the band-width of a filter in width_type units.
4111 Applies only to double-pole filter.
4112 The default is 0.707q and gives a Butterworth response.
4115 How much to use filtered signal in output. Default is 1.
4116 Range is between 0 and 1.
4119 Specify which channels to filter, by default all available are filtered.
4122 @subsection Examples
4125 Lowpass only LFE channel, it LFE is not present it does nothing:
4131 @subsection Commands
4133 This filter supports the following commands:
4136 Change lowpass frequency.
4137 Syntax for the command is : "@var{frequency}"
4140 Change lowpass width_type.
4141 Syntax for the command is : "@var{width_type}"
4144 Change lowpass width.
4145 Syntax for the command is : "@var{width}"
4149 Syntax for the command is : "@var{mix}"
4154 Load a LV2 (LADSPA Version 2) plugin.
4156 To enable compilation of this filter you need to configure FFmpeg with
4157 @code{--enable-lv2}.
4161 Specifies the plugin URI. You may need to escape ':'.
4164 Set the '|' separated list of controls which are zero or more floating point
4165 values that determine the behavior of the loaded plugin (for example delay,
4167 If @option{controls} is set to @code{help}, all available controls and
4168 their valid ranges are printed.
4170 @item sample_rate, s
4171 Specify the sample rate, default to 44100. Only used if plugin have
4175 Set the number of samples per channel per each output frame, default
4176 is 1024. Only used if plugin have zero inputs.
4179 Set the minimum duration of the sourced audio. See
4180 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4181 for the accepted syntax.
4182 Note that the resulting duration may be greater than the specified duration,
4183 as the generated audio is always cut at the end of a complete frame.
4184 If not specified, or the expressed duration is negative, the audio is
4185 supposed to be generated forever.
4186 Only used if plugin have zero inputs.
4189 @subsection Examples
4193 Apply bass enhancer plugin from Calf:
4195 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4199 Apply vinyl plugin from Calf:
4201 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4205 Apply bit crusher plugin from ArtyFX:
4207 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4212 Multiband Compress or expand the audio's dynamic range.
4214 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4215 This is akin to the crossover of a loudspeaker, and results in flat frequency
4216 response when absent compander action.
4218 It accepts the following parameters:
4222 This option syntax is:
4223 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4224 For explanation of each item refer to compand filter documentation.
4230 Mix channels with specific gain levels. The filter accepts the output
4231 channel layout followed by a set of channels definitions.
4233 This filter is also designed to efficiently remap the channels of an audio
4236 The filter accepts parameters of the form:
4237 "@var{l}|@var{outdef}|@var{outdef}|..."
4241 output channel layout or number of channels
4244 output channel specification, of the form:
4245 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4248 output channel to define, either a channel name (FL, FR, etc.) or a channel
4249 number (c0, c1, etc.)
4252 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4255 input channel to use, see out_name for details; it is not possible to mix
4256 named and numbered input channels
4259 If the `=' in a channel specification is replaced by `<', then the gains for
4260 that specification will be renormalized so that the total is 1, thus
4261 avoiding clipping noise.
4263 @subsection Mixing examples
4265 For example, if you want to down-mix from stereo to mono, but with a bigger
4266 factor for the left channel:
4268 pan=1c|c0=0.9*c0+0.1*c1
4271 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4272 7-channels surround:
4274 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4277 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4278 that should be preferred (see "-ac" option) unless you have very specific
4281 @subsection Remapping examples
4283 The channel remapping will be effective if, and only if:
4286 @item gain coefficients are zeroes or ones,
4287 @item only one input per channel output,
4290 If all these conditions are satisfied, the filter will notify the user ("Pure
4291 channel mapping detected"), and use an optimized and lossless method to do the
4294 For example, if you have a 5.1 source and want a stereo audio stream by
4295 dropping the extra channels:
4297 pan="stereo| c0=FL | c1=FR"
4300 Given the same source, you can also switch front left and front right channels
4301 and keep the input channel layout:
4303 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4306 If the input is a stereo audio stream, you can mute the front left channel (and
4307 still keep the stereo channel layout) with:
4312 Still with a stereo audio stream input, you can copy the right channel in both
4313 front left and right:
4315 pan="stereo| c0=FR | c1=FR"
4320 ReplayGain scanner filter. This filter takes an audio stream as an input and
4321 outputs it unchanged.
4322 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4326 Convert the audio sample format, sample rate and channel layout. It is
4327 not meant to be used directly.
4330 Apply time-stretching and pitch-shifting with librubberband.
4332 To enable compilation of this filter, you need to configure FFmpeg with
4333 @code{--enable-librubberband}.
4335 The filter accepts the following options:
4339 Set tempo scale factor.
4342 Set pitch scale factor.
4345 Set transients detector.
4346 Possible values are:
4355 Possible values are:
4364 Possible values are:
4371 Set processing window size.
4372 Possible values are:
4381 Possible values are:
4388 Enable formant preservation when shift pitching.
4389 Possible values are:
4397 Possible values are:
4406 Possible values are:
4413 @section sidechaincompress
4415 This filter acts like normal compressor but has the ability to compress
4416 detected signal using second input signal.
4417 It needs two input streams and returns one output stream.
4418 First input stream will be processed depending on second stream signal.
4419 The filtered signal then can be filtered with other filters in later stages of
4420 processing. See @ref{pan} and @ref{amerge} filter.
4422 The filter accepts the following options:
4426 Set input gain. Default is 1. Range is between 0.015625 and 64.
4429 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4430 Default is @code{downward}.
4433 If a signal of second stream raises above this level it will affect the gain
4434 reduction of first stream.
4435 By default is 0.125. Range is between 0.00097563 and 1.
4438 Set a ratio about which the signal is reduced. 1:2 means that if the level
4439 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4440 Default is 2. Range is between 1 and 20.
4443 Amount of milliseconds the signal has to rise above the threshold before gain
4444 reduction starts. Default is 20. Range is between 0.01 and 2000.
4447 Amount of milliseconds the signal has to fall below the threshold before
4448 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4451 Set the amount by how much signal will be amplified after processing.
4452 Default is 1. Range is from 1 to 64.
4455 Curve the sharp knee around the threshold to enter gain reduction more softly.
4456 Default is 2.82843. Range is between 1 and 8.
4459 Choose if the @code{average} level between all channels of side-chain stream
4460 or the louder(@code{maximum}) channel of side-chain stream affects the
4461 reduction. Default is @code{average}.
4464 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4465 of @code{rms}. Default is @code{rms} which is mainly smoother.
4468 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4471 How much to use compressed signal in output. Default is 1.
4472 Range is between 0 and 1.
4475 @subsection Examples
4479 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4480 depending on the signal of 2nd input and later compressed signal to be
4481 merged with 2nd input:
4483 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4487 @section sidechaingate
4489 A sidechain gate acts like a normal (wideband) gate but has the ability to
4490 filter the detected signal before sending it to the gain reduction stage.
4491 Normally a gate uses the full range signal to detect a level above the
4493 For example: If you cut all lower frequencies from your sidechain signal
4494 the gate will decrease the volume of your track only if not enough highs
4495 appear. With this technique you are able to reduce the resonation of a
4496 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4498 It needs two input streams and returns one output stream.
4499 First input stream will be processed depending on second stream signal.
4501 The filter accepts the following options:
4505 Set input level before filtering.
4506 Default is 1. Allowed range is from 0.015625 to 64.
4509 Set the mode of operation. Can be @code{upward} or @code{downward}.
4510 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
4511 will be amplified, expanding dynamic range in upward direction.
4512 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
4515 Set the level of gain reduction when the signal is below the threshold.
4516 Default is 0.06125. Allowed range is from 0 to 1.
4517 Setting this to 0 disables reduction and then filter behaves like expander.
4520 If a signal rises above this level the gain reduction is released.
4521 Default is 0.125. Allowed range is from 0 to 1.
4524 Set a ratio about which the signal is reduced.
4525 Default is 2. Allowed range is from 1 to 9000.
4528 Amount of milliseconds the signal has to rise above the threshold before gain
4530 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4533 Amount of milliseconds the signal has to fall below the threshold before the
4534 reduction is increased again. Default is 250 milliseconds.
4535 Allowed range is from 0.01 to 9000.
4538 Set amount of amplification of signal after processing.
4539 Default is 1. Allowed range is from 1 to 64.
4542 Curve the sharp knee around the threshold to enter gain reduction more softly.
4543 Default is 2.828427125. Allowed range is from 1 to 8.
4546 Choose if exact signal should be taken for detection or an RMS like one.
4547 Default is rms. Can be peak or rms.
4550 Choose if the average level between all channels or the louder channel affects
4552 Default is average. Can be average or maximum.
4555 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4558 @section silencedetect
4560 Detect silence in an audio stream.
4562 This filter logs a message when it detects that the input audio volume is less
4563 or equal to a noise tolerance value for a duration greater or equal to the
4564 minimum detected noise duration.
4566 The printed times and duration are expressed in seconds.
4568 The filter accepts the following options:
4572 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4573 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4576 Set silence duration until notification (default is 2 seconds).
4579 Process each channel separately, instead of combined. By default is disabled.
4582 @subsection Examples
4586 Detect 5 seconds of silence with -50dB noise tolerance:
4588 silencedetect=n=-50dB:d=5
4592 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4593 tolerance in @file{silence.mp3}:
4595 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4599 @section silenceremove
4601 Remove silence from the beginning, middle or end of the audio.
4603 The filter accepts the following options:
4607 This value is used to indicate if audio should be trimmed at beginning of
4608 the audio. A value of zero indicates no silence should be trimmed from the
4609 beginning. When specifying a non-zero value, it trims audio up until it
4610 finds non-silence. Normally, when trimming silence from beginning of audio
4611 the @var{start_periods} will be @code{1} but it can be increased to higher
4612 values to trim all audio up to specific count of non-silence periods.
4613 Default value is @code{0}.
4615 @item start_duration
4616 Specify the amount of time that non-silence must be detected before it stops
4617 trimming audio. By increasing the duration, bursts of noises can be treated
4618 as silence and trimmed off. Default value is @code{0}.
4620 @item start_threshold
4621 This indicates what sample value should be treated as silence. For digital
4622 audio, a value of @code{0} may be fine but for audio recorded from analog,
4623 you may wish to increase the value to account for background noise.
4624 Can be specified in dB (in case "dB" is appended to the specified value)
4625 or amplitude ratio. Default value is @code{0}.
4628 Specify max duration of silence at beginning that will be kept after
4629 trimming. Default is 0, which is equal to trimming all samples detected
4633 Specify mode of detection of silence end in start of multi-channel audio.
4634 Can be @var{any} or @var{all}. Default is @var{any}.
4635 With @var{any}, any sample that is detected as non-silence will cause
4636 stopped trimming of silence.
4637 With @var{all}, only if all channels are detected as non-silence will cause
4638 stopped trimming of silence.
4641 Set the count for trimming silence from the end of audio.
4642 To remove silence from the middle of a file, specify a @var{stop_periods}
4643 that is negative. This value is then treated as a positive value and is
4644 used to indicate the effect should restart processing as specified by
4645 @var{start_periods}, making it suitable for removing periods of silence
4646 in the middle of the audio.
4647 Default value is @code{0}.
4650 Specify a duration of silence that must exist before audio is not copied any
4651 more. By specifying a higher duration, silence that is wanted can be left in
4653 Default value is @code{0}.
4655 @item stop_threshold
4656 This is the same as @option{start_threshold} but for trimming silence from
4658 Can be specified in dB (in case "dB" is appended to the specified value)
4659 or amplitude ratio. Default value is @code{0}.
4662 Specify max duration of silence at end that will be kept after
4663 trimming. Default is 0, which is equal to trimming all samples detected
4667 Specify mode of detection of silence start in end of multi-channel audio.
4668 Can be @var{any} or @var{all}. Default is @var{any}.
4669 With @var{any}, any sample that is detected as non-silence will cause
4670 stopped trimming of silence.
4671 With @var{all}, only if all channels are detected as non-silence will cause
4672 stopped trimming of silence.
4675 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4676 and works better with digital silence which is exactly 0.
4677 Default value is @code{rms}.
4680 Set duration in number of seconds used to calculate size of window in number
4681 of samples for detecting silence.
4682 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4685 @subsection Examples
4689 The following example shows how this filter can be used to start a recording
4690 that does not contain the delay at the start which usually occurs between
4691 pressing the record button and the start of the performance:
4693 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
4697 Trim all silence encountered from beginning to end where there is more than 1
4698 second of silence in audio:
4700 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
4704 Trim all digital silence samples, using peak detection, from beginning to end
4705 where there is more than 0 samples of digital silence in audio and digital
4706 silence is detected in all channels at same positions in stream:
4708 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
4714 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4715 loudspeakers around the user for binaural listening via headphones (audio
4716 formats up to 9 channels supported).
4717 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4718 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4719 Austrian Academy of Sciences.
4721 To enable compilation of this filter you need to configure FFmpeg with
4722 @code{--enable-libmysofa}.
4724 The filter accepts the following options:
4728 Set the SOFA file used for rendering.
4731 Set gain applied to audio. Value is in dB. Default is 0.
4734 Set rotation of virtual loudspeakers in deg. Default is 0.
4737 Set elevation of virtual speakers in deg. Default is 0.
4740 Set distance in meters between loudspeakers and the listener with near-field
4741 HRTFs. Default is 1.
4744 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4745 processing audio in time domain which is slow.
4746 @var{freq} is processing audio in frequency domain which is fast.
4747 Default is @var{freq}.
4750 Set custom positions of virtual loudspeakers. Syntax for this option is:
4751 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4752 Each virtual loudspeaker is described with short channel name following with
4753 azimuth and elevation in degrees.
4754 Each virtual loudspeaker description is separated by '|'.
4755 For example to override front left and front right channel positions use:
4756 'speakers=FL 45 15|FR 345 15'.
4757 Descriptions with unrecognised channel names are ignored.
4760 Set custom gain for LFE channels. Value is in dB. Default is 0.
4763 Set custom frame size in number of samples. Default is 1024.
4764 Allowed range is from 1024 to 96000. Only used if option @samp{type}
4765 is set to @var{freq}.
4768 Should all IRs be normalized upon importing SOFA file.
4769 By default is enabled.
4772 Should nearest IRs be interpolated with neighbor IRs if exact position
4773 does not match. By default is disabled.
4776 Minphase all IRs upon loading of SOFA file. By default is disabled.
4779 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
4782 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
4785 @subsection Examples
4789 Using ClubFritz6 sofa file:
4791 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
4795 Using ClubFritz12 sofa file and bigger radius with small rotation:
4797 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
4801 Similar as above but with custom speaker positions for front left, front right, back left and back right
4802 and also with custom gain:
4804 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
4808 @section stereotools
4810 This filter has some handy utilities to manage stereo signals, for converting
4811 M/S stereo recordings to L/R signal while having control over the parameters
4812 or spreading the stereo image of master track.
4814 The filter accepts the following options:
4818 Set input level before filtering for both channels. Defaults is 1.
4819 Allowed range is from 0.015625 to 64.
4822 Set output level after filtering for both channels. Defaults is 1.
4823 Allowed range is from 0.015625 to 64.
4826 Set input balance between both channels. Default is 0.
4827 Allowed range is from -1 to 1.
4830 Set output balance between both channels. Default is 0.
4831 Allowed range is from -1 to 1.
4834 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
4835 clipping. Disabled by default.
4838 Mute the left channel. Disabled by default.
4841 Mute the right channel. Disabled by default.
4844 Change the phase of the left channel. Disabled by default.
4847 Change the phase of the right channel. Disabled by default.
4850 Set stereo mode. Available values are:
4854 Left/Right to Left/Right, this is default.
4857 Left/Right to Mid/Side.
4860 Mid/Side to Left/Right.
4863 Left/Right to Left/Left.
4866 Left/Right to Right/Right.
4869 Left/Right to Left + Right.
4872 Left/Right to Right/Left.
4875 Mid/Side to Left/Left.
4878 Mid/Side to Right/Right.
4882 Set level of side signal. Default is 1.
4883 Allowed range is from 0.015625 to 64.
4886 Set balance of side signal. Default is 0.
4887 Allowed range is from -1 to 1.
4890 Set level of the middle signal. Default is 1.
4891 Allowed range is from 0.015625 to 64.
4894 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
4897 Set stereo base between mono and inversed channels. Default is 0.
4898 Allowed range is from -1 to 1.
4901 Set delay in milliseconds how much to delay left from right channel and
4902 vice versa. Default is 0. Allowed range is from -20 to 20.
4905 Set S/C level. Default is 1. Allowed range is from 1 to 100.
4908 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
4910 @item bmode_in, bmode_out
4911 Set balance mode for balance_in/balance_out option.
4913 Can be one of the following:
4917 Classic balance mode. Attenuate one channel at time.
4918 Gain is raised up to 1.
4921 Similar as classic mode above but gain is raised up to 2.
4924 Equal power distribution, from -6dB to +6dB range.
4928 @subsection Examples
4932 Apply karaoke like effect:
4934 stereotools=mlev=0.015625
4938 Convert M/S signal to L/R:
4940 "stereotools=mode=ms>lr"
4944 @section stereowiden
4946 This filter enhance the stereo effect by suppressing signal common to both
4947 channels and by delaying the signal of left into right and vice versa,
4948 thereby widening the stereo effect.
4950 The filter accepts the following options:
4954 Time in milliseconds of the delay of left signal into right and vice versa.
4955 Default is 20 milliseconds.
4958 Amount of gain in delayed signal into right and vice versa. Gives a delay
4959 effect of left signal in right output and vice versa which gives widening
4960 effect. Default is 0.3.
4963 Cross feed of left into right with inverted phase. This helps in suppressing
4964 the mono. If the value is 1 it will cancel all the signal common to both
4965 channels. Default is 0.3.
4968 Set level of input signal of original channel. Default is 0.8.
4971 @section superequalizer
4972 Apply 18 band equalizer.
4974 The filter accepts the following options:
4981 Set 131Hz band gain.
4983 Set 185Hz band gain.
4985 Set 262Hz band gain.
4987 Set 370Hz band gain.
4989 Set 523Hz band gain.
4991 Set 740Hz band gain.
4993 Set 1047Hz band gain.
4995 Set 1480Hz band gain.
4997 Set 2093Hz band gain.
4999 Set 2960Hz band gain.
5001 Set 4186Hz band gain.
5003 Set 5920Hz band gain.
5005 Set 8372Hz band gain.
5007 Set 11840Hz band gain.
5009 Set 16744Hz band gain.
5011 Set 20000Hz band gain.
5015 Apply audio surround upmix filter.
5017 This filter allows to produce multichannel output from audio stream.
5019 The filter accepts the following options:
5023 Set output channel layout. By default, this is @var{5.1}.
5025 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5026 for the required syntax.
5029 Set input channel layout. By default, this is @var{stereo}.
5031 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5032 for the required syntax.
5035 Set input volume level. By default, this is @var{1}.
5038 Set output volume level. By default, this is @var{1}.
5041 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5044 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5047 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5050 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5051 In @var{add} mode, LFE channel is created from input audio and added to output.
5052 In @var{sub} mode, LFE channel is created from input audio and added to output but
5053 also all non-LFE output channels are subtracted with output LFE channel.
5056 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5057 Default is @var{90}.
5060 Set front center input volume. By default, this is @var{1}.
5063 Set front center output volume. By default, this is @var{1}.
5066 Set front left input volume. By default, this is @var{1}.
5069 Set front left output volume. By default, this is @var{1}.
5072 Set front right input volume. By default, this is @var{1}.
5075 Set front right output volume. By default, this is @var{1}.
5078 Set side left input volume. By default, this is @var{1}.
5081 Set side left output volume. By default, this is @var{1}.
5084 Set side right input volume. By default, this is @var{1}.
5087 Set side right output volume. By default, this is @var{1}.
5090 Set back left input volume. By default, this is @var{1}.
5093 Set back left output volume. By default, this is @var{1}.
5096 Set back right input volume. By default, this is @var{1}.
5099 Set back right output volume. By default, this is @var{1}.
5102 Set back center input volume. By default, this is @var{1}.
5105 Set back center output volume. By default, this is @var{1}.
5108 Set LFE input volume. By default, this is @var{1}.
5111 Set LFE output volume. By default, this is @var{1}.
5114 Set spread usage of stereo image across X axis for all channels.
5117 Set spread usage of stereo image across Y axis for all channels.
5119 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5120 Set spread usage of stereo image across X axis for each channel.
5122 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5123 Set spread usage of stereo image across Y axis for each channel.
5126 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5129 Set window function.
5131 It accepts the following values:
5154 Default is @code{hann}.
5157 Set window overlap. If set to 1, the recommended overlap for selected
5158 window function will be picked. Default is @code{0.5}.
5161 @section treble, highshelf
5163 Boost or cut treble (upper) frequencies of the audio using a two-pole
5164 shelving filter with a response similar to that of a standard
5165 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5167 The filter accepts the following options:
5171 Give the gain at whichever is the lower of ~22 kHz and the
5172 Nyquist frequency. Its useful range is about -20 (for a large cut)
5173 to +20 (for a large boost). Beware of clipping when using a positive gain.
5176 Set the filter's central frequency and so can be used
5177 to extend or reduce the frequency range to be boosted or cut.
5178 The default value is @code{3000} Hz.
5181 Set method to specify band-width of filter.
5196 Determine how steep is the filter's shelf transition.
5199 How much to use filtered signal in output. Default is 1.
5200 Range is between 0 and 1.
5203 Specify which channels to filter, by default all available are filtered.
5206 @subsection Commands
5208 This filter supports the following commands:
5211 Change treble frequency.
5212 Syntax for the command is : "@var{frequency}"
5215 Change treble width_type.
5216 Syntax for the command is : "@var{width_type}"
5219 Change treble width.
5220 Syntax for the command is : "@var{width}"
5224 Syntax for the command is : "@var{gain}"
5228 Syntax for the command is : "@var{mix}"
5233 Sinusoidal amplitude modulation.
5235 The filter accepts the following options:
5239 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5240 (20 Hz or lower) will result in a tremolo effect.
5241 This filter may also be used as a ring modulator by specifying
5242 a modulation frequency higher than 20 Hz.
5243 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5246 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5247 Default value is 0.5.
5252 Sinusoidal phase modulation.
5254 The filter accepts the following options:
5258 Modulation frequency in Hertz.
5259 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5262 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5263 Default value is 0.5.
5268 Adjust the input audio volume.
5270 It accepts the following parameters:
5274 Set audio volume expression.
5276 Output values are clipped to the maximum value.
5278 The output audio volume is given by the relation:
5280 @var{output_volume} = @var{volume} * @var{input_volume}
5283 The default value for @var{volume} is "1.0".
5286 This parameter represents the mathematical precision.
5288 It determines which input sample formats will be allowed, which affects the
5289 precision of the volume scaling.
5293 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5295 32-bit floating-point; this limits input sample format to FLT. (default)
5297 64-bit floating-point; this limits input sample format to DBL.
5301 Choose the behaviour on encountering ReplayGain side data in input frames.
5305 Remove ReplayGain side data, ignoring its contents (the default).
5308 Ignore ReplayGain side data, but leave it in the frame.
5311 Prefer the track gain, if present.
5314 Prefer the album gain, if present.
5317 @item replaygain_preamp
5318 Pre-amplification gain in dB to apply to the selected replaygain gain.
5320 Default value for @var{replaygain_preamp} is 0.0.
5323 Set when the volume expression is evaluated.
5325 It accepts the following values:
5328 only evaluate expression once during the filter initialization, or
5329 when the @samp{volume} command is sent
5332 evaluate expression for each incoming frame
5335 Default value is @samp{once}.
5338 The volume expression can contain the following parameters.
5342 frame number (starting at zero)
5345 @item nb_consumed_samples
5346 number of samples consumed by the filter
5348 number of samples in the current frame
5350 original frame position in the file
5356 PTS at start of stream
5358 time at start of stream
5364 last set volume value
5367 Note that when @option{eval} is set to @samp{once} only the
5368 @var{sample_rate} and @var{tb} variables are available, all other
5369 variables will evaluate to NAN.
5371 @subsection Commands
5373 This filter supports the following commands:
5376 Modify the volume expression.
5377 The command accepts the same syntax of the corresponding option.
5379 If the specified expression is not valid, it is kept at its current
5381 @item replaygain_noclip
5382 Prevent clipping by limiting the gain applied.
5384 Default value for @var{replaygain_noclip} is 1.
5388 @subsection Examples
5392 Halve the input audio volume:
5396 volume=volume=-6.0206dB
5399 In all the above example the named key for @option{volume} can be
5400 omitted, for example like in:
5406 Increase input audio power by 6 decibels using fixed-point precision:
5408 volume=volume=6dB:precision=fixed
5412 Fade volume after time 10 with an annihilation period of 5 seconds:
5414 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5418 @section volumedetect
5420 Detect the volume of the input video.
5422 The filter has no parameters. The input is not modified. Statistics about
5423 the volume will be printed in the log when the input stream end is reached.
5425 In particular it will show the mean volume (root mean square), maximum
5426 volume (on a per-sample basis), and the beginning of a histogram of the
5427 registered volume values (from the maximum value to a cumulated 1/1000 of
5430 All volumes are in decibels relative to the maximum PCM value.
5432 @subsection Examples
5434 Here is an excerpt of the output:
5436 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5437 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5438 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5439 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5440 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5441 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5442 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5443 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5444 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5450 The mean square energy is approximately -27 dB, or 10^-2.7.
5452 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5454 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5457 In other words, raising the volume by +4 dB does not cause any clipping,
5458 raising it by +5 dB causes clipping for 6 samples, etc.
5460 @c man end AUDIO FILTERS
5462 @chapter Audio Sources
5463 @c man begin AUDIO SOURCES
5465 Below is a description of the currently available audio sources.
5469 Buffer audio frames, and make them available to the filter chain.
5471 This source is mainly intended for a programmatic use, in particular
5472 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5474 It accepts the following parameters:
5478 The timebase which will be used for timestamps of submitted frames. It must be
5479 either a floating-point number or in @var{numerator}/@var{denominator} form.
5482 The sample rate of the incoming audio buffers.
5485 The sample format of the incoming audio buffers.
5486 Either a sample format name or its corresponding integer representation from
5487 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5489 @item channel_layout
5490 The channel layout of the incoming audio buffers.
5491 Either a channel layout name from channel_layout_map in
5492 @file{libavutil/channel_layout.c} or its corresponding integer representation
5493 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5496 The number of channels of the incoming audio buffers.
5497 If both @var{channels} and @var{channel_layout} are specified, then they
5502 @subsection Examples
5505 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5508 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5509 Since the sample format with name "s16p" corresponds to the number
5510 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5513 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5518 Generate an audio signal specified by an expression.
5520 This source accepts in input one or more expressions (one for each
5521 channel), which are evaluated and used to generate a corresponding
5524 This source accepts the following options:
5528 Set the '|'-separated expressions list for each separate channel. In case the
5529 @option{channel_layout} option is not specified, the selected channel layout
5530 depends on the number of provided expressions. Otherwise the last
5531 specified expression is applied to the remaining output channels.
5533 @item channel_layout, c
5534 Set the channel layout. The number of channels in the specified layout
5535 must be equal to the number of specified expressions.
5538 Set the minimum duration of the sourced audio. See
5539 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5540 for the accepted syntax.
5541 Note that the resulting duration may be greater than the specified
5542 duration, as the generated audio is always cut at the end of a
5545 If not specified, or the expressed duration is negative, the audio is
5546 supposed to be generated forever.
5549 Set the number of samples per channel per each output frame,
5552 @item sample_rate, s
5553 Specify the sample rate, default to 44100.
5556 Each expression in @var{exprs} can contain the following constants:
5560 number of the evaluated sample, starting from 0
5563 time of the evaluated sample expressed in seconds, starting from 0
5570 @subsection Examples
5580 Generate a sin signal with frequency of 440 Hz, set sample rate to
5583 aevalsrc="sin(440*2*PI*t):s=8000"
5587 Generate a two channels signal, specify the channel layout (Front
5588 Center + Back Center) explicitly:
5590 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5594 Generate white noise:
5596 aevalsrc="-2+random(0)"
5600 Generate an amplitude modulated signal:
5602 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5606 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5608 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5615 The null audio source, return unprocessed audio frames. It is mainly useful
5616 as a template and to be employed in analysis / debugging tools, or as
5617 the source for filters which ignore the input data (for example the sox
5620 This source accepts the following options:
5624 @item channel_layout, cl
5626 Specifies the channel layout, and can be either an integer or a string
5627 representing a channel layout. The default value of @var{channel_layout}
5630 Check the channel_layout_map definition in
5631 @file{libavutil/channel_layout.c} for the mapping between strings and
5632 channel layout values.
5634 @item sample_rate, r
5635 Specifies the sample rate, and defaults to 44100.
5638 Set the number of samples per requested frames.
5642 @subsection Examples
5646 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5648 anullsrc=r=48000:cl=4
5652 Do the same operation with a more obvious syntax:
5654 anullsrc=r=48000:cl=mono
5658 All the parameters need to be explicitly defined.
5662 Synthesize a voice utterance using the libflite library.
5664 To enable compilation of this filter you need to configure FFmpeg with
5665 @code{--enable-libflite}.
5667 Note that versions of the flite library prior to 2.0 are not thread-safe.
5669 The filter accepts the following options:
5674 If set to 1, list the names of the available voices and exit
5675 immediately. Default value is 0.
5678 Set the maximum number of samples per frame. Default value is 512.
5681 Set the filename containing the text to speak.
5684 Set the text to speak.
5687 Set the voice to use for the speech synthesis. Default value is
5688 @code{kal}. See also the @var{list_voices} option.
5691 @subsection Examples
5695 Read from file @file{speech.txt}, and synthesize the text using the
5696 standard flite voice:
5698 flite=textfile=speech.txt
5702 Read the specified text selecting the @code{slt} voice:
5704 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5708 Input text to ffmpeg:
5710 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5714 Make @file{ffplay} speak the specified text, using @code{flite} and
5715 the @code{lavfi} device:
5717 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
5721 For more information about libflite, check:
5722 @url{http://www.festvox.org/flite/}
5726 Generate a noise audio signal.
5728 The filter accepts the following options:
5731 @item sample_rate, r
5732 Specify the sample rate. Default value is 48000 Hz.
5735 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
5739 Specify the duration of the generated audio stream. Not specifying this option
5740 results in noise with an infinite length.
5742 @item color, colour, c
5743 Specify the color of noise. Available noise colors are white, pink, brown,
5744 blue and violet. Default color is white.
5747 Specify a value used to seed the PRNG.
5750 Set the number of samples per each output frame, default is 1024.
5753 @subsection Examples
5758 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
5760 anoisesrc=d=60:c=pink:r=44100:a=0.5
5766 Generate odd-tap Hilbert transform FIR coefficients.
5768 The resulting stream can be used with @ref{afir} filter for phase-shifting
5769 the signal by 90 degrees.
5771 This is used in many matrix coding schemes and for analytic signal generation.
5772 The process is often written as a multiplication by i (or j), the imaginary unit.
5774 The filter accepts the following options:
5778 @item sample_rate, s
5779 Set sample rate, default is 44100.
5782 Set length of FIR filter, default is 22051.
5785 Set number of samples per each frame.
5788 Set window function to be used when generating FIR coefficients.
5793 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
5795 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
5797 The filter accepts the following options:
5800 @item sample_rate, r
5801 Set sample rate, default is 44100.
5804 Set number of samples per each frame. Default is 1024.
5807 Set high-pass frequency. Default is 0.
5810 Set low-pass frequency. Default is 0.
5811 If high-pass frequency is lower than low-pass frequency and low-pass frequency
5812 is higher than 0 then filter will create band-pass filter coefficients,
5813 otherwise band-reject filter coefficients.
5816 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
5819 Set Kaiser window beta.
5822 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
5825 Enable rounding, by default is disabled.
5828 Set number of taps for high-pass filter.
5831 Set number of taps for low-pass filter.
5836 Generate an audio signal made of a sine wave with amplitude 1/8.
5838 The audio signal is bit-exact.
5840 The filter accepts the following options:
5845 Set the carrier frequency. Default is 440 Hz.
5847 @item beep_factor, b
5848 Enable a periodic beep every second with frequency @var{beep_factor} times
5849 the carrier frequency. Default is 0, meaning the beep is disabled.
5851 @item sample_rate, r
5852 Specify the sample rate, default is 44100.
5855 Specify the duration of the generated audio stream.
5857 @item samples_per_frame
5858 Set the number of samples per output frame.
5860 The expression can contain the following constants:
5864 The (sequential) number of the output audio frame, starting from 0.
5867 The PTS (Presentation TimeStamp) of the output audio frame,
5868 expressed in @var{TB} units.
5871 The PTS of the output audio frame, expressed in seconds.
5874 The timebase of the output audio frames.
5877 Default is @code{1024}.
5880 @subsection Examples
5885 Generate a simple 440 Hz sine wave:
5891 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
5895 sine=frequency=220:beep_factor=4:duration=5
5899 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
5902 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
5906 @c man end AUDIO SOURCES
5908 @chapter Audio Sinks
5909 @c man begin AUDIO SINKS
5911 Below is a description of the currently available audio sinks.
5913 @section abuffersink
5915 Buffer audio frames, and make them available to the end of filter chain.
5917 This sink is mainly intended for programmatic use, in particular
5918 through the interface defined in @file{libavfilter/buffersink.h}
5919 or the options system.
5921 It accepts a pointer to an AVABufferSinkContext structure, which
5922 defines the incoming buffers' formats, to be passed as the opaque
5923 parameter to @code{avfilter_init_filter} for initialization.
5926 Null audio sink; do absolutely nothing with the input audio. It is
5927 mainly useful as a template and for use in analysis / debugging
5930 @c man end AUDIO SINKS
5932 @chapter Video Filters
5933 @c man begin VIDEO FILTERS
5935 When you configure your FFmpeg build, you can disable any of the
5936 existing filters using @code{--disable-filters}.
5937 The configure output will show the video filters included in your
5940 Below is a description of the currently available video filters.
5944 Mark a region of interest in a video frame.
5946 The frame data is passed through unchanged, but metadata is attached
5947 to the frame indicating regions of interest which can affect the
5948 behaviour of later encoding. Multiple regions can be marked by
5949 applying the filter multiple times.
5953 Region distance in pixels from the left edge of the frame.
5955 Region distance in pixels from the top edge of the frame.
5957 Region width in pixels.
5959 Region height in pixels.
5961 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
5962 and may contain the following variables:
5965 Width of the input frame.
5967 Height of the input frame.
5971 Quantisation offset to apply within the region.
5973 This must be a real value in the range -1 to +1. A value of zero
5974 indicates no quality change. A negative value asks for better quality
5975 (less quantisation), while a positive value asks for worse quality
5976 (greater quantisation).
5978 The range is calibrated so that the extreme values indicate the
5979 largest possible offset - if the rest of the frame is encoded with the
5980 worst possible quality, an offset of -1 indicates that this region
5981 should be encoded with the best possible quality anyway. Intermediate
5982 values are then interpolated in some codec-dependent way.
5984 For example, in 10-bit H.264 the quantisation parameter varies between
5985 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
5986 this region should be encoded with a QP around one-tenth of the full
5987 range better than the rest of the frame. So, if most of the frame
5988 were to be encoded with a QP of around 30, this region would get a QP
5989 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
5990 An extreme value of -1 would indicate that this region should be
5991 encoded with the best possible quality regardless of the treatment of
5992 the rest of the frame - that is, should be encoded at a QP of -12.
5994 If set to true, remove any existing regions of interest marked on the
5995 frame before adding the new one.
5998 @subsection Examples
6002 Mark the centre quarter of the frame as interesting.
6004 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6007 Mark the 100-pixel-wide region on the left edge of the frame as very
6008 uninteresting (to be encoded at much lower quality than the rest of
6011 addroi=0:0:100:ih:+1/5
6015 @section alphaextract
6017 Extract the alpha component from the input as a grayscale video. This
6018 is especially useful with the @var{alphamerge} filter.
6022 Add or replace the alpha component of the primary input with the
6023 grayscale value of a second input. This is intended for use with
6024 @var{alphaextract} to allow the transmission or storage of frame
6025 sequences that have alpha in a format that doesn't support an alpha
6028 For example, to reconstruct full frames from a normal YUV-encoded video
6029 and a separate video created with @var{alphaextract}, you might use:
6031 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6034 Since this filter is designed for reconstruction, it operates on frame
6035 sequences without considering timestamps, and terminates when either
6036 input reaches end of stream. This will cause problems if your encoding
6037 pipeline drops frames. If you're trying to apply an image as an
6038 overlay to a video stream, consider the @var{overlay} filter instead.
6042 Amplify differences between current pixel and pixels of adjacent frames in
6043 same pixel location.
6045 This filter accepts the following options:
6049 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6050 For example radius of 3 will instruct filter to calculate average of 7 frames.
6053 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6056 Set threshold for difference amplification. Any difference greater or equal to
6057 this value will not alter source pixel. Default is 10.
6058 Allowed range is from 0 to 65535.
6061 Set tolerance for difference amplification. Any difference lower to
6062 this value will not alter source pixel. Default is 0.
6063 Allowed range is from 0 to 65535.
6066 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6067 This option controls maximum possible value that will decrease source pixel value.
6070 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6071 This option controls maximum possible value that will increase source pixel value.
6074 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6079 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6080 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6081 Substation Alpha) subtitles files.
6083 This filter accepts the following option in addition to the common options from
6084 the @ref{subtitles} filter:
6088 Set the shaping engine
6090 Available values are:
6093 The default libass shaping engine, which is the best available.
6095 Fast, font-agnostic shaper that can do only substitutions
6097 Slower shaper using OpenType for substitutions and positioning
6100 The default is @code{auto}.
6104 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6106 The filter accepts the following options:
6110 Set threshold A for 1st plane. Default is 0.02.
6111 Valid range is 0 to 0.3.
6114 Set threshold B for 1st plane. Default is 0.04.
6115 Valid range is 0 to 5.
6118 Set threshold A for 2nd plane. Default is 0.02.
6119 Valid range is 0 to 0.3.
6122 Set threshold B for 2nd plane. Default is 0.04.
6123 Valid range is 0 to 5.
6126 Set threshold A for 3rd plane. Default is 0.02.
6127 Valid range is 0 to 0.3.
6130 Set threshold B for 3rd plane. Default is 0.04.
6131 Valid range is 0 to 5.
6133 Threshold A is designed to react on abrupt changes in the input signal and
6134 threshold B is designed to react on continuous changes in the input signal.
6137 Set number of frames filter will use for averaging. Default is 9. Must be odd
6138 number in range [5, 129].
6141 Set what planes of frame filter will use for averaging. Default is all.
6146 Apply average blur filter.
6148 The filter accepts the following options:
6152 Set horizontal radius size.
6155 Set which planes to filter. By default all planes are filtered.
6158 Set vertical radius size, if zero it will be same as @code{sizeX}.
6159 Default is @code{0}.
6164 Compute the bounding box for the non-black pixels in the input frame
6167 This filter computes the bounding box containing all the pixels with a
6168 luminance value greater than the minimum allowed value.
6169 The parameters describing the bounding box are printed on the filter
6172 The filter accepts the following option:
6176 Set the minimal luminance value. Default is @code{16}.
6179 @section bitplanenoise
6181 Show and measure bit plane noise.
6183 The filter accepts the following options:
6187 Set which plane to analyze. Default is @code{1}.
6190 Filter out noisy pixels from @code{bitplane} set above.
6191 Default is disabled.
6194 @section blackdetect
6196 Detect video intervals that are (almost) completely black. Can be
6197 useful to detect chapter transitions, commercials, or invalid
6198 recordings. Output lines contains the time for the start, end and
6199 duration of the detected black interval expressed in seconds.
6201 In order to display the output lines, you need to set the loglevel at
6202 least to the AV_LOG_INFO value.
6204 The filter accepts the following options:
6207 @item black_min_duration, d
6208 Set the minimum detected black duration expressed in seconds. It must
6209 be a non-negative floating point number.
6211 Default value is 2.0.
6213 @item picture_black_ratio_th, pic_th
6214 Set the threshold for considering a picture "black".
6215 Express the minimum value for the ratio:
6217 @var{nb_black_pixels} / @var{nb_pixels}
6220 for which a picture is considered black.
6221 Default value is 0.98.
6223 @item pixel_black_th, pix_th
6224 Set the threshold for considering a pixel "black".
6226 The threshold expresses the maximum pixel luminance value for which a
6227 pixel is considered "black". The provided value is scaled according to
6228 the following equation:
6230 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6233 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6234 the input video format, the range is [0-255] for YUV full-range
6235 formats and [16-235] for YUV non full-range formats.
6237 Default value is 0.10.
6240 The following example sets the maximum pixel threshold to the minimum
6241 value, and detects only black intervals of 2 or more seconds:
6243 blackdetect=d=2:pix_th=0.00
6248 Detect frames that are (almost) completely black. Can be useful to
6249 detect chapter transitions or commercials. Output lines consist of
6250 the frame number of the detected frame, the percentage of blackness,
6251 the position in the file if known or -1 and the timestamp in seconds.
6253 In order to display the output lines, you need to set the loglevel at
6254 least to the AV_LOG_INFO value.
6256 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6257 The value represents the percentage of pixels in the picture that
6258 are below the threshold value.
6260 It accepts the following parameters:
6265 The percentage of the pixels that have to be below the threshold; it defaults to
6268 @item threshold, thresh
6269 The threshold below which a pixel value is considered black; it defaults to
6274 @section blend, tblend
6276 Blend two video frames into each other.
6278 The @code{blend} filter takes two input streams and outputs one
6279 stream, the first input is the "top" layer and second input is
6280 "bottom" layer. By default, the output terminates when the longest input terminates.
6282 The @code{tblend} (time blend) filter takes two consecutive frames
6283 from one single stream, and outputs the result obtained by blending
6284 the new frame on top of the old frame.
6286 A description of the accepted options follows.
6294 Set blend mode for specific pixel component or all pixel components in case
6295 of @var{all_mode}. Default value is @code{normal}.
6297 Available values for component modes are:
6339 Set blend opacity for specific pixel component or all pixel components in case
6340 of @var{all_opacity}. Only used in combination with pixel component blend modes.
6347 Set blend expression for specific pixel component or all pixel components in case
6348 of @var{all_expr}. Note that related mode options will be ignored if those are set.
6350 The expressions can use the following variables:
6354 The sequential number of the filtered frame, starting from @code{0}.
6358 the coordinates of the current sample
6362 the width and height of currently filtered plane
6366 Width and height scale for the plane being filtered. It is the
6367 ratio between the dimensions of the current plane to the luma plane,
6368 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
6369 the luma plane and @code{0.5,0.5} for the chroma planes.
6372 Time of the current frame, expressed in seconds.
6375 Value of pixel component at current location for first video frame (top layer).
6378 Value of pixel component at current location for second video frame (bottom layer).
6382 The @code{blend} filter also supports the @ref{framesync} options.
6384 @subsection Examples
6388 Apply transition from bottom layer to top layer in first 10 seconds:
6390 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
6394 Apply linear horizontal transition from top layer to bottom layer:
6396 blend=all_expr='A*(X/W)+B*(1-X/W)'
6400 Apply 1x1 checkerboard effect:
6402 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
6406 Apply uncover left effect:
6408 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6412 Apply uncover down effect:
6414 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6418 Apply uncover up-left effect:
6420 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6424 Split diagonally video and shows top and bottom layer on each side:
6426 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6430 Display differences between the current and the previous frame:
6432 tblend=all_mode=grainextract
6438 Denoise frames using Block-Matching 3D algorithm.
6440 The filter accepts the following options.
6444 Set denoising strength. Default value is 1.
6445 Allowed range is from 0 to 999.9.
6446 The denoising algorithm is very sensitive to sigma, so adjust it
6447 according to the source.
6450 Set local patch size. This sets dimensions in 2D.
6453 Set sliding step for processing blocks. Default value is 4.
6454 Allowed range is from 1 to 64.
6455 Smaller values allows processing more reference blocks and is slower.
6458 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6459 When set to 1, no block matching is done. Larger values allows more blocks
6461 Allowed range is from 1 to 256.
6464 Set radius for search block matching. Default is 9.
6465 Allowed range is from 1 to INT32_MAX.
6468 Set step between two search locations for block matching. Default is 1.
6469 Allowed range is from 1 to 64. Smaller is slower.
6472 Set threshold of mean square error for block matching. Valid range is 0 to
6476 Set thresholding parameter for hard thresholding in 3D transformed domain.
6477 Larger values results in stronger hard-thresholding filtering in frequency
6481 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6482 Default is @code{basic}.
6485 If enabled, filter will use 2nd stream for block matching.
6486 Default is disabled for @code{basic} value of @var{estim} option,
6487 and always enabled if value of @var{estim} is @code{final}.
6490 Set planes to filter. Default is all available except alpha.
6493 @subsection Examples
6497 Basic filtering with bm3d:
6499 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
6503 Same as above, but filtering only luma:
6505 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
6509 Same as above, but with both estimation modes:
6511 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
6515 Same as above, but prefilter with @ref{nlmeans} filter instead:
6517 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
6523 Apply a boxblur algorithm to the input video.
6525 It accepts the following parameters:
6529 @item luma_radius, lr
6530 @item luma_power, lp
6531 @item chroma_radius, cr
6532 @item chroma_power, cp
6533 @item alpha_radius, ar
6534 @item alpha_power, ap
6538 A description of the accepted options follows.
6541 @item luma_radius, lr
6542 @item chroma_radius, cr
6543 @item alpha_radius, ar
6544 Set an expression for the box radius in pixels used for blurring the
6545 corresponding input plane.
6547 The radius value must be a non-negative number, and must not be
6548 greater than the value of the expression @code{min(w,h)/2} for the
6549 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
6552 Default value for @option{luma_radius} is "2". If not specified,
6553 @option{chroma_radius} and @option{alpha_radius} default to the
6554 corresponding value set for @option{luma_radius}.
6556 The expressions can contain the following constants:
6560 The input width and height in pixels.
6564 The input chroma image width and height in pixels.
6568 The horizontal and vertical chroma subsample values. For example, for the
6569 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6572 @item luma_power, lp
6573 @item chroma_power, cp
6574 @item alpha_power, ap
6575 Specify how many times the boxblur filter is applied to the
6576 corresponding plane.
6578 Default value for @option{luma_power} is 2. If not specified,
6579 @option{chroma_power} and @option{alpha_power} default to the
6580 corresponding value set for @option{luma_power}.
6582 A value of 0 will disable the effect.
6585 @subsection Examples
6589 Apply a boxblur filter with the luma, chroma, and alpha radii
6592 boxblur=luma_radius=2:luma_power=1
6597 Set the luma radius to 2, and alpha and chroma radius to 0:
6599 boxblur=2:1:cr=0:ar=0
6603 Set the luma and chroma radii to a fraction of the video dimension:
6605 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6611 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6612 Deinterlacing Filter").
6614 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6615 interpolation algorithms.
6616 It accepts the following parameters:
6620 The interlacing mode to adopt. It accepts one of the following values:
6624 Output one frame for each frame.
6626 Output one frame for each field.
6629 The default value is @code{send_field}.
6632 The picture field parity assumed for the input interlaced video. It accepts one
6633 of the following values:
6637 Assume the top field is first.
6639 Assume the bottom field is first.
6641 Enable automatic detection of field parity.
6644 The default value is @code{auto}.
6645 If the interlacing is unknown or the decoder does not export this information,
6646 top field first will be assumed.
6649 Specify which frames to deinterlace. Accepts one of the following
6654 Deinterlace all frames.
6656 Only deinterlace frames marked as interlaced.
6659 The default value is @code{all}.
6663 Remove all color information for all colors except for certain one.
6665 The filter accepts the following options:
6669 The color which will not be replaced with neutral chroma.
6672 Similarity percentage with the above color.
6673 0.01 matches only the exact key color, while 1.0 matches everything.
6677 0.0 makes pixels either fully gray, or not gray at all.
6678 Higher values result in more preserved color.
6681 Signals that the color passed is already in YUV instead of RGB.
6683 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6684 This can be used to pass exact YUV values as hexadecimal numbers.
6688 YUV colorspace color/chroma keying.
6690 The filter accepts the following options:
6694 The color which will be replaced with transparency.
6697 Similarity percentage with the key color.
6699 0.01 matches only the exact key color, while 1.0 matches everything.
6704 0.0 makes pixels either fully transparent, or not transparent at all.
6706 Higher values result in semi-transparent pixels, with a higher transparency
6707 the more similar the pixels color is to the key color.
6710 Signals that the color passed is already in YUV instead of RGB.
6712 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6713 This can be used to pass exact YUV values as hexadecimal numbers.
6716 @subsection Examples
6720 Make every green pixel in the input image transparent:
6722 ffmpeg -i input.png -vf chromakey=green out.png
6726 Overlay a greenscreen-video on top of a static black background.
6728 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
6732 @section chromashift
6733 Shift chroma pixels horizontally and/or vertically.
6735 The filter accepts the following options:
6738 Set amount to shift chroma-blue horizontally.
6740 Set amount to shift chroma-blue vertically.
6742 Set amount to shift chroma-red horizontally.
6744 Set amount to shift chroma-red vertically.
6746 Set edge mode, can be @var{smear}, default, or @var{warp}.
6751 Display CIE color diagram with pixels overlaid onto it.
6753 The filter accepts the following options:
6768 @item uhdtv, rec2020
6782 Set what gamuts to draw.
6784 See @code{system} option for available values.
6787 Set ciescope size, by default set to 512.
6790 Set intensity used to map input pixel values to CIE diagram.
6793 Set contrast used to draw tongue colors that are out of active color system gamut.
6796 Correct gamma displayed on scope, by default enabled.
6799 Show white point on CIE diagram, by default disabled.
6802 Set input gamma. Used only with XYZ input color space.
6807 Visualize information exported by some codecs.
6809 Some codecs can export information through frames using side-data or other
6810 means. For example, some MPEG based codecs export motion vectors through the
6811 @var{export_mvs} flag in the codec @option{flags2} option.
6813 The filter accepts the following option:
6817 Set motion vectors to visualize.
6819 Available flags for @var{mv} are:
6823 forward predicted MVs of P-frames
6825 forward predicted MVs of B-frames
6827 backward predicted MVs of B-frames
6831 Display quantization parameters using the chroma planes.
6834 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
6836 Available flags for @var{mv_type} are:
6840 forward predicted MVs
6842 backward predicted MVs
6845 @item frame_type, ft
6846 Set frame type to visualize motion vectors of.
6848 Available flags for @var{frame_type} are:
6852 intra-coded frames (I-frames)
6854 predicted frames (P-frames)
6856 bi-directionally predicted frames (B-frames)
6860 @subsection Examples
6864 Visualize forward predicted MVs of all frames using @command{ffplay}:
6866 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
6870 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
6872 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
6876 @section colorbalance
6877 Modify intensity of primary colors (red, green and blue) of input frames.
6879 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
6880 regions for the red-cyan, green-magenta or blue-yellow balance.
6882 A positive adjustment value shifts the balance towards the primary color, a negative
6883 value towards the complementary color.
6885 The filter accepts the following options:
6891 Adjust red, green and blue shadows (darkest pixels).
6896 Adjust red, green and blue midtones (medium pixels).
6901 Adjust red, green and blue highlights (brightest pixels).
6903 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
6906 @subsection Examples
6910 Add red color cast to shadows:
6916 @section colorchannelmixer
6918 Adjust video input frames by re-mixing color channels.
6920 This filter modifies a color channel by adding the values associated to
6921 the other channels of the same pixels. For example if the value to
6922 modify is red, the output value will be:
6924 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
6927 The filter accepts the following options:
6934 Adjust contribution of input red, green, blue and alpha channels for output red channel.
6935 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
6941 Adjust contribution of input red, green, blue and alpha channels for output green channel.
6942 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
6948 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
6949 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
6955 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
6956 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
6958 Allowed ranges for options are @code{[-2.0, 2.0]}.
6961 @subsection Examples
6965 Convert source to grayscale:
6967 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
6970 Simulate sepia tones:
6972 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
6977 RGB colorspace color keying.
6979 The filter accepts the following options:
6983 The color which will be replaced with transparency.
6986 Similarity percentage with the key color.
6988 0.01 matches only the exact key color, while 1.0 matches everything.
6993 0.0 makes pixels either fully transparent, or not transparent at all.
6995 Higher values result in semi-transparent pixels, with a higher transparency
6996 the more similar the pixels color is to the key color.
6999 @subsection Examples
7003 Make every green pixel in the input image transparent:
7005 ffmpeg -i input.png -vf colorkey=green out.png
7009 Overlay a greenscreen-video on top of a static background image.
7011 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
7016 Remove all color information for all RGB colors except for certain one.
7018 The filter accepts the following options:
7022 The color which will not be replaced with neutral gray.
7025 Similarity percentage with the above color.
7026 0.01 matches only the exact key color, while 1.0 matches everything.
7029 Blend percentage. 0.0 makes pixels fully gray.
7030 Higher values result in more preserved color.
7033 @section colorlevels
7035 Adjust video input frames using levels.
7037 The filter accepts the following options:
7044 Adjust red, green, blue and alpha input black point.
7045 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7051 Adjust red, green, blue and alpha input white point.
7052 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7054 Input levels are used to lighten highlights (bright tones), darken shadows
7055 (dark tones), change the balance of bright and dark tones.
7061 Adjust red, green, blue and alpha output black point.
7062 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7068 Adjust red, green, blue and alpha output white point.
7069 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7071 Output levels allows manual selection of a constrained output level range.
7074 @subsection Examples
7078 Make video output darker:
7080 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7086 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7090 Make video output lighter:
7092 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7096 Increase brightness:
7098 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7102 @section colormatrix
7104 Convert color matrix.
7106 The filter accepts the following options:
7111 Specify the source and destination color matrix. Both values must be
7114 The accepted values are:
7142 For example to convert from BT.601 to SMPTE-240M, use the command:
7144 colormatrix=bt601:smpte240m
7149 Convert colorspace, transfer characteristics or color primaries.
7150 Input video needs to have an even size.
7152 The filter accepts the following options:
7157 Specify all color properties at once.
7159 The accepted values are:
7189 Specify output colorspace.
7191 The accepted values are:
7200 BT.470BG or BT.601-6 625
7203 SMPTE-170M or BT.601-6 525
7212 BT.2020 with non-constant luminance
7218 Specify output transfer characteristics.
7220 The accepted values are:
7232 Constant gamma of 2.2
7235 Constant gamma of 2.8
7238 SMPTE-170M, BT.601-6 625 or BT.601-6 525
7256 BT.2020 for 10-bits content
7259 BT.2020 for 12-bits content
7265 Specify output color primaries.
7267 The accepted values are:
7276 BT.470BG or BT.601-6 625
7279 SMPTE-170M or BT.601-6 525
7303 Specify output color range.
7305 The accepted values are:
7308 TV (restricted) range
7311 MPEG (restricted) range
7322 Specify output color format.
7324 The accepted values are:
7327 YUV 4:2:0 planar 8-bits
7330 YUV 4:2:0 planar 10-bits
7333 YUV 4:2:0 planar 12-bits
7336 YUV 4:2:2 planar 8-bits
7339 YUV 4:2:2 planar 10-bits
7342 YUV 4:2:2 planar 12-bits
7345 YUV 4:4:4 planar 8-bits
7348 YUV 4:4:4 planar 10-bits
7351 YUV 4:4:4 planar 12-bits
7356 Do a fast conversion, which skips gamma/primary correction. This will take
7357 significantly less CPU, but will be mathematically incorrect. To get output
7358 compatible with that produced by the colormatrix filter, use fast=1.
7361 Specify dithering mode.
7363 The accepted values are:
7369 Floyd-Steinberg dithering
7373 Whitepoint adaptation mode.
7375 The accepted values are:
7378 Bradford whitepoint adaptation
7381 von Kries whitepoint adaptation
7384 identity whitepoint adaptation (i.e. no whitepoint adaptation)
7388 Override all input properties at once. Same accepted values as @ref{all}.
7391 Override input colorspace. Same accepted values as @ref{space}.
7394 Override input color primaries. Same accepted values as @ref{primaries}.
7397 Override input transfer characteristics. Same accepted values as @ref{trc}.
7400 Override input color range. Same accepted values as @ref{range}.
7404 The filter converts the transfer characteristics, color space and color
7405 primaries to the specified user values. The output value, if not specified,
7406 is set to a default value based on the "all" property. If that property is
7407 also not specified, the filter will log an error. The output color range and
7408 format default to the same value as the input color range and format. The
7409 input transfer characteristics, color space, color primaries and color range
7410 should be set on the input data. If any of these are missing, the filter will
7411 log an error and no conversion will take place.
7413 For example to convert the input to SMPTE-240M, use the command:
7415 colorspace=smpte240m
7418 @section convolution
7420 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
7422 The filter accepts the following options:
7429 Set matrix for each plane.
7430 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
7431 and from 1 to 49 odd number of signed integers in @var{row} mode.
7437 Set multiplier for calculated value for each plane.
7438 If unset or 0, it will be sum of all matrix elements.
7444 Set bias for each plane. This value is added to the result of the multiplication.
7445 Useful for making the overall image brighter or darker. Default is 0.0.
7451 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
7452 Default is @var{square}.
7455 @subsection Examples
7461 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"
7467 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"
7473 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"
7479 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"
7483 Apply laplacian edge detector which includes diagonals:
7485 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"
7491 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"
7497 Apply 2D convolution of video stream in frequency domain using second stream
7500 The filter accepts the following options:
7504 Set which planes to process.
7507 Set which impulse video frames will be processed, can be @var{first}
7508 or @var{all}. Default is @var{all}.
7511 The @code{convolve} filter also supports the @ref{framesync} options.
7515 Copy the input video source unchanged to the output. This is mainly useful for
7520 Video filtering on GPU using Apple's CoreImage API on OSX.
7522 Hardware acceleration is based on an OpenGL context. Usually, this means it is
7523 processed by video hardware. However, software-based OpenGL implementations
7524 exist which means there is no guarantee for hardware processing. It depends on
7527 There are many filters and image generators provided by Apple that come with a
7528 large variety of options. The filter has to be referenced by its name along
7531 The coreimage filter accepts the following options:
7534 List all available filters and generators along with all their respective
7535 options as well as possible minimum and maximum values along with the default
7542 Specify all filters by their respective name and options.
7543 Use @var{list_filters} to determine all valid filter names and options.
7544 Numerical options are specified by a float value and are automatically clamped
7545 to their respective value range. Vector and color options have to be specified
7546 by a list of space separated float values. Character escaping has to be done.
7547 A special option name @code{default} is available to use default options for a
7550 It is required to specify either @code{default} or at least one of the filter options.
7551 All omitted options are used with their default values.
7552 The syntax of the filter string is as follows:
7554 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
7558 Specify a rectangle where the output of the filter chain is copied into the
7559 input image. It is given by a list of space separated float values:
7561 output_rect=x\ y\ width\ height
7563 If not given, the output rectangle equals the dimensions of the input image.
7564 The output rectangle is automatically cropped at the borders of the input
7565 image. Negative values are valid for each component.
7567 output_rect=25\ 25\ 100\ 100
7571 Several filters can be chained for successive processing without GPU-HOST
7572 transfers allowing for fast processing of complex filter chains.
7573 Currently, only filters with zero (generators) or exactly one (filters) input
7574 image and one output image are supported. Also, transition filters are not yet
7577 Some filters generate output images with additional padding depending on the
7578 respective filter kernel. The padding is automatically removed to ensure the
7579 filter output has the same size as the input image.
7581 For image generators, the size of the output image is determined by the
7582 previous output image of the filter chain or the input image of the whole
7583 filterchain, respectively. The generators do not use the pixel information of
7584 this image to generate their output. However, the generated output is
7585 blended onto this image, resulting in partial or complete coverage of the
7588 The @ref{coreimagesrc} video source can be used for generating input images
7589 which are directly fed into the filter chain. By using it, providing input
7590 images by another video source or an input video is not required.
7592 @subsection Examples
7597 List all filters available:
7599 coreimage=list_filters=true
7603 Use the CIBoxBlur filter with default options to blur an image:
7605 coreimage=filter=CIBoxBlur@@default
7609 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
7610 its center at 100x100 and a radius of 50 pixels:
7612 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
7616 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
7617 given as complete and escaped command-line for Apple's standard bash shell:
7619 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
7625 Cover a rectangular object
7627 It accepts the following options:
7631 Filepath of the optional cover image, needs to be in yuv420.
7636 It accepts the following values:
7639 cover it by the supplied image
7641 cover it by interpolating the surrounding pixels
7644 Default value is @var{blur}.
7647 @subsection Examples
7651 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
7653 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
7659 Crop the input video to given dimensions.
7661 It accepts the following parameters:
7665 The width of the output video. It defaults to @code{iw}.
7666 This expression is evaluated only once during the filter
7667 configuration, or when the @samp{w} or @samp{out_w} command is sent.
7670 The height of the output video. It defaults to @code{ih}.
7671 This expression is evaluated only once during the filter
7672 configuration, or when the @samp{h} or @samp{out_h} command is sent.
7675 The horizontal position, in the input video, of the left edge of the output
7676 video. It defaults to @code{(in_w-out_w)/2}.
7677 This expression is evaluated per-frame.
7680 The vertical position, in the input video, of the top edge of the output video.
7681 It defaults to @code{(in_h-out_h)/2}.
7682 This expression is evaluated per-frame.
7685 If set to 1 will force the output display aspect ratio
7686 to be the same of the input, by changing the output sample aspect
7687 ratio. It defaults to 0.
7690 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
7691 width/height/x/y as specified and will not be rounded to nearest smaller value.
7695 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
7696 expressions containing the following constants:
7701 The computed values for @var{x} and @var{y}. They are evaluated for
7706 The input width and height.
7710 These are the same as @var{in_w} and @var{in_h}.
7714 The output (cropped) width and height.
7718 These are the same as @var{out_w} and @var{out_h}.
7721 same as @var{iw} / @var{ih}
7724 input sample aspect ratio
7727 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
7731 horizontal and vertical chroma subsample values. For example for the
7732 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
7735 The number of the input frame, starting from 0.
7738 the position in the file of the input frame, NAN if unknown
7741 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
7745 The expression for @var{out_w} may depend on the value of @var{out_h},
7746 and the expression for @var{out_h} may depend on @var{out_w}, but they
7747 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
7748 evaluated after @var{out_w} and @var{out_h}.
7750 The @var{x} and @var{y} parameters specify the expressions for the
7751 position of the top-left corner of the output (non-cropped) area. They
7752 are evaluated for each frame. If the evaluated value is not valid, it
7753 is approximated to the nearest valid value.
7755 The expression for @var{x} may depend on @var{y}, and the expression
7756 for @var{y} may depend on @var{x}.
7758 @subsection Examples
7762 Crop area with size 100x100 at position (12,34).
7767 Using named options, the example above becomes:
7769 crop=w=100:h=100:x=12:y=34
7773 Crop the central input area with size 100x100:
7779 Crop the central input area with size 2/3 of the input video:
7781 crop=2/3*in_w:2/3*in_h
7785 Crop the input video central square:
7792 Delimit the rectangle with the top-left corner placed at position
7793 100:100 and the right-bottom corner corresponding to the right-bottom
7794 corner of the input image.
7796 crop=in_w-100:in_h-100:100:100
7800 Crop 10 pixels from the left and right borders, and 20 pixels from
7801 the top and bottom borders
7803 crop=in_w-2*10:in_h-2*20
7807 Keep only the bottom right quarter of the input image:
7809 crop=in_w/2:in_h/2:in_w/2:in_h/2
7813 Crop height for getting Greek harmony:
7815 crop=in_w:1/PHI*in_w
7819 Apply trembling effect:
7821 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)
7825 Apply erratic camera effect depending on timestamp:
7827 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)"
7831 Set x depending on the value of y:
7833 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
7837 @subsection Commands
7839 This filter supports the following commands:
7845 Set width/height of the output video and the horizontal/vertical position
7847 The command accepts the same syntax of the corresponding option.
7849 If the specified expression is not valid, it is kept at its current
7855 Auto-detect the crop size.
7857 It calculates the necessary cropping parameters and prints the
7858 recommended parameters via the logging system. The detected dimensions
7859 correspond to the non-black area of the input video.
7861 It accepts the following parameters:
7866 Set higher black value threshold, which can be optionally specified
7867 from nothing (0) to everything (255 for 8-bit based formats). An intensity
7868 value greater to the set value is considered non-black. It defaults to 24.
7869 You can also specify a value between 0.0 and 1.0 which will be scaled depending
7870 on the bitdepth of the pixel format.
7873 The value which the width/height should be divisible by. It defaults to
7874 16. The offset is automatically adjusted to center the video. Use 2 to
7875 get only even dimensions (needed for 4:2:2 video). 16 is best when
7876 encoding to most video codecs.
7878 @item reset_count, reset
7879 Set the counter that determines after how many frames cropdetect will
7880 reset the previously detected largest video area and start over to
7881 detect the current optimal crop area. Default value is 0.
7883 This can be useful when channel logos distort the video area. 0
7884 indicates 'never reset', and returns the largest area encountered during
7891 Delay video filtering until a given wallclock timestamp. The filter first
7892 passes on @option{preroll} amount of frames, then it buffers at most
7893 @option{buffer} amount of frames and waits for the cue. After reaching the cue
7894 it forwards the buffered frames and also any subsequent frames coming in its
7897 The filter can be used synchronize the output of multiple ffmpeg processes for
7898 realtime output devices like decklink. By putting the delay in the filtering
7899 chain and pre-buffering frames the process can pass on data to output almost
7900 immediately after the target wallclock timestamp is reached.
7902 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
7908 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
7911 The duration of content to pass on as preroll expressed in seconds. Default is 0.
7914 The maximum duration of content to buffer before waiting for the cue expressed
7915 in seconds. Default is 0.
7922 Apply color adjustments using curves.
7924 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
7925 component (red, green and blue) has its values defined by @var{N} key points
7926 tied from each other using a smooth curve. The x-axis represents the pixel
7927 values from the input frame, and the y-axis the new pixel values to be set for
7930 By default, a component curve is defined by the two points @var{(0;0)} and
7931 @var{(1;1)}. This creates a straight line where each original pixel value is
7932 "adjusted" to its own value, which means no change to the image.
7934 The filter allows you to redefine these two points and add some more. A new
7935 curve (using a natural cubic spline interpolation) will be define to pass
7936 smoothly through all these new coordinates. The new defined points needs to be
7937 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
7938 be in the @var{[0;1]} interval. If the computed curves happened to go outside
7939 the vector spaces, the values will be clipped accordingly.
7941 The filter accepts the following options:
7945 Select one of the available color presets. This option can be used in addition
7946 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
7947 options takes priority on the preset values.
7948 Available presets are:
7951 @item color_negative
7954 @item increase_contrast
7956 @item linear_contrast
7957 @item medium_contrast
7959 @item strong_contrast
7962 Default is @code{none}.
7964 Set the master key points. These points will define a second pass mapping. It
7965 is sometimes called a "luminance" or "value" mapping. It can be used with
7966 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
7967 post-processing LUT.
7969 Set the key points for the red component.
7971 Set the key points for the green component.
7973 Set the key points for the blue component.
7975 Set the key points for all components (not including master).
7976 Can be used in addition to the other key points component
7977 options. In this case, the unset component(s) will fallback on this
7978 @option{all} setting.
7980 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
7982 Save Gnuplot script of the curves in specified file.
7985 To avoid some filtergraph syntax conflicts, each key points list need to be
7986 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
7988 @subsection Examples
7992 Increase slightly the middle level of blue:
7994 curves=blue='0/0 0.5/0.58 1/1'
8000 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'
8002 Here we obtain the following coordinates for each components:
8005 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8007 @code{(0;0) (0.50;0.48) (1;1)}
8009 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8013 The previous example can also be achieved with the associated built-in preset:
8015 curves=preset=vintage
8025 Use a Photoshop preset and redefine the points of the green component:
8027 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8031 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8032 and @command{gnuplot}:
8034 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8035 gnuplot -p /tmp/curves.plt
8041 Video data analysis filter.
8043 This filter shows hexadecimal pixel values of part of video.
8045 The filter accepts the following options:
8049 Set output video size.
8052 Set x offset from where to pick pixels.
8055 Set y offset from where to pick pixels.
8058 Set scope mode, can be one of the following:
8061 Draw hexadecimal pixel values with white color on black background.
8064 Draw hexadecimal pixel values with input video pixel color on black
8068 Draw hexadecimal pixel values on color background picked from input video,
8069 the text color is picked in such way so its always visible.
8073 Draw rows and columns numbers on left and top of video.
8076 Set background opacity.
8081 Denoise frames using 2D DCT (frequency domain filtering).
8083 This filter is not designed for real time.
8085 The filter accepts the following options:
8089 Set the noise sigma constant.
8091 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8092 coefficient (absolute value) below this threshold with be dropped.
8094 If you need a more advanced filtering, see @option{expr}.
8096 Default is @code{0}.
8099 Set number overlapping pixels for each block. Since the filter can be slow, you
8100 may want to reduce this value, at the cost of a less effective filter and the
8101 risk of various artefacts.
8103 If the overlapping value doesn't permit processing the whole input width or
8104 height, a warning will be displayed and according borders won't be denoised.
8106 Default value is @var{blocksize}-1, which is the best possible setting.
8109 Set the coefficient factor expression.
8111 For each coefficient of a DCT block, this expression will be evaluated as a
8112 multiplier value for the coefficient.
8114 If this is option is set, the @option{sigma} option will be ignored.
8116 The absolute value of the coefficient can be accessed through the @var{c}
8120 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8121 @var{blocksize}, which is the width and height of the processed blocks.
8123 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8124 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8125 on the speed processing. Also, a larger block size does not necessarily means a
8129 @subsection Examples
8131 Apply a denoise with a @option{sigma} of @code{4.5}:
8136 The same operation can be achieved using the expression system:
8138 dctdnoiz=e='gte(c, 4.5*3)'
8141 Violent denoise using a block size of @code{16x16}:
8148 Remove banding artifacts from input video.
8149 It works by replacing banded pixels with average value of referenced pixels.
8151 The filter accepts the following options:
8158 Set banding detection threshold for each plane. Default is 0.02.
8159 Valid range is 0.00003 to 0.5.
8160 If difference between current pixel and reference pixel is less than threshold,
8161 it will be considered as banded.
8164 Banding detection range in pixels. Default is 16. If positive, random number
8165 in range 0 to set value will be used. If negative, exact absolute value
8167 The range defines square of four pixels around current pixel.
8170 Set direction in radians from which four pixel will be compared. If positive,
8171 random direction from 0 to set direction will be picked. If negative, exact of
8172 absolute value will be picked. For example direction 0, -PI or -2*PI radians
8173 will pick only pixels on same row and -PI/2 will pick only pixels on same
8177 If enabled, current pixel is compared with average value of all four
8178 surrounding pixels. The default is enabled. If disabled current pixel is
8179 compared with all four surrounding pixels. The pixel is considered banded
8180 if only all four differences with surrounding pixels are less than threshold.
8183 If enabled, current pixel is changed if and only if all pixel components are banded,
8184 e.g. banding detection threshold is triggered for all color components.
8185 The default is disabled.
8190 Remove blocking artifacts from input video.
8192 The filter accepts the following options:
8196 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
8197 This controls what kind of deblocking is applied.
8200 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
8206 Set blocking detection thresholds. Allowed range is 0 to 1.
8207 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
8208 Using higher threshold gives more deblocking strength.
8209 Setting @var{alpha} controls threshold detection at exact edge of block.
8210 Remaining options controls threshold detection near the edge. Each one for
8211 below/above or left/right. Setting any of those to @var{0} disables
8215 Set planes to filter. Default is to filter all available planes.
8218 @subsection Examples
8222 Deblock using weak filter and block size of 4 pixels.
8224 deblock=filter=weak:block=4
8228 Deblock using strong filter, block size of 4 pixels and custom thresholds for
8229 deblocking more edges.
8231 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
8235 Similar as above, but filter only first plane.
8237 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
8241 Similar as above, but filter only second and third plane.
8243 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
8250 Drop duplicated frames at regular intervals.
8252 The filter accepts the following options:
8256 Set the number of frames from which one will be dropped. Setting this to
8257 @var{N} means one frame in every batch of @var{N} frames will be dropped.
8258 Default is @code{5}.
8261 Set the threshold for duplicate detection. If the difference metric for a frame
8262 is less than or equal to this value, then it is declared as duplicate. Default
8266 Set scene change threshold. Default is @code{15}.
8270 Set the size of the x and y-axis blocks used during metric calculations.
8271 Larger blocks give better noise suppression, but also give worse detection of
8272 small movements. Must be a power of two. Default is @code{32}.
8275 Mark main input as a pre-processed input and activate clean source input
8276 stream. This allows the input to be pre-processed with various filters to help
8277 the metrics calculation while keeping the frame selection lossless. When set to
8278 @code{1}, the first stream is for the pre-processed input, and the second
8279 stream is the clean source from where the kept frames are chosen. Default is
8283 Set whether or not chroma is considered in the metric calculations. Default is
8289 Apply 2D deconvolution of video stream in frequency domain using second stream
8292 The filter accepts the following options:
8296 Set which planes to process.
8299 Set which impulse video frames will be processed, can be @var{first}
8300 or @var{all}. Default is @var{all}.
8303 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
8304 and height are not same and not power of 2 or if stream prior to convolving
8308 The @code{deconvolve} filter also supports the @ref{framesync} options.
8312 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
8314 It accepts the following options:
8318 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
8319 @var{rainbows} for cross-color reduction.
8322 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
8325 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
8328 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
8331 Set temporal chroma threshold. Lower values increases reduction of cross-color.
8336 Apply deflate effect to the video.
8338 This filter replaces the pixel by the local(3x3) average by taking into account
8339 only values lower than the pixel.
8341 It accepts the following options:
8348 Limit the maximum change for each plane, default is 65535.
8349 If 0, plane will remain unchanged.
8354 Remove temporal frame luminance variations.
8356 It accepts the following options:
8360 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
8363 Set averaging mode to smooth temporal luminance variations.
8365 Available values are:
8390 Do not actually modify frame. Useful when one only wants metadata.
8395 Remove judder produced by partially interlaced telecined content.
8397 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
8398 source was partially telecined content then the output of @code{pullup,dejudder}
8399 will have a variable frame rate. May change the recorded frame rate of the
8400 container. Aside from that change, this filter will not affect constant frame
8403 The option available in this filter is:
8407 Specify the length of the window over which the judder repeats.
8409 Accepts any integer greater than 1. Useful values are:
8413 If the original was telecined from 24 to 30 fps (Film to NTSC).
8416 If the original was telecined from 25 to 30 fps (PAL to NTSC).
8419 If a mixture of the two.
8422 The default is @samp{4}.
8427 Suppress a TV station logo by a simple interpolation of the surrounding
8428 pixels. Just set a rectangle covering the logo and watch it disappear
8429 (and sometimes something even uglier appear - your mileage may vary).
8431 It accepts the following parameters:
8436 Specify the top left corner coordinates of the logo. They must be
8441 Specify the width and height of the logo to clear. They must be
8445 Specify the thickness of the fuzzy edge of the rectangle (added to
8446 @var{w} and @var{h}). The default value is 1. This option is
8447 deprecated, setting higher values should no longer be necessary and
8451 When set to 1, a green rectangle is drawn on the screen to simplify
8452 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
8453 The default value is 0.
8455 The rectangle is drawn on the outermost pixels which will be (partly)
8456 replaced with interpolated values. The values of the next pixels
8457 immediately outside this rectangle in each direction will be used to
8458 compute the interpolated pixel values inside the rectangle.
8462 @subsection Examples
8466 Set a rectangle covering the area with top left corner coordinates 0,0
8467 and size 100x77, and a band of size 10:
8469 delogo=x=0:y=0:w=100:h=77:band=10
8476 Remove the rain in the input image/video by applying the derain methods based on
8477 convolutional neural networks. Supported models:
8481 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
8482 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
8485 Training as well as model generation scripts are provided in
8486 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
8488 Native model files (.model) can be generated from TensorFlow model
8489 files (.pb) by using tools/python/convert.py
8491 The filter accepts the following options:
8495 Specify which filter to use. This option accepts the following values:
8499 Derain filter. To conduct derain filter, you need to use a derain model.
8502 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
8504 Default value is @samp{derain}.
8507 Specify which DNN backend to use for model loading and execution. This option accepts
8508 the following values:
8512 Native implementation of DNN loading and execution.
8515 TensorFlow backend. To enable this backend you
8516 need to install the TensorFlow for C library (see
8517 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
8518 @code{--enable-libtensorflow}
8520 Default value is @samp{native}.
8523 Set path to model file specifying network architecture and its parameters.
8524 Note that different backends use different file formats. TensorFlow and native
8525 backend can load files for only its format.
8530 Attempt to fix small changes in horizontal and/or vertical shift. This
8531 filter helps remove camera shake from hand-holding a camera, bumping a
8532 tripod, moving on a vehicle, etc.
8534 The filter accepts the following options:
8542 Specify a rectangular area where to limit the search for motion
8544 If desired the search for motion vectors can be limited to a
8545 rectangular area of the frame defined by its top left corner, width
8546 and height. These parameters have the same meaning as the drawbox
8547 filter which can be used to visualise the position of the bounding
8550 This is useful when simultaneous movement of subjects within the frame
8551 might be confused for camera motion by the motion vector search.
8553 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
8554 then the full frame is used. This allows later options to be set
8555 without specifying the bounding box for the motion vector search.
8557 Default - search the whole frame.
8561 Specify the maximum extent of movement in x and y directions in the
8562 range 0-64 pixels. Default 16.
8565 Specify how to generate pixels to fill blanks at the edge of the
8566 frame. Available values are:
8569 Fill zeroes at blank locations
8571 Original image at blank locations
8573 Extruded edge value at blank locations
8575 Mirrored edge at blank locations
8577 Default value is @samp{mirror}.
8580 Specify the blocksize to use for motion search. Range 4-128 pixels,
8584 Specify the contrast threshold for blocks. Only blocks with more than
8585 the specified contrast (difference between darkest and lightest
8586 pixels) will be considered. Range 1-255, default 125.
8589 Specify the search strategy. Available values are:
8592 Set exhaustive search
8594 Set less exhaustive search.
8596 Default value is @samp{exhaustive}.
8599 If set then a detailed log of the motion search is written to the
8606 Remove unwanted contamination of foreground colors, caused by reflected color of
8607 greenscreen or bluescreen.
8609 This filter accepts the following options:
8613 Set what type of despill to use.
8616 Set how spillmap will be generated.
8619 Set how much to get rid of still remaining spill.
8622 Controls amount of red in spill area.
8625 Controls amount of green in spill area.
8626 Should be -1 for greenscreen.
8629 Controls amount of blue in spill area.
8630 Should be -1 for bluescreen.
8633 Controls brightness of spill area, preserving colors.
8636 Modify alpha from generated spillmap.
8641 Apply an exact inverse of the telecine operation. It requires a predefined
8642 pattern specified using the pattern option which must be the same as that passed
8643 to the telecine filter.
8645 This filter accepts the following options:
8654 The default value is @code{top}.
8658 A string of numbers representing the pulldown pattern you wish to apply.
8659 The default value is @code{23}.
8662 A number representing position of the first frame with respect to the telecine
8663 pattern. This is to be used if the stream is cut. The default value is @code{0}.
8668 Apply dilation effect to the video.
8670 This filter replaces the pixel by the local(3x3) maximum.
8672 It accepts the following options:
8679 Limit the maximum change for each plane, default is 65535.
8680 If 0, plane will remain unchanged.
8683 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
8686 Flags to local 3x3 coordinates maps like this:
8695 Displace pixels as indicated by second and third input stream.
8697 It takes three input streams and outputs one stream, the first input is the
8698 source, and second and third input are displacement maps.
8700 The second input specifies how much to displace pixels along the
8701 x-axis, while the third input specifies how much to displace pixels
8703 If one of displacement map streams terminates, last frame from that
8704 displacement map will be used.
8706 Note that once generated, displacements maps can be reused over and over again.
8708 A description of the accepted options follows.
8712 Set displace behavior for pixels that are out of range.
8714 Available values are:
8717 Missing pixels are replaced by black pixels.
8720 Adjacent pixels will spread out to replace missing pixels.
8723 Out of range pixels are wrapped so they point to pixels of other side.
8726 Out of range pixels will be replaced with mirrored pixels.
8728 Default is @samp{smear}.
8732 @subsection Examples
8736 Add ripple effect to rgb input of video size hd720:
8738 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
8742 Add wave effect to rgb input of video size hd720:
8744 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
8750 Draw a colored box on the input image.
8752 It accepts the following parameters:
8757 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
8761 The expressions which specify the width and height of the box; if 0 they are interpreted as
8762 the input width and height. It defaults to 0.
8765 Specify the color of the box to write. For the general syntax of this option,
8766 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8767 value @code{invert} is used, the box edge color is the same as the
8768 video with inverted luma.
8771 The expression which sets the thickness of the box edge.
8772 A value of @code{fill} will create a filled box. Default value is @code{3}.
8774 See below for the list of accepted constants.
8777 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
8778 will overwrite the video's color and alpha pixels.
8779 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
8782 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8783 following constants:
8787 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8791 horizontal and vertical chroma subsample values. For example for the
8792 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8796 The input width and height.
8799 The input sample aspect ratio.
8803 The x and y offset coordinates where the box is drawn.
8807 The width and height of the drawn box.
8810 The thickness of the drawn box.
8812 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8813 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8817 @subsection Examples
8821 Draw a black box around the edge of the input image:
8827 Draw a box with color red and an opacity of 50%:
8829 drawbox=10:20:200:60:red@@0.5
8832 The previous example can be specified as:
8834 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
8838 Fill the box with pink color:
8840 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
8844 Draw a 2-pixel red 2.40:1 mask:
8846 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
8852 Draw a grid on the input image.
8854 It accepts the following parameters:
8859 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
8863 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
8864 input width and height, respectively, minus @code{thickness}, so image gets
8865 framed. Default to 0.
8868 Specify the color of the grid. For the general syntax of this option,
8869 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8870 value @code{invert} is used, the grid color is the same as the
8871 video with inverted luma.
8874 The expression which sets the thickness of the grid line. Default value is @code{1}.
8876 See below for the list of accepted constants.
8879 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
8880 will overwrite the video's color and alpha pixels.
8881 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
8884 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8885 following constants:
8889 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8893 horizontal and vertical chroma subsample values. For example for the
8894 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8898 The input grid cell width and height.
8901 The input sample aspect ratio.
8905 The x and y coordinates of some point of grid intersection (meant to configure offset).
8909 The width and height of the drawn cell.
8912 The thickness of the drawn cell.
8914 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8915 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8919 @subsection Examples
8923 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
8925 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
8929 Draw a white 3x3 grid with an opacity of 50%:
8931 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
8938 Draw a text string or text from a specified file on top of a video, using the
8939 libfreetype library.
8941 To enable compilation of this filter, you need to configure FFmpeg with
8942 @code{--enable-libfreetype}.
8943 To enable default font fallback and the @var{font} option you need to
8944 configure FFmpeg with @code{--enable-libfontconfig}.
8945 To enable the @var{text_shaping} option, you need to configure FFmpeg with
8946 @code{--enable-libfribidi}.
8950 It accepts the following parameters:
8955 Used to draw a box around text using the background color.
8956 The value must be either 1 (enable) or 0 (disable).
8957 The default value of @var{box} is 0.
8960 Set the width of the border to be drawn around the box using @var{boxcolor}.
8961 The default value of @var{boxborderw} is 0.
8964 The color to be used for drawing box around text. For the syntax of this
8965 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8967 The default value of @var{boxcolor} is "white".
8970 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
8971 The default value of @var{line_spacing} is 0.
8974 Set the width of the border to be drawn around the text using @var{bordercolor}.
8975 The default value of @var{borderw} is 0.
8978 Set the color to be used for drawing border around text. For the syntax of this
8979 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8981 The default value of @var{bordercolor} is "black".
8984 Select how the @var{text} is expanded. Can be either @code{none},
8985 @code{strftime} (deprecated) or
8986 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
8990 Set a start time for the count. Value is in microseconds. Only applied
8991 in the deprecated strftime expansion mode. To emulate in normal expansion
8992 mode use the @code{pts} function, supplying the start time (in seconds)
8993 as the second argument.
8996 If true, check and fix text coords to avoid clipping.
8999 The color to be used for drawing fonts. For the syntax of this option, check
9000 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9002 The default value of @var{fontcolor} is "black".
9004 @item fontcolor_expr
9005 String which is expanded the same way as @var{text} to obtain dynamic
9006 @var{fontcolor} value. By default this option has empty value and is not
9007 processed. When this option is set, it overrides @var{fontcolor} option.
9010 The font family to be used for drawing text. By default Sans.
9013 The font file to be used for drawing text. The path must be included.
9014 This parameter is mandatory if the fontconfig support is disabled.
9017 Draw the text applying alpha blending. The value can
9018 be a number between 0.0 and 1.0.
9019 The expression accepts the same variables @var{x, y} as well.
9020 The default value is 1.
9021 Please see @var{fontcolor_expr}.
9024 The font size to be used for drawing text.
9025 The default value of @var{fontsize} is 16.
9028 If set to 1, attempt to shape the text (for example, reverse the order of
9029 right-to-left text and join Arabic characters) before drawing it.
9030 Otherwise, just draw the text exactly as given.
9031 By default 1 (if supported).
9034 The flags to be used for loading the fonts.
9036 The flags map the corresponding flags supported by libfreetype, and are
9037 a combination of the following values:
9044 @item vertical_layout
9045 @item force_autohint
9048 @item ignore_global_advance_width
9050 @item ignore_transform
9056 Default value is "default".
9058 For more information consult the documentation for the FT_LOAD_*
9062 The color to be used for drawing a shadow behind the drawn text. For the
9063 syntax of this option, check the @ref{color syntax,,"Color" section in the
9064 ffmpeg-utils manual,ffmpeg-utils}.
9066 The default value of @var{shadowcolor} is "black".
9070 The x and y offsets for the text shadow position with respect to the
9071 position of the text. They can be either positive or negative
9072 values. The default value for both is "0".
9075 The starting frame number for the n/frame_num variable. The default value
9079 The size in number of spaces to use for rendering the tab.
9083 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
9084 format. It can be used with or without text parameter. @var{timecode_rate}
9085 option must be specified.
9087 @item timecode_rate, rate, r
9088 Set the timecode frame rate (timecode only). Value will be rounded to nearest
9089 integer. Minimum value is "1".
9090 Drop-frame timecode is supported for frame rates 30 & 60.
9093 If set to 1, the output of the timecode option will wrap around at 24 hours.
9094 Default is 0 (disabled).
9097 The text string to be drawn. The text must be a sequence of UTF-8
9099 This parameter is mandatory if no file is specified with the parameter
9103 A text file containing text to be drawn. The text must be a sequence
9104 of UTF-8 encoded characters.
9106 This parameter is mandatory if no text string is specified with the
9107 parameter @var{text}.
9109 If both @var{text} and @var{textfile} are specified, an error is thrown.
9112 If set to 1, the @var{textfile} will be reloaded before each frame.
9113 Be sure to update it atomically, or it may be read partially, or even fail.
9117 The expressions which specify the offsets where text will be drawn
9118 within the video frame. They are relative to the top/left border of the
9121 The default value of @var{x} and @var{y} is "0".
9123 See below for the list of accepted constants and functions.
9126 The parameters for @var{x} and @var{y} are expressions containing the
9127 following constants and functions:
9131 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
9135 horizontal and vertical chroma subsample values. For example for the
9136 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9139 the height of each text line
9147 @item max_glyph_a, ascent
9148 the maximum distance from the baseline to the highest/upper grid
9149 coordinate used to place a glyph outline point, for all the rendered
9151 It is a positive value, due to the grid's orientation with the Y axis
9154 @item max_glyph_d, descent
9155 the maximum distance from the baseline to the lowest grid coordinate
9156 used to place a glyph outline point, for all the rendered glyphs.
9157 This is a negative value, due to the grid's orientation, with the Y axis
9161 maximum glyph height, that is the maximum height for all the glyphs
9162 contained in the rendered text, it is equivalent to @var{ascent} -
9166 maximum glyph width, that is the maximum width for all the glyphs
9167 contained in the rendered text
9170 the number of input frame, starting from 0
9172 @item rand(min, max)
9173 return a random number included between @var{min} and @var{max}
9176 The input sample aspect ratio.
9179 timestamp expressed in seconds, NAN if the input timestamp is unknown
9182 the height of the rendered text
9185 the width of the rendered text
9189 the x and y offset coordinates where the text is drawn.
9191 These parameters allow the @var{x} and @var{y} expressions to refer
9192 to each other, so you can for example specify @code{y=x/dar}.
9195 A one character description of the current frame's picture type.
9198 The current packet's position in the input file or stream
9199 (in bytes, from the start of the input). A value of -1 indicates
9200 this info is not available.
9203 The current packet's duration, in seconds.
9206 The current packet's size (in bytes).
9209 @anchor{drawtext_expansion}
9210 @subsection Text expansion
9212 If @option{expansion} is set to @code{strftime},
9213 the filter recognizes strftime() sequences in the provided text and
9214 expands them accordingly. Check the documentation of strftime(). This
9215 feature is deprecated.
9217 If @option{expansion} is set to @code{none}, the text is printed verbatim.
9219 If @option{expansion} is set to @code{normal} (which is the default),
9220 the following expansion mechanism is used.
9222 The backslash character @samp{\}, followed by any character, always expands to
9223 the second character.
9225 Sequences of the form @code{%@{...@}} are expanded. The text between the
9226 braces is a function name, possibly followed by arguments separated by ':'.
9227 If the arguments contain special characters or delimiters (':' or '@}'),
9228 they should be escaped.
9230 Note that they probably must also be escaped as the value for the
9231 @option{text} option in the filter argument string and as the filter
9232 argument in the filtergraph description, and possibly also for the shell,
9233 that makes up to four levels of escaping; using a text file avoids these
9236 The following functions are available:
9241 The expression evaluation result.
9243 It must take one argument specifying the expression to be evaluated,
9244 which accepts the same constants and functions as the @var{x} and
9245 @var{y} values. Note that not all constants should be used, for
9246 example the text size is not known when evaluating the expression, so
9247 the constants @var{text_w} and @var{text_h} will have an undefined
9250 @item expr_int_format, eif
9251 Evaluate the expression's value and output as formatted integer.
9253 The first argument is the expression to be evaluated, just as for the @var{expr} function.
9254 The second argument specifies the output format. Allowed values are @samp{x},
9255 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
9256 @code{printf} function.
9257 The third parameter is optional and sets the number of positions taken by the output.
9258 It can be used to add padding with zeros from the left.
9261 The time at which the filter is running, expressed in UTC.
9262 It can accept an argument: a strftime() format string.
9265 The time at which the filter is running, expressed in the local time zone.
9266 It can accept an argument: a strftime() format string.
9269 Frame metadata. Takes one or two arguments.
9271 The first argument is mandatory and specifies the metadata key.
9273 The second argument is optional and specifies a default value, used when the
9274 metadata key is not found or empty.
9276 Available metadata can be identified by inspecting entries
9277 starting with TAG included within each frame section
9278 printed by running @code{ffprobe -show_frames}.
9280 String metadata generated in filters leading to
9281 the drawtext filter are also available.
9284 The frame number, starting from 0.
9287 A one character description of the current picture type.
9290 The timestamp of the current frame.
9291 It can take up to three arguments.
9293 The first argument is the format of the timestamp; it defaults to @code{flt}
9294 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
9295 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
9296 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
9297 @code{localtime} stands for the timestamp of the frame formatted as
9298 local time zone time.
9300 The second argument is an offset added to the timestamp.
9302 If the format is set to @code{hms}, a third argument @code{24HH} may be
9303 supplied to present the hour part of the formatted timestamp in 24h format
9306 If the format is set to @code{localtime} or @code{gmtime},
9307 a third argument may be supplied: a strftime() format string.
9308 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
9311 @subsection Commands
9313 This filter supports altering parameters via commands:
9316 Alter existing filter parameters.
9318 Syntax for the argument is the same as for filter invocation, e.g.
9321 fontsize=56:fontcolor=green:text='Hello World'
9324 Full filter invocation with sendcmd would look like this:
9327 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
9331 If the entire argument can't be parsed or applied as valid values then the filter will
9332 continue with its existing parameters.
9334 @subsection Examples
9338 Draw "Test Text" with font FreeSerif, using the default values for the
9339 optional parameters.
9342 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
9346 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
9347 and y=50 (counting from the top-left corner of the screen), text is
9348 yellow with a red box around it. Both the text and the box have an
9352 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
9353 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
9356 Note that the double quotes are not necessary if spaces are not used
9357 within the parameter list.
9360 Show the text at the center of the video frame:
9362 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
9366 Show the text at a random position, switching to a new position every 30 seconds:
9368 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)"
9372 Show a text line sliding from right to left in the last row of the video
9373 frame. The file @file{LONG_LINE} is assumed to contain a single line
9376 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
9380 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
9382 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
9386 Draw a single green letter "g", at the center of the input video.
9387 The glyph baseline is placed at half screen height.
9389 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
9393 Show text for 1 second every 3 seconds:
9395 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
9399 Use fontconfig to set the font. Note that the colons need to be escaped.
9401 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
9405 Print the date of a real-time encoding (see strftime(3)):
9407 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
9411 Show text fading in and out (appearing/disappearing):
9414 DS=1.0 # display start
9415 DE=10.0 # display end
9416 FID=1.5 # fade in duration
9417 FOD=5 # fade out duration
9418 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 @}"
9422 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
9423 and the @option{fontsize} value are included in the @option{y} offset.
9425 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
9426 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
9431 For more information about libfreetype, check:
9432 @url{http://www.freetype.org/}.
9434 For more information about fontconfig, check:
9435 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
9437 For more information about libfribidi, check:
9438 @url{http://fribidi.org/}.
9442 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
9444 The filter accepts the following options:
9449 Set low and high threshold values used by the Canny thresholding
9452 The high threshold selects the "strong" edge pixels, which are then
9453 connected through 8-connectivity with the "weak" edge pixels selected
9454 by the low threshold.
9456 @var{low} and @var{high} threshold values must be chosen in the range
9457 [0,1], and @var{low} should be lesser or equal to @var{high}.
9459 Default value for @var{low} is @code{20/255}, and default value for @var{high}
9463 Define the drawing mode.
9467 Draw white/gray wires on black background.
9470 Mix the colors to create a paint/cartoon effect.
9473 Apply Canny edge detector on all selected planes.
9475 Default value is @var{wires}.
9478 Select planes for filtering. By default all available planes are filtered.
9481 @subsection Examples
9485 Standard edge detection with custom values for the hysteresis thresholding:
9487 edgedetect=low=0.1:high=0.4
9491 Painting effect without thresholding:
9493 edgedetect=mode=colormix:high=0
9499 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
9501 For each input image, the filter will compute the optimal mapping from
9502 the input to the output given the codebook length, that is the number
9503 of distinct output colors.
9505 This filter accepts the following options.
9508 @item codebook_length, l
9509 Set codebook length. The value must be a positive integer, and
9510 represents the number of distinct output colors. Default value is 256.
9513 Set the maximum number of iterations to apply for computing the optimal
9514 mapping. The higher the value the better the result and the higher the
9515 computation time. Default value is 1.
9518 Set a random seed, must be an integer included between 0 and
9519 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
9520 will try to use a good random seed on a best effort basis.
9523 Set pal8 output pixel format. This option does not work with codebook
9524 length greater than 256.
9529 Measure graylevel entropy in histogram of color channels of video frames.
9531 It accepts the following parameters:
9535 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
9537 @var{diff} mode measures entropy of histogram delta values, absolute differences
9538 between neighbour histogram values.
9542 Set brightness, contrast, saturation and approximate gamma adjustment.
9544 The filter accepts the following options:
9548 Set the contrast expression. The value must be a float value in range
9549 @code{-1000.0} to @code{1000.0}. The default value is "1".
9552 Set the brightness expression. The value must be a float value in
9553 range @code{-1.0} to @code{1.0}. The default value is "0".
9556 Set the saturation expression. The value must be a float in
9557 range @code{0.0} to @code{3.0}. The default value is "1".
9560 Set the gamma expression. The value must be a float in range
9561 @code{0.1} to @code{10.0}. The default value is "1".
9564 Set the gamma expression for red. The value must be a float in
9565 range @code{0.1} to @code{10.0}. The default value is "1".
9568 Set the gamma expression for green. The value must be a float in range
9569 @code{0.1} to @code{10.0}. The default value is "1".
9572 Set the gamma expression for blue. The value must be a float in range
9573 @code{0.1} to @code{10.0}. The default value is "1".
9576 Set the gamma weight expression. It can be used to reduce the effect
9577 of a high gamma value on bright image areas, e.g. keep them from
9578 getting overamplified and just plain white. The value must be a float
9579 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
9580 gamma correction all the way down while @code{1.0} leaves it at its
9581 full strength. Default is "1".
9584 Set when the expressions for brightness, contrast, saturation and
9585 gamma expressions are evaluated.
9587 It accepts the following values:
9590 only evaluate expressions once during the filter initialization or
9591 when a command is processed
9594 evaluate expressions for each incoming frame
9597 Default value is @samp{init}.
9600 The expressions accept the following parameters:
9603 frame count of the input frame starting from 0
9606 byte position of the corresponding packet in the input file, NAN if
9610 frame rate of the input video, NAN if the input frame rate is unknown
9613 timestamp expressed in seconds, NAN if the input timestamp is unknown
9616 @subsection Commands
9617 The filter supports the following commands:
9621 Set the contrast expression.
9624 Set the brightness expression.
9627 Set the saturation expression.
9630 Set the gamma expression.
9633 Set the gamma_r expression.
9636 Set gamma_g expression.
9639 Set gamma_b expression.
9642 Set gamma_weight expression.
9644 The command accepts the same syntax of the corresponding option.
9646 If the specified expression is not valid, it is kept at its current
9653 Apply erosion effect to the video.
9655 This filter replaces the pixel by the local(3x3) minimum.
9657 It accepts the following options:
9664 Limit the maximum change for each plane, default is 65535.
9665 If 0, plane will remain unchanged.
9668 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9671 Flags to local 3x3 coordinates maps like this:
9678 @section extractplanes
9680 Extract color channel components from input video stream into
9681 separate grayscale video streams.
9683 The filter accepts the following option:
9687 Set plane(s) to extract.
9689 Available values for planes are:
9700 Choosing planes not available in the input will result in an error.
9701 That means you cannot select @code{r}, @code{g}, @code{b} planes
9702 with @code{y}, @code{u}, @code{v} planes at same time.
9705 @subsection Examples
9709 Extract luma, u and v color channel component from input video frame
9710 into 3 grayscale outputs:
9712 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
9718 Apply a fade-in/out effect to the input video.
9720 It accepts the following parameters:
9724 The effect type can be either "in" for a fade-in, or "out" for a fade-out
9726 Default is @code{in}.
9728 @item start_frame, s
9729 Specify the number of the frame to start applying the fade
9730 effect at. Default is 0.
9733 The number of frames that the fade effect lasts. At the end of the
9734 fade-in effect, the output video will have the same intensity as the input video.
9735 At the end of the fade-out transition, the output video will be filled with the
9736 selected @option{color}.
9740 If set to 1, fade only alpha channel, if one exists on the input.
9743 @item start_time, st
9744 Specify the timestamp (in seconds) of the frame to start to apply the fade
9745 effect. If both start_frame and start_time are specified, the fade will start at
9746 whichever comes last. Default is 0.
9749 The number of seconds for which the fade effect has to last. At the end of the
9750 fade-in effect the output video will have the same intensity as the input video,
9751 at the end of the fade-out transition the output video will be filled with the
9752 selected @option{color}.
9753 If both duration and nb_frames are specified, duration is used. Default is 0
9754 (nb_frames is used by default).
9757 Specify the color of the fade. Default is "black".
9760 @subsection Examples
9764 Fade in the first 30 frames of video:
9769 The command above is equivalent to:
9775 Fade out the last 45 frames of a 200-frame video:
9778 fade=type=out:start_frame=155:nb_frames=45
9782 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
9784 fade=in:0:25, fade=out:975:25
9788 Make the first 5 frames yellow, then fade in from frame 5-24:
9790 fade=in:5:20:color=yellow
9794 Fade in alpha over first 25 frames of video:
9796 fade=in:0:25:alpha=1
9800 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
9802 fade=t=in:st=5.5:d=0.5
9808 Denoise frames using 3D FFT (frequency domain filtering).
9810 The filter accepts the following options:
9814 Set the noise sigma constant. This sets denoising strength.
9815 Default value is 1. Allowed range is from 0 to 30.
9816 Using very high sigma with low overlap may give blocking artifacts.
9819 Set amount of denoising. By default all detected noise is reduced.
9820 Default value is 1. Allowed range is from 0 to 1.
9823 Set size of block, Default is 4, can be 3, 4, 5 or 6.
9824 Actual size of block in pixels is 2 to power of @var{block}, so by default
9825 block size in pixels is 2^4 which is 16.
9828 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
9831 Set number of previous frames to use for denoising. By default is set to 0.
9834 Set number of next frames to to use for denoising. By default is set to 0.
9837 Set planes which will be filtered, by default are all available filtered
9842 Apply arbitrary expressions to samples in frequency domain
9846 Adjust the dc value (gain) of the luma plane of the image. The filter
9847 accepts an integer value in range @code{0} to @code{1000}. The default
9848 value is set to @code{0}.
9851 Adjust the dc value (gain) of the 1st chroma plane of the image. The
9852 filter accepts an integer value in range @code{0} to @code{1000}. The
9853 default value is set to @code{0}.
9856 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
9857 filter accepts an integer value in range @code{0} to @code{1000}. The
9858 default value is set to @code{0}.
9861 Set the frequency domain weight expression for the luma plane.
9864 Set the frequency domain weight expression for the 1st chroma plane.
9867 Set the frequency domain weight expression for the 2nd chroma plane.
9870 Set when the expressions are evaluated.
9872 It accepts the following values:
9875 Only evaluate expressions once during the filter initialization.
9878 Evaluate expressions for each incoming frame.
9881 Default value is @samp{init}.
9883 The filter accepts the following variables:
9886 The coordinates of the current sample.
9890 The width and height of the image.
9893 The number of input frame, starting from 0.
9896 @subsection Examples
9902 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
9908 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
9914 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
9920 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
9927 Extract a single field from an interlaced image using stride
9928 arithmetic to avoid wasting CPU time. The output frames are marked as
9931 The filter accepts the following options:
9935 Specify whether to extract the top (if the value is @code{0} or
9936 @code{top}) or the bottom field (if the value is @code{1} or
9942 Create new frames by copying the top and bottom fields from surrounding frames
9943 supplied as numbers by the hint file.
9947 Set file containing hints: absolute/relative frame numbers.
9949 There must be one line for each frame in a clip. Each line must contain two
9950 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
9951 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
9952 is current frame number for @code{absolute} mode or out of [-1, 1] range
9953 for @code{relative} mode. First number tells from which frame to pick up top
9954 field and second number tells from which frame to pick up bottom field.
9956 If optionally followed by @code{+} output frame will be marked as interlaced,
9957 else if followed by @code{-} output frame will be marked as progressive, else
9958 it will be marked same as input frame.
9959 If line starts with @code{#} or @code{;} that line is skipped.
9962 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
9965 Example of first several lines of @code{hint} file for @code{relative} mode:
9968 1,0 - # second frame, use third's frame top field and second's frame bottom field
9969 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
9986 Field matching filter for inverse telecine. It is meant to reconstruct the
9987 progressive frames from a telecined stream. The filter does not drop duplicated
9988 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
9989 followed by a decimation filter such as @ref{decimate} in the filtergraph.
9991 The separation of the field matching and the decimation is notably motivated by
9992 the possibility of inserting a de-interlacing filter fallback between the two.
9993 If the source has mixed telecined and real interlaced content,
9994 @code{fieldmatch} will not be able to match fields for the interlaced parts.
9995 But these remaining combed frames will be marked as interlaced, and thus can be
9996 de-interlaced by a later filter such as @ref{yadif} before decimation.
9998 In addition to the various configuration options, @code{fieldmatch} can take an
9999 optional second stream, activated through the @option{ppsrc} option. If
10000 enabled, the frames reconstruction will be based on the fields and frames from
10001 this second stream. This allows the first input to be pre-processed in order to
10002 help the various algorithms of the filter, while keeping the output lossless
10003 (assuming the fields are matched properly). Typically, a field-aware denoiser,
10004 or brightness/contrast adjustments can help.
10006 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
10007 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
10008 which @code{fieldmatch} is based on. While the semantic and usage are very
10009 close, some behaviour and options names can differ.
10011 The @ref{decimate} filter currently only works for constant frame rate input.
10012 If your input has mixed telecined (30fps) and progressive content with a lower
10013 framerate like 24fps use the following filterchain to produce the necessary cfr
10014 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
10016 The filter accepts the following options:
10020 Specify the assumed field order of the input stream. Available values are:
10024 Auto detect parity (use FFmpeg's internal parity value).
10026 Assume bottom field first.
10028 Assume top field first.
10031 Note that it is sometimes recommended not to trust the parity announced by the
10034 Default value is @var{auto}.
10037 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
10038 sense that it won't risk creating jerkiness due to duplicate frames when
10039 possible, but if there are bad edits or blended fields it will end up
10040 outputting combed frames when a good match might actually exist. On the other
10041 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
10042 but will almost always find a good frame if there is one. The other values are
10043 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
10044 jerkiness and creating duplicate frames versus finding good matches in sections
10045 with bad edits, orphaned fields, blended fields, etc.
10047 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
10049 Available values are:
10053 2-way matching (p/c)
10055 2-way matching, and trying 3rd match if still combed (p/c + n)
10057 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
10059 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
10060 still combed (p/c + n + u/b)
10062 3-way matching (p/c/n)
10064 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
10065 detected as combed (p/c/n + u/b)
10068 The parenthesis at the end indicate the matches that would be used for that
10069 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
10072 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
10075 Default value is @var{pc_n}.
10078 Mark the main input stream as a pre-processed input, and enable the secondary
10079 input stream as the clean source to pick the fields from. See the filter
10080 introduction for more details. It is similar to the @option{clip2} feature from
10083 Default value is @code{0} (disabled).
10086 Set the field to match from. It is recommended to set this to the same value as
10087 @option{order} unless you experience matching failures with that setting. In
10088 certain circumstances changing the field that is used to match from can have a
10089 large impact on matching performance. Available values are:
10093 Automatic (same value as @option{order}).
10095 Match from the bottom field.
10097 Match from the top field.
10100 Default value is @var{auto}.
10103 Set whether or not chroma is included during the match comparisons. In most
10104 cases it is recommended to leave this enabled. You should set this to @code{0}
10105 only if your clip has bad chroma problems such as heavy rainbowing or other
10106 artifacts. Setting this to @code{0} could also be used to speed things up at
10107 the cost of some accuracy.
10109 Default value is @code{1}.
10113 These define an exclusion band which excludes the lines between @option{y0} and
10114 @option{y1} from being included in the field matching decision. An exclusion
10115 band can be used to ignore subtitles, a logo, or other things that may
10116 interfere with the matching. @option{y0} sets the starting scan line and
10117 @option{y1} sets the ending line; all lines in between @option{y0} and
10118 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
10119 @option{y0} and @option{y1} to the same value will disable the feature.
10120 @option{y0} and @option{y1} defaults to @code{0}.
10123 Set the scene change detection threshold as a percentage of maximum change on
10124 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
10125 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
10126 @option{scthresh} is @code{[0.0, 100.0]}.
10128 Default value is @code{12.0}.
10131 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
10132 account the combed scores of matches when deciding what match to use as the
10133 final match. Available values are:
10137 No final matching based on combed scores.
10139 Combed scores are only used when a scene change is detected.
10141 Use combed scores all the time.
10144 Default is @var{sc}.
10147 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
10148 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
10149 Available values are:
10153 No forced calculation.
10155 Force p/c/n calculations.
10157 Force p/c/n/u/b calculations.
10160 Default value is @var{none}.
10163 This is the area combing threshold used for combed frame detection. This
10164 essentially controls how "strong" or "visible" combing must be to be detected.
10165 Larger values mean combing must be more visible and smaller values mean combing
10166 can be less visible or strong and still be detected. Valid settings are from
10167 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
10168 be detected as combed). This is basically a pixel difference value. A good
10169 range is @code{[8, 12]}.
10171 Default value is @code{9}.
10174 Sets whether or not chroma is considered in the combed frame decision. Only
10175 disable this if your source has chroma problems (rainbowing, etc.) that are
10176 causing problems for the combed frame detection with chroma enabled. Actually,
10177 using @option{chroma}=@var{0} is usually more reliable, except for the case
10178 where there is chroma only combing in the source.
10180 Default value is @code{0}.
10184 Respectively set the x-axis and y-axis size of the window used during combed
10185 frame detection. This has to do with the size of the area in which
10186 @option{combpel} pixels are required to be detected as combed for a frame to be
10187 declared combed. See the @option{combpel} parameter description for more info.
10188 Possible values are any number that is a power of 2 starting at 4 and going up
10191 Default value is @code{16}.
10194 The number of combed pixels inside any of the @option{blocky} by
10195 @option{blockx} size blocks on the frame for the frame to be detected as
10196 combed. While @option{cthresh} controls how "visible" the combing must be, this
10197 setting controls "how much" combing there must be in any localized area (a
10198 window defined by the @option{blockx} and @option{blocky} settings) on the
10199 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
10200 which point no frames will ever be detected as combed). This setting is known
10201 as @option{MI} in TFM/VFM vocabulary.
10203 Default value is @code{80}.
10206 @anchor{p/c/n/u/b meaning}
10207 @subsection p/c/n/u/b meaning
10209 @subsubsection p/c/n
10211 We assume the following telecined stream:
10214 Top fields: 1 2 2 3 4
10215 Bottom fields: 1 2 3 4 4
10218 The numbers correspond to the progressive frame the fields relate to. Here, the
10219 first two frames are progressive, the 3rd and 4th are combed, and so on.
10221 When @code{fieldmatch} is configured to run a matching from bottom
10222 (@option{field}=@var{bottom}) this is how this input stream get transformed:
10227 B 1 2 3 4 4 <-- matching reference
10236 As a result of the field matching, we can see that some frames get duplicated.
10237 To perform a complete inverse telecine, you need to rely on a decimation filter
10238 after this operation. See for instance the @ref{decimate} filter.
10240 The same operation now matching from top fields (@option{field}=@var{top})
10245 T 1 2 2 3 4 <-- matching reference
10255 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
10256 basically, they refer to the frame and field of the opposite parity:
10259 @item @var{p} matches the field of the opposite parity in the previous frame
10260 @item @var{c} matches the field of the opposite parity in the current frame
10261 @item @var{n} matches the field of the opposite parity in the next frame
10266 The @var{u} and @var{b} matching are a bit special in the sense that they match
10267 from the opposite parity flag. In the following examples, we assume that we are
10268 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
10269 'x' is placed above and below each matched fields.
10271 With bottom matching (@option{field}=@var{bottom}):
10276 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10277 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10285 With top matching (@option{field}=@var{top}):
10290 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10291 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10299 @subsection Examples
10301 Simple IVTC of a top field first telecined stream:
10303 fieldmatch=order=tff:combmatch=none, decimate
10306 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
10308 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
10311 @section fieldorder
10313 Transform the field order of the input video.
10315 It accepts the following parameters:
10320 The output field order. Valid values are @var{tff} for top field first or @var{bff}
10321 for bottom field first.
10324 The default value is @samp{tff}.
10326 The transformation is done by shifting the picture content up or down
10327 by one line, and filling the remaining line with appropriate picture content.
10328 This method is consistent with most broadcast field order converters.
10330 If the input video is not flagged as being interlaced, or it is already
10331 flagged as being of the required output field order, then this filter does
10332 not alter the incoming video.
10334 It is very useful when converting to or from PAL DV material,
10335 which is bottom field first.
10339 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
10342 @section fifo, afifo
10344 Buffer input images and send them when they are requested.
10346 It is mainly useful when auto-inserted by the libavfilter
10349 It does not take parameters.
10351 @section fillborders
10353 Fill borders of the input video, without changing video stream dimensions.
10354 Sometimes video can have garbage at the four edges and you may not want to
10355 crop video input to keep size multiple of some number.
10357 This filter accepts the following options:
10361 Number of pixels to fill from left border.
10364 Number of pixels to fill from right border.
10367 Number of pixels to fill from top border.
10370 Number of pixels to fill from bottom border.
10375 It accepts the following values:
10378 fill pixels using outermost pixels
10381 fill pixels using mirroring
10384 fill pixels with constant value
10387 Default is @var{smear}.
10390 Set color for pixels in fixed mode. Default is @var{black}.
10395 Find a rectangular object
10397 It accepts the following options:
10401 Filepath of the object image, needs to be in gray8.
10404 Detection threshold, default is 0.5.
10407 Number of mipmaps, default is 3.
10409 @item xmin, ymin, xmax, ymax
10410 Specifies the rectangle in which to search.
10413 @subsection Examples
10417 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
10419 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
10425 Flood area with values of same pixel components with another values.
10427 It accepts the following options:
10430 Set pixel x coordinate.
10433 Set pixel y coordinate.
10436 Set source #0 component value.
10439 Set source #1 component value.
10442 Set source #2 component value.
10445 Set source #3 component value.
10448 Set destination #0 component value.
10451 Set destination #1 component value.
10454 Set destination #2 component value.
10457 Set destination #3 component value.
10463 Convert the input video to one of the specified pixel formats.
10464 Libavfilter will try to pick one that is suitable as input to
10467 It accepts the following parameters:
10471 A '|'-separated list of pixel format names, such as
10472 "pix_fmts=yuv420p|monow|rgb24".
10476 @subsection Examples
10480 Convert the input video to the @var{yuv420p} format
10482 format=pix_fmts=yuv420p
10485 Convert the input video to any of the formats in the list
10487 format=pix_fmts=yuv420p|yuv444p|yuv410p
10494 Convert the video to specified constant frame rate by duplicating or dropping
10495 frames as necessary.
10497 It accepts the following parameters:
10501 The desired output frame rate. The default is @code{25}.
10504 Assume the first PTS should be the given value, in seconds. This allows for
10505 padding/trimming at the start of stream. By default, no assumption is made
10506 about the first frame's expected PTS, so no padding or trimming is done.
10507 For example, this could be set to 0 to pad the beginning with duplicates of
10508 the first frame if a video stream starts after the audio stream or to trim any
10509 frames with a negative PTS.
10512 Timestamp (PTS) rounding method.
10514 Possible values are:
10521 round towards -infinity
10523 round towards +infinity
10527 The default is @code{near}.
10530 Action performed when reading the last frame.
10532 Possible values are:
10535 Use same timestamp rounding method as used for other frames.
10537 Pass through last frame if input duration has not been reached yet.
10539 The default is @code{round}.
10543 Alternatively, the options can be specified as a flat string:
10544 @var{fps}[:@var{start_time}[:@var{round}]].
10546 See also the @ref{setpts} filter.
10548 @subsection Examples
10552 A typical usage in order to set the fps to 25:
10558 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
10560 fps=fps=film:round=near
10566 Pack two different video streams into a stereoscopic video, setting proper
10567 metadata on supported codecs. The two views should have the same size and
10568 framerate and processing will stop when the shorter video ends. Please note
10569 that you may conveniently adjust view properties with the @ref{scale} and
10572 It accepts the following parameters:
10576 The desired packing format. Supported values are:
10581 The views are next to each other (default).
10584 The views are on top of each other.
10587 The views are packed by line.
10590 The views are packed by column.
10593 The views are temporally interleaved.
10602 # Convert left and right views into a frame-sequential video
10603 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
10605 # Convert views into a side-by-side video with the same output resolution as the input
10606 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
10611 Change the frame rate by interpolating new video output frames from the source
10614 This filter is not designed to function correctly with interlaced media. If
10615 you wish to change the frame rate of interlaced media then you are required
10616 to deinterlace before this filter and re-interlace after this filter.
10618 A description of the accepted options follows.
10622 Specify the output frames per second. This option can also be specified
10623 as a value alone. The default is @code{50}.
10626 Specify the start of a range where the output frame will be created as a
10627 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10628 the default is @code{15}.
10631 Specify the end of a range where the output frame will be created as a
10632 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10633 the default is @code{240}.
10636 Specify the level at which a scene change is detected as a value between
10637 0 and 100 to indicate a new scene; a low value reflects a low
10638 probability for the current frame to introduce a new scene, while a higher
10639 value means the current frame is more likely to be one.
10640 The default is @code{8.2}.
10643 Specify flags influencing the filter process.
10645 Available value for @var{flags} is:
10648 @item scene_change_detect, scd
10649 Enable scene change detection using the value of the option @var{scene}.
10650 This flag is enabled by default.
10656 Select one frame every N-th frame.
10658 This filter accepts the following option:
10661 Select frame after every @code{step} frames.
10662 Allowed values are positive integers higher than 0. Default value is @code{1}.
10665 @section freezedetect
10667 Detect frozen video.
10669 This filter logs a message and sets frame metadata when it detects that the
10670 input video has no significant change in content during a specified duration.
10671 Video freeze detection calculates the mean average absolute difference of all
10672 the components of video frames and compares it to a noise floor.
10674 The printed times and duration are expressed in seconds. The
10675 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
10676 whose timestamp equals or exceeds the detection duration and it contains the
10677 timestamp of the first frame of the freeze. The
10678 @code{lavfi.freezedetect.freeze_duration} and
10679 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
10682 The filter accepts the following options:
10686 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
10687 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
10691 Set freeze duration until notification (default is 2 seconds).
10697 Apply a frei0r effect to the input video.
10699 To enable the compilation of this filter, you need to install the frei0r
10700 header and configure FFmpeg with @code{--enable-frei0r}.
10702 It accepts the following parameters:
10707 The name of the frei0r effect to load. If the environment variable
10708 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
10709 directories specified by the colon-separated list in @env{FREI0R_PATH}.
10710 Otherwise, the standard frei0r paths are searched, in this order:
10711 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
10712 @file{/usr/lib/frei0r-1/}.
10714 @item filter_params
10715 A '|'-separated list of parameters to pass to the frei0r effect.
10719 A frei0r effect parameter can be a boolean (its value is either
10720 "y" or "n"), a double, a color (specified as
10721 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
10722 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
10723 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
10724 a position (specified as @var{X}/@var{Y}, where
10725 @var{X} and @var{Y} are floating point numbers) and/or a string.
10727 The number and types of parameters depend on the loaded effect. If an
10728 effect parameter is not specified, the default value is set.
10730 @subsection Examples
10734 Apply the distort0r effect, setting the first two double parameters:
10736 frei0r=filter_name=distort0r:filter_params=0.5|0.01
10740 Apply the colordistance effect, taking a color as the first parameter:
10742 frei0r=colordistance:0.2/0.3/0.4
10743 frei0r=colordistance:violet
10744 frei0r=colordistance:0x112233
10748 Apply the perspective effect, specifying the top left and top right image
10751 frei0r=perspective:0.2/0.2|0.8/0.2
10755 For more information, see
10756 @url{http://frei0r.dyne.org}
10760 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
10762 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
10763 processing filter, one of them is performed once per block, not per pixel.
10764 This allows for much higher speed.
10766 The filter accepts the following options:
10770 Set quality. This option defines the number of levels for averaging. It accepts
10771 an integer in the range 4-5. Default value is @code{4}.
10774 Force a constant quantization parameter. It accepts an integer in range 0-63.
10775 If not set, the filter will use the QP from the video stream (if available).
10778 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
10779 more details but also more artifacts, while higher values make the image smoother
10780 but also blurrier. Default value is @code{0} − PSNR optimal.
10782 @item use_bframe_qp
10783 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
10784 option may cause flicker since the B-Frames have often larger QP. Default is
10785 @code{0} (not enabled).
10791 Apply Gaussian blur filter.
10793 The filter accepts the following options:
10797 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
10800 Set number of steps for Gaussian approximation. Default is @code{1}.
10803 Set which planes to filter. By default all planes are filtered.
10806 Set vertical sigma, if negative it will be same as @code{sigma}.
10807 Default is @code{-1}.
10812 Apply generic equation to each pixel.
10814 The filter accepts the following options:
10817 @item lum_expr, lum
10818 Set the luminance expression.
10820 Set the chrominance blue expression.
10822 Set the chrominance red expression.
10823 @item alpha_expr, a
10824 Set the alpha expression.
10826 Set the red expression.
10827 @item green_expr, g
10828 Set the green expression.
10830 Set the blue expression.
10833 The colorspace is selected according to the specified options. If one
10834 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
10835 options is specified, the filter will automatically select a YCbCr
10836 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
10837 @option{blue_expr} options is specified, it will select an RGB
10840 If one of the chrominance expression is not defined, it falls back on the other
10841 one. If no alpha expression is specified it will evaluate to opaque value.
10842 If none of chrominance expressions are specified, they will evaluate
10843 to the luminance expression.
10845 The expressions can use the following variables and functions:
10849 The sequential number of the filtered frame, starting from @code{0}.
10853 The coordinates of the current sample.
10857 The width and height of the image.
10861 Width and height scale depending on the currently filtered plane. It is the
10862 ratio between the corresponding luma plane number of pixels and the current
10863 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
10864 @code{0.5,0.5} for chroma planes.
10867 Time of the current frame, expressed in seconds.
10870 Return the value of the pixel at location (@var{x},@var{y}) of the current
10874 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
10878 Return the value of the pixel at location (@var{x},@var{y}) of the
10879 blue-difference chroma plane. Return 0 if there is no such plane.
10882 Return the value of the pixel at location (@var{x},@var{y}) of the
10883 red-difference chroma plane. Return 0 if there is no such plane.
10888 Return the value of the pixel at location (@var{x},@var{y}) of the
10889 red/green/blue component. Return 0 if there is no such component.
10892 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
10893 plane. Return 0 if there is no such plane.
10896 For functions, if @var{x} and @var{y} are outside the area, the value will be
10897 automatically clipped to the closer edge.
10899 @subsection Examples
10903 Flip the image horizontally:
10909 Generate a bidimensional sine wave, with angle @code{PI/3} and a
10910 wavelength of 100 pixels:
10912 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
10916 Generate a fancy enigmatic moving light:
10918 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
10922 Generate a quick emboss effect:
10924 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
10928 Modify RGB components depending on pixel position:
10930 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
10934 Create a radial gradient that is the same size as the input (also see
10935 the @ref{vignette} filter):
10937 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
10943 Fix the banding artifacts that are sometimes introduced into nearly flat
10944 regions by truncation to 8-bit color depth.
10945 Interpolate the gradients that should go where the bands are, and
10948 It is designed for playback only. Do not use it prior to
10949 lossy compression, because compression tends to lose the dither and
10950 bring back the bands.
10952 It accepts the following parameters:
10957 The maximum amount by which the filter will change any one pixel. This is also
10958 the threshold for detecting nearly flat regions. Acceptable values range from
10959 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
10963 The neighborhood to fit the gradient to. A larger radius makes for smoother
10964 gradients, but also prevents the filter from modifying the pixels near detailed
10965 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
10966 values will be clipped to the valid range.
10970 Alternatively, the options can be specified as a flat string:
10971 @var{strength}[:@var{radius}]
10973 @subsection Examples
10977 Apply the filter with a @code{3.5} strength and radius of @code{8}:
10983 Specify radius, omitting the strength (which will fall-back to the default
10991 @section graphmonitor, agraphmonitor
10992 Show various filtergraph stats.
10994 With this filter one can debug complete filtergraph.
10995 Especially issues with links filling with queued frames.
10997 The filter accepts the following options:
11001 Set video output size. Default is @var{hd720}.
11004 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
11007 Set output mode, can be @var{fulll} or @var{compact}.
11008 In @var{compact} mode only filters with some queued frames have displayed stats.
11011 Set flags which enable which stats are shown in video.
11013 Available values for flags are:
11016 Display number of queued frames in each link.
11018 @item frame_count_in
11019 Display number of frames taken from filter.
11021 @item frame_count_out
11022 Display number of frames given out from filter.
11025 Display current filtered frame pts.
11028 Display current filtered frame time.
11031 Display time base for filter link.
11034 Display used format for filter link.
11037 Display video size or number of audio channels in case of audio used by filter link.
11040 Display video frame rate or sample rate in case of audio used by filter link.
11044 Set upper limit for video rate of output stream, Default value is @var{25}.
11045 This guarantee that output video frame rate will not be higher than this value.
11049 A color constancy variation filter which estimates scene illumination via grey edge algorithm
11050 and corrects the scene colors accordingly.
11052 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
11054 The filter accepts the following options:
11058 The order of differentiation to be applied on the scene. Must be chosen in the range
11059 [0,2] and default value is 1.
11062 The Minkowski parameter to be used for calculating the Minkowski distance. Must
11063 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
11064 max value instead of calculating Minkowski distance.
11067 The standard deviation of Gaussian blur to be applied on the scene. Must be
11068 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
11069 can't be equal to 0 if @var{difford} is greater than 0.
11072 @subsection Examples
11078 greyedge=difford=1:minknorm=5:sigma=2
11084 greyedge=difford=1:minknorm=0:sigma=2
11092 Apply a Hald CLUT to a video stream.
11094 First input is the video stream to process, and second one is the Hald CLUT.
11095 The Hald CLUT input can be a simple picture or a complete video stream.
11097 The filter accepts the following options:
11101 Force termination when the shortest input terminates. Default is @code{0}.
11103 Continue applying the last CLUT after the end of the stream. A value of
11104 @code{0} disable the filter after the last frame of the CLUT is reached.
11105 Default is @code{1}.
11108 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
11109 filters share the same internals).
11111 This filter also supports the @ref{framesync} options.
11113 More information about the Hald CLUT can be found on Eskil Steenberg's website
11114 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
11116 @subsection Workflow examples
11118 @subsubsection Hald CLUT video stream
11120 Generate an identity Hald CLUT stream altered with various effects:
11122 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
11125 Note: make sure you use a lossless codec.
11127 Then use it with @code{haldclut} to apply it on some random stream:
11129 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
11132 The Hald CLUT will be applied to the 10 first seconds (duration of
11133 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
11134 to the remaining frames of the @code{mandelbrot} stream.
11136 @subsubsection Hald CLUT with preview
11138 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
11139 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
11140 biggest possible square starting at the top left of the picture. The remaining
11141 padding pixels (bottom or right) will be ignored. This area can be used to add
11142 a preview of the Hald CLUT.
11144 Typically, the following generated Hald CLUT will be supported by the
11145 @code{haldclut} filter:
11148 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
11149 pad=iw+320 [padded_clut];
11150 smptebars=s=320x256, split [a][b];
11151 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
11152 [main][b] overlay=W-320" -frames:v 1 clut.png
11155 It contains the original and a preview of the effect of the CLUT: SMPTE color
11156 bars are displayed on the right-top, and below the same color bars processed by
11159 Then, the effect of this Hald CLUT can be visualized with:
11161 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
11166 Flip the input video horizontally.
11168 For example, to horizontally flip the input video with @command{ffmpeg}:
11170 ffmpeg -i in.avi -vf "hflip" out.avi
11174 This filter applies a global color histogram equalization on a
11177 It can be used to correct video that has a compressed range of pixel
11178 intensities. The filter redistributes the pixel intensities to
11179 equalize their distribution across the intensity range. It may be
11180 viewed as an "automatically adjusting contrast filter". This filter is
11181 useful only for correcting degraded or poorly captured source
11184 The filter accepts the following options:
11188 Determine the amount of equalization to be applied. As the strength
11189 is reduced, the distribution of pixel intensities more-and-more
11190 approaches that of the input frame. The value must be a float number
11191 in the range [0,1] and defaults to 0.200.
11194 Set the maximum intensity that can generated and scale the output
11195 values appropriately. The strength should be set as desired and then
11196 the intensity can be limited if needed to avoid washing-out. The value
11197 must be a float number in the range [0,1] and defaults to 0.210.
11200 Set the antibanding level. If enabled the filter will randomly vary
11201 the luminance of output pixels by a small amount to avoid banding of
11202 the histogram. Possible values are @code{none}, @code{weak} or
11203 @code{strong}. It defaults to @code{none}.
11208 Compute and draw a color distribution histogram for the input video.
11210 The computed histogram is a representation of the color component
11211 distribution in an image.
11213 Standard histogram displays the color components distribution in an image.
11214 Displays color graph for each color component. Shows distribution of
11215 the Y, U, V, A or R, G, B components, depending on input format, in the
11216 current frame. Below each graph a color component scale meter is shown.
11218 The filter accepts the following options:
11222 Set height of level. Default value is @code{200}.
11223 Allowed range is [50, 2048].
11226 Set height of color scale. Default value is @code{12}.
11227 Allowed range is [0, 40].
11231 It accepts the following values:
11234 Per color component graphs are placed below each other.
11237 Per color component graphs are placed side by side.
11240 Presents information identical to that in the @code{parade}, except
11241 that the graphs representing color components are superimposed directly
11244 Default is @code{stack}.
11247 Set mode. Can be either @code{linear}, or @code{logarithmic}.
11248 Default is @code{linear}.
11251 Set what color components to display.
11252 Default is @code{7}.
11255 Set foreground opacity. Default is @code{0.7}.
11258 Set background opacity. Default is @code{0.5}.
11261 @subsection Examples
11266 Calculate and draw histogram:
11268 ffplay -i input -vf histogram
11276 This is a high precision/quality 3d denoise filter. It aims to reduce
11277 image noise, producing smooth images and making still images really
11278 still. It should enhance compressibility.
11280 It accepts the following optional parameters:
11284 A non-negative floating point number which specifies spatial luma strength.
11285 It defaults to 4.0.
11287 @item chroma_spatial
11288 A non-negative floating point number which specifies spatial chroma strength.
11289 It defaults to 3.0*@var{luma_spatial}/4.0.
11292 A floating point number which specifies luma temporal strength. It defaults to
11293 6.0*@var{luma_spatial}/4.0.
11296 A floating point number which specifies chroma temporal strength. It defaults to
11297 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
11300 @anchor{hwdownload}
11301 @section hwdownload
11303 Download hardware frames to system memory.
11305 The input must be in hardware frames, and the output a non-hardware format.
11306 Not all formats will be supported on the output - it may be necessary to insert
11307 an additional @option{format} filter immediately following in the graph to get
11308 the output in a supported format.
11312 Map hardware frames to system memory or to another device.
11314 This filter has several different modes of operation; which one is used depends
11315 on the input and output formats:
11318 Hardware frame input, normal frame output
11320 Map the input frames to system memory and pass them to the output. If the
11321 original hardware frame is later required (for example, after overlaying
11322 something else on part of it), the @option{hwmap} filter can be used again
11323 in the next mode to retrieve it.
11325 Normal frame input, hardware frame output
11327 If the input is actually a software-mapped hardware frame, then unmap it -
11328 that is, return the original hardware frame.
11330 Otherwise, a device must be provided. Create new hardware surfaces on that
11331 device for the output, then map them back to the software format at the input
11332 and give those frames to the preceding filter. This will then act like the
11333 @option{hwupload} filter, but may be able to avoid an additional copy when
11334 the input is already in a compatible format.
11336 Hardware frame input and output
11338 A device must be supplied for the output, either directly or with the
11339 @option{derive_device} option. The input and output devices must be of
11340 different types and compatible - the exact meaning of this is
11341 system-dependent, but typically it means that they must refer to the same
11342 underlying hardware context (for example, refer to the same graphics card).
11344 If the input frames were originally created on the output device, then unmap
11345 to retrieve the original frames.
11347 Otherwise, map the frames to the output device - create new hardware frames
11348 on the output corresponding to the frames on the input.
11351 The following additional parameters are accepted:
11355 Set the frame mapping mode. Some combination of:
11358 The mapped frame should be readable.
11360 The mapped frame should be writeable.
11362 The mapping will always overwrite the entire frame.
11364 This may improve performance in some cases, as the original contents of the
11365 frame need not be loaded.
11367 The mapping must not involve any copying.
11369 Indirect mappings to copies of frames are created in some cases where either
11370 direct mapping is not possible or it would have unexpected properties.
11371 Setting this flag ensures that the mapping is direct and will fail if that is
11374 Defaults to @var{read+write} if not specified.
11376 @item derive_device @var{type}
11377 Rather than using the device supplied at initialisation, instead derive a new
11378 device of type @var{type} from the device the input frames exist on.
11381 In a hardware to hardware mapping, map in reverse - create frames in the sink
11382 and map them back to the source. This may be necessary in some cases where
11383 a mapping in one direction is required but only the opposite direction is
11384 supported by the devices being used.
11386 This option is dangerous - it may break the preceding filter in undefined
11387 ways if there are any additional constraints on that filter's output.
11388 Do not use it without fully understanding the implications of its use.
11394 Upload system memory frames to hardware surfaces.
11396 The device to upload to must be supplied when the filter is initialised. If
11397 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
11400 @anchor{hwupload_cuda}
11401 @section hwupload_cuda
11403 Upload system memory frames to a CUDA device.
11405 It accepts the following optional parameters:
11409 The number of the CUDA device to use
11414 Apply a high-quality magnification filter designed for pixel art. This filter
11415 was originally created by Maxim Stepin.
11417 It accepts the following option:
11421 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
11422 @code{hq3x} and @code{4} for @code{hq4x}.
11423 Default is @code{3}.
11427 Stack input videos horizontally.
11429 All streams must be of same pixel format and of same height.
11431 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
11432 to create same output.
11434 The filter accepts the following option:
11438 Set number of input streams. Default is 2.
11441 If set to 1, force the output to terminate when the shortest input
11442 terminates. Default value is 0.
11447 Modify the hue and/or the saturation of the input.
11449 It accepts the following parameters:
11453 Specify the hue angle as a number of degrees. It accepts an expression,
11454 and defaults to "0".
11457 Specify the saturation in the [-10,10] range. It accepts an expression and
11461 Specify the hue angle as a number of radians. It accepts an
11462 expression, and defaults to "0".
11465 Specify the brightness in the [-10,10] range. It accepts an expression and
11469 @option{h} and @option{H} are mutually exclusive, and can't be
11470 specified at the same time.
11472 The @option{b}, @option{h}, @option{H} and @option{s} option values are
11473 expressions containing the following constants:
11477 frame count of the input frame starting from 0
11480 presentation timestamp of the input frame expressed in time base units
11483 frame rate of the input video, NAN if the input frame rate is unknown
11486 timestamp expressed in seconds, NAN if the input timestamp is unknown
11489 time base of the input video
11492 @subsection Examples
11496 Set the hue to 90 degrees and the saturation to 1.0:
11502 Same command but expressing the hue in radians:
11508 Rotate hue and make the saturation swing between 0
11509 and 2 over a period of 1 second:
11511 hue="H=2*PI*t: s=sin(2*PI*t)+1"
11515 Apply a 3 seconds saturation fade-in effect starting at 0:
11517 hue="s=min(t/3\,1)"
11520 The general fade-in expression can be written as:
11522 hue="s=min(0\, max((t-START)/DURATION\, 1))"
11526 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
11528 hue="s=max(0\, min(1\, (8-t)/3))"
11531 The general fade-out expression can be written as:
11533 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
11538 @subsection Commands
11540 This filter supports the following commands:
11546 Modify the hue and/or the saturation and/or brightness of the input video.
11547 The command accepts the same syntax of the corresponding option.
11549 If the specified expression is not valid, it is kept at its current
11553 @section hysteresis
11555 Grow first stream into second stream by connecting components.
11556 This makes it possible to build more robust edge masks.
11558 This filter accepts the following options:
11562 Set which planes will be processed as bitmap, unprocessed planes will be
11563 copied from first stream.
11564 By default value 0xf, all planes will be processed.
11567 Set threshold which is used in filtering. If pixel component value is higher than
11568 this value filter algorithm for connecting components is activated.
11569 By default value is 0.
11574 Detect video interlacing type.
11576 This filter tries to detect if the input frames are interlaced, progressive,
11577 top or bottom field first. It will also try to detect fields that are
11578 repeated between adjacent frames (a sign of telecine).
11580 Single frame detection considers only immediately adjacent frames when classifying each frame.
11581 Multiple frame detection incorporates the classification history of previous frames.
11583 The filter will log these metadata values:
11586 @item single.current_frame
11587 Detected type of current frame using single-frame detection. One of:
11588 ``tff'' (top field first), ``bff'' (bottom field first),
11589 ``progressive'', or ``undetermined''
11592 Cumulative number of frames detected as top field first using single-frame detection.
11595 Cumulative number of frames detected as top field first using multiple-frame detection.
11598 Cumulative number of frames detected as bottom field first using single-frame detection.
11600 @item multiple.current_frame
11601 Detected type of current frame using multiple-frame detection. One of:
11602 ``tff'' (top field first), ``bff'' (bottom field first),
11603 ``progressive'', or ``undetermined''
11606 Cumulative number of frames detected as bottom field first using multiple-frame detection.
11608 @item single.progressive
11609 Cumulative number of frames detected as progressive using single-frame detection.
11611 @item multiple.progressive
11612 Cumulative number of frames detected as progressive using multiple-frame detection.
11614 @item single.undetermined
11615 Cumulative number of frames that could not be classified using single-frame detection.
11617 @item multiple.undetermined
11618 Cumulative number of frames that could not be classified using multiple-frame detection.
11620 @item repeated.current_frame
11621 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
11623 @item repeated.neither
11624 Cumulative number of frames with no repeated field.
11627 Cumulative number of frames with the top field repeated from the previous frame's top field.
11629 @item repeated.bottom
11630 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
11633 The filter accepts the following options:
11637 Set interlacing threshold.
11639 Set progressive threshold.
11641 Threshold for repeated field detection.
11643 Number of frames after which a given frame's contribution to the
11644 statistics is halved (i.e., it contributes only 0.5 to its
11645 classification). The default of 0 means that all frames seen are given
11646 full weight of 1.0 forever.
11647 @item analyze_interlaced_flag
11648 When this is not 0 then idet will use the specified number of frames to determine
11649 if the interlaced flag is accurate, it will not count undetermined frames.
11650 If the flag is found to be accurate it will be used without any further
11651 computations, if it is found to be inaccurate it will be cleared without any
11652 further computations. This allows inserting the idet filter as a low computational
11653 method to clean up the interlaced flag
11658 Deinterleave or interleave fields.
11660 This filter allows one to process interlaced images fields without
11661 deinterlacing them. Deinterleaving splits the input frame into 2
11662 fields (so called half pictures). Odd lines are moved to the top
11663 half of the output image, even lines to the bottom half.
11664 You can process (filter) them independently and then re-interleave them.
11666 The filter accepts the following options:
11670 @item chroma_mode, c
11671 @item alpha_mode, a
11672 Available values for @var{luma_mode}, @var{chroma_mode} and
11673 @var{alpha_mode} are:
11679 @item deinterleave, d
11680 Deinterleave fields, placing one above the other.
11682 @item interleave, i
11683 Interleave fields. Reverse the effect of deinterleaving.
11685 Default value is @code{none}.
11687 @item luma_swap, ls
11688 @item chroma_swap, cs
11689 @item alpha_swap, as
11690 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
11695 Apply inflate effect to the video.
11697 This filter replaces the pixel by the local(3x3) average by taking into account
11698 only values higher than the pixel.
11700 It accepts the following options:
11707 Limit the maximum change for each plane, default is 65535.
11708 If 0, plane will remain unchanged.
11713 Simple interlacing filter from progressive contents. This interleaves upper (or
11714 lower) lines from odd frames with lower (or upper) lines from even frames,
11715 halving the frame rate and preserving image height.
11718 Original Original New Frame
11719 Frame 'j' Frame 'j+1' (tff)
11720 ========== =========== ==================
11721 Line 0 --------------------> Frame 'j' Line 0
11722 Line 1 Line 1 ----> Frame 'j+1' Line 1
11723 Line 2 ---------------------> Frame 'j' Line 2
11724 Line 3 Line 3 ----> Frame 'j+1' Line 3
11726 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
11729 It accepts the following optional parameters:
11733 This determines whether the interlaced frame is taken from the even
11734 (tff - default) or odd (bff) lines of the progressive frame.
11737 Vertical lowpass filter to avoid twitter interlacing and
11738 reduce moire patterns.
11742 Disable vertical lowpass filter
11745 Enable linear filter (default)
11748 Enable complex filter. This will slightly less reduce twitter and moire
11749 but better retain detail and subjective sharpness impression.
11756 Deinterlace input video by applying Donald Graft's adaptive kernel
11757 deinterling. Work on interlaced parts of a video to produce
11758 progressive frames.
11760 The description of the accepted parameters follows.
11764 Set the threshold which affects the filter's tolerance when
11765 determining if a pixel line must be processed. It must be an integer
11766 in the range [0,255] and defaults to 10. A value of 0 will result in
11767 applying the process on every pixels.
11770 Paint pixels exceeding the threshold value to white if set to 1.
11774 Set the fields order. Swap fields if set to 1, leave fields alone if
11778 Enable additional sharpening if set to 1. Default is 0.
11781 Enable twoway sharpening if set to 1. Default is 0.
11784 @subsection Examples
11788 Apply default values:
11790 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
11794 Enable additional sharpening:
11800 Paint processed pixels in white:
11808 Slowly update darker pixels.
11810 This filter makes short flashes of light appear longer.
11811 This filter accepts the following options:
11815 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
11818 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
11821 @section lenscorrection
11823 Correct radial lens distortion
11825 This filter can be used to correct for radial distortion as can result from the use
11826 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
11827 one can use tools available for example as part of opencv or simply trial-and-error.
11828 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
11829 and extract the k1 and k2 coefficients from the resulting matrix.
11831 Note that effectively the same filter is available in the open-source tools Krita and
11832 Digikam from the KDE project.
11834 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
11835 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
11836 brightness distribution, so you may want to use both filters together in certain
11837 cases, though you will have to take care of ordering, i.e. whether vignetting should
11838 be applied before or after lens correction.
11840 @subsection Options
11842 The filter accepts the following options:
11846 Relative x-coordinate of the focal point of the image, and thereby the center of the
11847 distortion. This value has a range [0,1] and is expressed as fractions of the image
11848 width. Default is 0.5.
11850 Relative y-coordinate of the focal point of the image, and thereby the center of the
11851 distortion. This value has a range [0,1] and is expressed as fractions of the image
11852 height. Default is 0.5.
11854 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
11855 no correction. Default is 0.
11857 Coefficient of the double quadratic correction term. This value has a range [-1,1].
11858 0 means no correction. Default is 0.
11861 The formula that generates the correction is:
11863 @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)
11865 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
11866 distances from the focal point in the source and target images, respectively.
11870 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
11872 The @code{lensfun} filter requires the camera make, camera model, and lens model
11873 to apply the lens correction. The filter will load the lensfun database and
11874 query it to find the corresponding camera and lens entries in the database. As
11875 long as these entries can be found with the given options, the filter can
11876 perform corrections on frames. Note that incomplete strings will result in the
11877 filter choosing the best match with the given options, and the filter will
11878 output the chosen camera and lens models (logged with level "info"). You must
11879 provide the make, camera model, and lens model as they are required.
11881 The filter accepts the following options:
11885 The make of the camera (for example, "Canon"). This option is required.
11888 The model of the camera (for example, "Canon EOS 100D"). This option is
11892 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
11893 option is required.
11896 The type of correction to apply. The following values are valid options:
11900 Enables fixing lens vignetting.
11903 Enables fixing lens geometry. This is the default.
11906 Enables fixing chromatic aberrations.
11909 Enables fixing lens vignetting and lens geometry.
11912 Enables fixing lens vignetting and chromatic aberrations.
11915 Enables fixing both lens geometry and chromatic aberrations.
11918 Enables all possible corrections.
11922 The focal length of the image/video (zoom; expected constant for video). For
11923 example, a 18--55mm lens has focal length range of [18--55], so a value in that
11924 range should be chosen when using that lens. Default 18.
11927 The aperture of the image/video (expected constant for video). Note that
11928 aperture is only used for vignetting correction. Default 3.5.
11930 @item focus_distance
11931 The focus distance of the image/video (expected constant for video). Note that
11932 focus distance is only used for vignetting and only slightly affects the
11933 vignetting correction process. If unknown, leave it at the default value (which
11937 The scale factor which is applied after transformation. After correction the
11938 video is no longer necessarily rectangular. This parameter controls how much of
11939 the resulting image is visible. The value 0 means that a value will be chosen
11940 automatically such that there is little or no unmapped area in the output
11941 image. 1.0 means that no additional scaling is done. Lower values may result
11942 in more of the corrected image being visible, while higher values may avoid
11943 unmapped areas in the output.
11945 @item target_geometry
11946 The target geometry of the output image/video. The following values are valid
11950 @item rectilinear (default)
11953 @item equirectangular
11954 @item fisheye_orthographic
11955 @item fisheye_stereographic
11956 @item fisheye_equisolid
11957 @item fisheye_thoby
11960 Apply the reverse of image correction (instead of correcting distortion, apply
11963 @item interpolation
11964 The type of interpolation used when correcting distortion. The following values
11969 @item linear (default)
11974 @subsection Examples
11978 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
11979 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
11983 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
11987 Apply the same as before, but only for the first 5 seconds of video.
11990 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
11997 Obtain the VMAF (Video Multi-Method Assessment Fusion)
11998 score between two input videos.
12000 The obtained VMAF score is printed through the logging system.
12002 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
12003 After installing the library it can be enabled using:
12004 @code{./configure --enable-libvmaf --enable-version3}.
12005 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
12007 The filter has following options:
12011 Set the model path which is to be used for SVM.
12012 Default value: @code{"vmaf_v0.6.1.pkl"}
12015 Set the file path to be used to store logs.
12018 Set the format of the log file (xml or json).
12020 @item enable_transform
12021 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
12022 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
12023 Default value: @code{false}
12026 Invokes the phone model which will generate VMAF scores higher than in the
12027 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
12030 Enables computing psnr along with vmaf.
12033 Enables computing ssim along with vmaf.
12036 Enables computing ms_ssim along with vmaf.
12039 Set the pool method (mean, min or harmonic mean) to be used for computing vmaf.
12042 Set number of threads to be used when computing vmaf.
12045 Set interval for frame subsampling used when computing vmaf.
12047 @item enable_conf_interval
12048 Enables confidence interval.
12051 This filter also supports the @ref{framesync} options.
12053 On the below examples the input file @file{main.mpg} being processed is
12054 compared with the reference file @file{ref.mpg}.
12057 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
12060 Example with options:
12062 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
12067 Limits the pixel components values to the specified range [min, max].
12069 The filter accepts the following options:
12073 Lower bound. Defaults to the lowest allowed value for the input.
12076 Upper bound. Defaults to the highest allowed value for the input.
12079 Specify which planes will be processed. Defaults to all available.
12086 The filter accepts the following options:
12090 Set the number of loops. Setting this value to -1 will result in infinite loops.
12094 Set maximal size in number of frames. Default is 0.
12097 Set first frame of loop. Default is 0.
12100 @subsection Examples
12104 Loop single first frame infinitely:
12106 loop=loop=-1:size=1:start=0
12110 Loop single first frame 10 times:
12112 loop=loop=10:size=1:start=0
12116 Loop 10 first frames 5 times:
12118 loop=loop=5:size=10:start=0
12124 Apply a 1D LUT to an input video.
12126 The filter accepts the following options:
12130 Set the 1D LUT file name.
12132 Currently supported formats:
12141 Select interpolation mode.
12143 Available values are:
12147 Use values from the nearest defined point.
12149 Interpolate values using the linear interpolation.
12151 Interpolate values using the cosine interpolation.
12153 Interpolate values using the cubic interpolation.
12155 Interpolate values using the spline interpolation.
12162 Apply a 3D LUT to an input video.
12164 The filter accepts the following options:
12168 Set the 3D LUT file name.
12170 Currently supported formats:
12184 Select interpolation mode.
12186 Available values are:
12190 Use values from the nearest defined point.
12192 Interpolate values using the 8 points defining a cube.
12194 Interpolate values using a tetrahedron.
12200 Turn certain luma values into transparency.
12202 The filter accepts the following options:
12206 Set the luma which will be used as base for transparency.
12207 Default value is @code{0}.
12210 Set the range of luma values to be keyed out.
12211 Default value is @code{0}.
12214 Set the range of softness. Default value is @code{0}.
12215 Use this to control gradual transition from zero to full transparency.
12218 @section lut, lutrgb, lutyuv
12220 Compute a look-up table for binding each pixel component input value
12221 to an output value, and apply it to the input video.
12223 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
12224 to an RGB input video.
12226 These filters accept the following parameters:
12229 set first pixel component expression
12231 set second pixel component expression
12233 set third pixel component expression
12235 set fourth pixel component expression, corresponds to the alpha component
12238 set red component expression
12240 set green component expression
12242 set blue component expression
12244 alpha component expression
12247 set Y/luminance component expression
12249 set U/Cb component expression
12251 set V/Cr component expression
12254 Each of them specifies the expression to use for computing the lookup table for
12255 the corresponding pixel component values.
12257 The exact component associated to each of the @var{c*} options depends on the
12260 The @var{lut} filter requires either YUV or RGB pixel formats in input,
12261 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
12263 The expressions can contain the following constants and functions:
12268 The input width and height.
12271 The input value for the pixel component.
12274 The input value, clipped to the @var{minval}-@var{maxval} range.
12277 The maximum value for the pixel component.
12280 The minimum value for the pixel component.
12283 The negated value for the pixel component value, clipped to the
12284 @var{minval}-@var{maxval} range; it corresponds to the expression
12285 "maxval-clipval+minval".
12288 The computed value in @var{val}, clipped to the
12289 @var{minval}-@var{maxval} range.
12291 @item gammaval(gamma)
12292 The computed gamma correction value of the pixel component value,
12293 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
12295 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
12299 All expressions default to "val".
12301 @subsection Examples
12305 Negate input video:
12307 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
12308 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
12311 The above is the same as:
12313 lutrgb="r=negval:g=negval:b=negval"
12314 lutyuv="y=negval:u=negval:v=negval"
12324 Remove chroma components, turning the video into a graytone image:
12326 lutyuv="u=128:v=128"
12330 Apply a luma burning effect:
12336 Remove green and blue components:
12342 Set a constant alpha channel value on input:
12344 format=rgba,lutrgb=a="maxval-minval/2"
12348 Correct luminance gamma by a factor of 0.5:
12350 lutyuv=y=gammaval(0.5)
12354 Discard least significant bits of luma:
12356 lutyuv=y='bitand(val, 128+64+32)'
12360 Technicolor like effect:
12362 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
12366 @section lut2, tlut2
12368 The @code{lut2} filter takes two input streams and outputs one
12371 The @code{tlut2} (time lut2) filter takes two consecutive frames
12372 from one single stream.
12374 This filter accepts the following parameters:
12377 set first pixel component expression
12379 set second pixel component expression
12381 set third pixel component expression
12383 set fourth pixel component expression, corresponds to the alpha component
12386 set output bit depth, only available for @code{lut2} filter. By default is 0,
12387 which means bit depth is automatically picked from first input format.
12390 Each of them specifies the expression to use for computing the lookup table for
12391 the corresponding pixel component values.
12393 The exact component associated to each of the @var{c*} options depends on the
12396 The expressions can contain the following constants:
12401 The input width and height.
12404 The first input value for the pixel component.
12407 The second input value for the pixel component.
12410 The first input video bit depth.
12413 The second input video bit depth.
12416 All expressions default to "x".
12418 @subsection Examples
12422 Highlight differences between two RGB video streams:
12424 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)'
12428 Highlight differences between two YUV video streams:
12430 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)'
12434 Show max difference between two video streams:
12436 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)))'
12440 @section maskedclamp
12442 Clamp the first input stream with the second input and third input stream.
12444 Returns the value of first stream to be between second input
12445 stream - @code{undershoot} and third input stream + @code{overshoot}.
12447 This filter accepts the following options:
12450 Default value is @code{0}.
12453 Default value is @code{0}.
12456 Set which planes will be processed as bitmap, unprocessed planes will be
12457 copied from first stream.
12458 By default value 0xf, all planes will be processed.
12461 @section maskedmerge
12463 Merge the first input stream with the second input stream using per pixel
12464 weights in the third input stream.
12466 A value of 0 in the third stream pixel component means that pixel component
12467 from first stream is returned unchanged, while maximum value (eg. 255 for
12468 8-bit videos) means that pixel component from second stream is returned
12469 unchanged. Intermediate values define the amount of merging between both
12470 input stream's pixel components.
12472 This filter accepts the following options:
12475 Set which planes will be processed as bitmap, unprocessed planes will be
12476 copied from first stream.
12477 By default value 0xf, all planes will be processed.
12481 Create mask from input video.
12483 For example it is useful to create motion masks after @code{tblend} filter.
12485 This filter accepts the following options:
12489 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
12492 Set high threshold. Any pixel component higher than this value will be set to max value
12493 allowed for current pixel format.
12496 Set planes to filter, by default all available planes are filtered.
12499 Fill all frame pixels with this value.
12502 Set max average pixel value for frame. If sum of all pixel components is higher that this
12503 average, output frame will be completely filled with value set by @var{fill} option.
12504 Typically useful for scene changes when used in combination with @code{tblend} filter.
12509 Apply motion-compensation deinterlacing.
12511 It needs one field per frame as input and must thus be used together
12512 with yadif=1/3 or equivalent.
12514 This filter accepts the following options:
12517 Set the deinterlacing mode.
12519 It accepts one of the following values:
12524 use iterative motion estimation
12526 like @samp{slow}, but use multiple reference frames.
12528 Default value is @samp{fast}.
12531 Set the picture field parity assumed for the input video. It must be
12532 one of the following values:
12536 assume top field first
12538 assume bottom field first
12541 Default value is @samp{bff}.
12544 Set per-block quantization parameter (QP) used by the internal
12547 Higher values should result in a smoother motion vector field but less
12548 optimal individual vectors. Default value is 1.
12551 @section mergeplanes
12553 Merge color channel components from several video streams.
12555 The filter accepts up to 4 input streams, and merge selected input
12556 planes to the output video.
12558 This filter accepts the following options:
12561 Set input to output plane mapping. Default is @code{0}.
12563 The mappings is specified as a bitmap. It should be specified as a
12564 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
12565 mapping for the first plane of the output stream. 'A' sets the number of
12566 the input stream to use (from 0 to 3), and 'a' the plane number of the
12567 corresponding input to use (from 0 to 3). The rest of the mappings is
12568 similar, 'Bb' describes the mapping for the output stream second
12569 plane, 'Cc' describes the mapping for the output stream third plane and
12570 'Dd' describes the mapping for the output stream fourth plane.
12573 Set output pixel format. Default is @code{yuva444p}.
12576 @subsection Examples
12580 Merge three gray video streams of same width and height into single video stream:
12582 [a0][a1][a2]mergeplanes=0x001020:yuv444p
12586 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
12588 [a0][a1]mergeplanes=0x00010210:yuva444p
12592 Swap Y and A plane in yuva444p stream:
12594 format=yuva444p,mergeplanes=0x03010200:yuva444p
12598 Swap U and V plane in yuv420p stream:
12600 format=yuv420p,mergeplanes=0x000201:yuv420p
12604 Cast a rgb24 clip to yuv444p:
12606 format=rgb24,mergeplanes=0x000102:yuv444p
12612 Estimate and export motion vectors using block matching algorithms.
12613 Motion vectors are stored in frame side data to be used by other filters.
12615 This filter accepts the following options:
12618 Specify the motion estimation method. Accepts one of the following values:
12622 Exhaustive search algorithm.
12624 Three step search algorithm.
12626 Two dimensional logarithmic search algorithm.
12628 New three step search algorithm.
12630 Four step search algorithm.
12632 Diamond search algorithm.
12634 Hexagon-based search algorithm.
12636 Enhanced predictive zonal search algorithm.
12638 Uneven multi-hexagon search algorithm.
12640 Default value is @samp{esa}.
12643 Macroblock size. Default @code{16}.
12646 Search parameter. Default @code{7}.
12649 @section midequalizer
12651 Apply Midway Image Equalization effect using two video streams.
12653 Midway Image Equalization adjusts a pair of images to have the same
12654 histogram, while maintaining their dynamics as much as possible. It's
12655 useful for e.g. matching exposures from a pair of stereo cameras.
12657 This filter has two inputs and one output, which must be of same pixel format, but
12658 may be of different sizes. The output of filter is first input adjusted with
12659 midway histogram of both inputs.
12661 This filter accepts the following option:
12665 Set which planes to process. Default is @code{15}, which is all available planes.
12668 @section minterpolate
12670 Convert the video to specified frame rate using motion interpolation.
12672 This filter accepts the following options:
12675 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}.
12678 Motion interpolation mode. Following values are accepted:
12681 Duplicate previous or next frame for interpolating new ones.
12683 Blend source frames. Interpolated frame is mean of previous and next frames.
12685 Motion compensated interpolation. Following options are effective when this mode is selected:
12689 Motion compensation mode. Following values are accepted:
12692 Overlapped block motion compensation.
12694 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
12696 Default mode is @samp{obmc}.
12699 Motion estimation mode. Following values are accepted:
12702 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
12704 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
12706 Default mode is @samp{bilat}.
12709 The algorithm to be used for motion estimation. Following values are accepted:
12712 Exhaustive search algorithm.
12714 Three step search algorithm.
12716 Two dimensional logarithmic search algorithm.
12718 New three step search algorithm.
12720 Four step search algorithm.
12722 Diamond search algorithm.
12724 Hexagon-based search algorithm.
12726 Enhanced predictive zonal search algorithm.
12728 Uneven multi-hexagon search algorithm.
12730 Default algorithm is @samp{epzs}.
12733 Macroblock size. Default @code{16}.
12736 Motion estimation search parameter. Default @code{32}.
12739 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).
12744 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:
12747 Disable scene change detection.
12749 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
12751 Default method is @samp{fdiff}.
12753 @item scd_threshold
12754 Scene change detection threshold. Default is @code{5.0}.
12759 Mix several video input streams into one video stream.
12761 A description of the accepted options follows.
12765 The number of inputs. If unspecified, it defaults to 2.
12768 Specify weight of each input video stream as sequence.
12769 Each weight is separated by space. If number of weights
12770 is smaller than number of @var{frames} last specified
12771 weight will be used for all remaining unset weights.
12774 Specify scale, if it is set it will be multiplied with sum
12775 of each weight multiplied with pixel values to give final destination
12776 pixel value. By default @var{scale} is auto scaled to sum of weights.
12779 Specify how end of stream is determined.
12782 The duration of the longest input. (default)
12785 The duration of the shortest input.
12788 The duration of the first input.
12792 @section mpdecimate
12794 Drop frames that do not differ greatly from the previous frame in
12795 order to reduce frame rate.
12797 The main use of this filter is for very-low-bitrate encoding
12798 (e.g. streaming over dialup modem), but it could in theory be used for
12799 fixing movies that were inverse-telecined incorrectly.
12801 A description of the accepted options follows.
12805 Set the maximum number of consecutive frames which can be dropped (if
12806 positive), or the minimum interval between dropped frames (if
12807 negative). If the value is 0, the frame is dropped disregarding the
12808 number of previous sequentially dropped frames.
12810 Default value is 0.
12815 Set the dropping threshold values.
12817 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
12818 represent actual pixel value differences, so a threshold of 64
12819 corresponds to 1 unit of difference for each pixel, or the same spread
12820 out differently over the block.
12822 A frame is a candidate for dropping if no 8x8 blocks differ by more
12823 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
12824 meaning the whole image) differ by more than a threshold of @option{lo}.
12826 Default value for @option{hi} is 64*12, default value for @option{lo} is
12827 64*5, and default value for @option{frac} is 0.33.
12833 Negate (invert) the input video.
12835 It accepts the following option:
12840 With value 1, it negates the alpha component, if present. Default value is 0.
12846 Denoise frames using Non-Local Means algorithm.
12848 Each pixel is adjusted by looking for other pixels with similar contexts. This
12849 context similarity is defined by comparing their surrounding patches of size
12850 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
12853 Note that the research area defines centers for patches, which means some
12854 patches will be made of pixels outside that research area.
12856 The filter accepts the following options.
12860 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
12863 Set patch size. Default is 7. Must be odd number in range [0, 99].
12866 Same as @option{p} but for chroma planes.
12868 The default value is @var{0} and means automatic.
12871 Set research size. Default is 15. Must be odd number in range [0, 99].
12874 Same as @option{r} but for chroma planes.
12876 The default value is @var{0} and means automatic.
12881 Deinterlace video using neural network edge directed interpolation.
12883 This filter accepts the following options:
12887 Mandatory option, without binary file filter can not work.
12888 Currently file can be found here:
12889 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
12892 Set which frames to deinterlace, by default it is @code{all}.
12893 Can be @code{all} or @code{interlaced}.
12896 Set mode of operation.
12898 Can be one of the following:
12902 Use frame flags, both fields.
12904 Use frame flags, single field.
12906 Use top field only.
12908 Use bottom field only.
12910 Use both fields, top first.
12912 Use both fields, bottom first.
12916 Set which planes to process, by default filter process all frames.
12919 Set size of local neighborhood around each pixel, used by the predictor neural
12922 Can be one of the following:
12935 Set the number of neurons in predictor neural network.
12936 Can be one of the following:
12947 Controls the number of different neural network predictions that are blended
12948 together to compute the final output value. Can be @code{fast}, default or
12952 Set which set of weights to use in the predictor.
12953 Can be one of the following:
12957 weights trained to minimize absolute error
12959 weights trained to minimize squared error
12963 Controls whether or not the prescreener neural network is used to decide
12964 which pixels should be processed by the predictor neural network and which
12965 can be handled by simple cubic interpolation.
12966 The prescreener is trained to know whether cubic interpolation will be
12967 sufficient for a pixel or whether it should be predicted by the predictor nn.
12968 The computational complexity of the prescreener nn is much less than that of
12969 the predictor nn. Since most pixels can be handled by cubic interpolation,
12970 using the prescreener generally results in much faster processing.
12971 The prescreener is pretty accurate, so the difference between using it and not
12972 using it is almost always unnoticeable.
12974 Can be one of the following:
12982 Default is @code{new}.
12985 Set various debugging flags.
12990 Force libavfilter not to use any of the specified pixel formats for the
12991 input to the next filter.
12993 It accepts the following parameters:
12997 A '|'-separated list of pixel format names, such as
12998 pix_fmts=yuv420p|monow|rgb24".
13002 @subsection Examples
13006 Force libavfilter to use a format different from @var{yuv420p} for the
13007 input to the vflip filter:
13009 noformat=pix_fmts=yuv420p,vflip
13013 Convert the input video to any of the formats not contained in the list:
13015 noformat=yuv420p|yuv444p|yuv410p
13021 Add noise on video input frame.
13023 The filter accepts the following options:
13031 Set noise seed for specific pixel component or all pixel components in case
13032 of @var{all_seed}. Default value is @code{123457}.
13034 @item all_strength, alls
13035 @item c0_strength, c0s
13036 @item c1_strength, c1s
13037 @item c2_strength, c2s
13038 @item c3_strength, c3s
13039 Set noise strength for specific pixel component or all pixel components in case
13040 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
13042 @item all_flags, allf
13043 @item c0_flags, c0f
13044 @item c1_flags, c1f
13045 @item c2_flags, c2f
13046 @item c3_flags, c3f
13047 Set pixel component flags or set flags for all components if @var{all_flags}.
13048 Available values for component flags are:
13051 averaged temporal noise (smoother)
13053 mix random noise with a (semi)regular pattern
13055 temporal noise (noise pattern changes between frames)
13057 uniform noise (gaussian otherwise)
13061 @subsection Examples
13063 Add temporal and uniform noise to input video:
13065 noise=alls=20:allf=t+u
13070 Normalize RGB video (aka histogram stretching, contrast stretching).
13071 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
13073 For each channel of each frame, the filter computes the input range and maps
13074 it linearly to the user-specified output range. The output range defaults
13075 to the full dynamic range from pure black to pure white.
13077 Temporal smoothing can be used on the input range to reduce flickering (rapid
13078 changes in brightness) caused when small dark or bright objects enter or leave
13079 the scene. This is similar to the auto-exposure (automatic gain control) on a
13080 video camera, and, like a video camera, it may cause a period of over- or
13081 under-exposure of the video.
13083 The R,G,B channels can be normalized independently, which may cause some
13084 color shifting, or linked together as a single channel, which prevents
13085 color shifting. Linked normalization preserves hue. Independent normalization
13086 does not, so it can be used to remove some color casts. Independent and linked
13087 normalization can be combined in any ratio.
13089 The normalize filter accepts the following options:
13094 Colors which define the output range. The minimum input value is mapped to
13095 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
13096 The defaults are black and white respectively. Specifying white for
13097 @var{blackpt} and black for @var{whitept} will give color-inverted,
13098 normalized video. Shades of grey can be used to reduce the dynamic range
13099 (contrast). Specifying saturated colors here can create some interesting
13103 The number of previous frames to use for temporal smoothing. The input range
13104 of each channel is smoothed using a rolling average over the current frame
13105 and the @var{smoothing} previous frames. The default is 0 (no temporal
13109 Controls the ratio of independent (color shifting) channel normalization to
13110 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
13111 independent. Defaults to 1.0 (fully independent).
13114 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
13115 expensive no-op. Defaults to 1.0 (full strength).
13119 @subsection Examples
13121 Stretch video contrast to use the full dynamic range, with no temporal
13122 smoothing; may flicker depending on the source content:
13124 normalize=blackpt=black:whitept=white:smoothing=0
13127 As above, but with 50 frames of temporal smoothing; flicker should be
13128 reduced, depending on the source content:
13130 normalize=blackpt=black:whitept=white:smoothing=50
13133 As above, but with hue-preserving linked channel normalization:
13135 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
13138 As above, but with half strength:
13140 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
13143 Map the darkest input color to red, the brightest input color to cyan:
13145 normalize=blackpt=red:whitept=cyan
13150 Pass the video source unchanged to the output.
13153 Optical Character Recognition
13155 This filter uses Tesseract for optical character recognition. To enable
13156 compilation of this filter, you need to configure FFmpeg with
13157 @code{--enable-libtesseract}.
13159 It accepts the following options:
13163 Set datapath to tesseract data. Default is to use whatever was
13164 set at installation.
13167 Set language, default is "eng".
13170 Set character whitelist.
13173 Set character blacklist.
13176 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
13177 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
13181 Apply a video transform using libopencv.
13183 To enable this filter, install the libopencv library and headers and
13184 configure FFmpeg with @code{--enable-libopencv}.
13186 It accepts the following parameters:
13191 The name of the libopencv filter to apply.
13193 @item filter_params
13194 The parameters to pass to the libopencv filter. If not specified, the default
13195 values are assumed.
13199 Refer to the official libopencv documentation for more precise
13201 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
13203 Several libopencv filters are supported; see the following subsections.
13208 Dilate an image by using a specific structuring element.
13209 It corresponds to the libopencv function @code{cvDilate}.
13211 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
13213 @var{struct_el} represents a structuring element, and has the syntax:
13214 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
13216 @var{cols} and @var{rows} represent the number of columns and rows of
13217 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
13218 point, and @var{shape} the shape for the structuring element. @var{shape}
13219 must be "rect", "cross", "ellipse", or "custom".
13221 If the value for @var{shape} is "custom", it must be followed by a
13222 string of the form "=@var{filename}". The file with name
13223 @var{filename} is assumed to represent a binary image, with each
13224 printable character corresponding to a bright pixel. When a custom
13225 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
13226 or columns and rows of the read file are assumed instead.
13228 The default value for @var{struct_el} is "3x3+0x0/rect".
13230 @var{nb_iterations} specifies the number of times the transform is
13231 applied to the image, and defaults to 1.
13235 # Use the default values
13238 # Dilate using a structuring element with a 5x5 cross, iterating two times
13239 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
13241 # Read the shape from the file diamond.shape, iterating two times.
13242 # The file diamond.shape may contain a pattern of characters like this
13248 # The specified columns and rows are ignored
13249 # but the anchor point coordinates are not
13250 ocv=dilate:0x0+2x2/custom=diamond.shape|2
13255 Erode an image by using a specific structuring element.
13256 It corresponds to the libopencv function @code{cvErode}.
13258 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
13259 with the same syntax and semantics as the @ref{dilate} filter.
13263 Smooth the input video.
13265 The filter takes the following parameters:
13266 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
13268 @var{type} is the type of smooth filter to apply, and must be one of
13269 the following values: "blur", "blur_no_scale", "median", "gaussian",
13270 or "bilateral". The default value is "gaussian".
13272 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
13273 depends on the smooth type. @var{param1} and
13274 @var{param2} accept integer positive values or 0. @var{param3} and
13275 @var{param4} accept floating point values.
13277 The default value for @var{param1} is 3. The default value for the
13278 other parameters is 0.
13280 These parameters correspond to the parameters assigned to the
13281 libopencv function @code{cvSmooth}.
13283 @section oscilloscope
13285 2D Video Oscilloscope.
13287 Useful to measure spatial impulse, step responses, chroma delays, etc.
13289 It accepts the following parameters:
13293 Set scope center x position.
13296 Set scope center y position.
13299 Set scope size, relative to frame diagonal.
13302 Set scope tilt/rotation.
13308 Set trace center x position.
13311 Set trace center y position.
13314 Set trace width, relative to width of frame.
13317 Set trace height, relative to height of frame.
13320 Set which components to trace. By default it traces first three components.
13323 Draw trace grid. By default is enabled.
13326 Draw some statistics. By default is enabled.
13329 Draw scope. By default is enabled.
13332 @subsection Examples
13336 Inspect full first row of video frame.
13338 oscilloscope=x=0.5:y=0:s=1
13342 Inspect full last row of video frame.
13344 oscilloscope=x=0.5:y=1:s=1
13348 Inspect full 5th line of video frame of height 1080.
13350 oscilloscope=x=0.5:y=5/1080:s=1
13354 Inspect full last column of video frame.
13356 oscilloscope=x=1:y=0.5:s=1:t=1
13364 Overlay one video on top of another.
13366 It takes two inputs and has one output. The first input is the "main"
13367 video on which the second input is overlaid.
13369 It accepts the following parameters:
13371 A description of the accepted options follows.
13376 Set the expression for the x and y coordinates of the overlaid video
13377 on the main video. Default value is "0" for both expressions. In case
13378 the expression is invalid, it is set to a huge value (meaning that the
13379 overlay will not be displayed within the output visible area).
13382 See @ref{framesync}.
13385 Set when the expressions for @option{x}, and @option{y} are evaluated.
13387 It accepts the following values:
13390 only evaluate expressions once during the filter initialization or
13391 when a command is processed
13394 evaluate expressions for each incoming frame
13397 Default value is @samp{frame}.
13400 See @ref{framesync}.
13403 Set the format for the output video.
13405 It accepts the following values:
13408 force YUV420 output
13411 force YUV422 output
13414 force YUV444 output
13417 force packed RGB output
13420 force planar RGB output
13423 automatically pick format
13426 Default value is @samp{yuv420}.
13429 See @ref{framesync}.
13432 Set format of alpha of the overlaid video, it can be @var{straight} or
13433 @var{premultiplied}. Default is @var{straight}.
13436 The @option{x}, and @option{y} expressions can contain the following
13442 The main input width and height.
13446 The overlay input width and height.
13450 The computed values for @var{x} and @var{y}. They are evaluated for
13455 horizontal and vertical chroma subsample values of the output
13456 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
13460 the number of input frame, starting from 0
13463 the position in the file of the input frame, NAN if unknown
13466 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
13470 This filter also supports the @ref{framesync} options.
13472 Note that the @var{n}, @var{pos}, @var{t} variables are available only
13473 when evaluation is done @emph{per frame}, and will evaluate to NAN
13474 when @option{eval} is set to @samp{init}.
13476 Be aware that frames are taken from each input video in timestamp
13477 order, hence, if their initial timestamps differ, it is a good idea
13478 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
13479 have them begin in the same zero timestamp, as the example for
13480 the @var{movie} filter does.
13482 You can chain together more overlays but you should test the
13483 efficiency of such approach.
13485 @subsection Commands
13487 This filter supports the following commands:
13491 Modify the x and y of the overlay input.
13492 The command accepts the same syntax of the corresponding option.
13494 If the specified expression is not valid, it is kept at its current
13498 @subsection Examples
13502 Draw the overlay at 10 pixels from the bottom right corner of the main
13505 overlay=main_w-overlay_w-10:main_h-overlay_h-10
13508 Using named options the example above becomes:
13510 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
13514 Insert a transparent PNG logo in the bottom left corner of the input,
13515 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
13517 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
13521 Insert 2 different transparent PNG logos (second logo on bottom
13522 right corner) using the @command{ffmpeg} tool:
13524 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
13528 Add a transparent color layer on top of the main video; @code{WxH}
13529 must specify the size of the main input to the overlay filter:
13531 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
13535 Play an original video and a filtered version (here with the deshake
13536 filter) side by side using the @command{ffplay} tool:
13538 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
13541 The above command is the same as:
13543 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
13547 Make a sliding overlay appearing from the left to the right top part of the
13548 screen starting since time 2:
13550 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
13554 Compose output by putting two input videos side to side:
13556 ffmpeg -i left.avi -i right.avi -filter_complex "
13557 nullsrc=size=200x100 [background];
13558 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
13559 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
13560 [background][left] overlay=shortest=1 [background+left];
13561 [background+left][right] overlay=shortest=1:x=100 [left+right]
13566 Mask 10-20 seconds of a video by applying the delogo filter to a section
13568 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
13569 -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]'
13574 Chain several overlays in cascade:
13576 nullsrc=s=200x200 [bg];
13577 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
13578 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
13579 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
13580 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
13581 [in3] null, [mid2] overlay=100:100 [out0]
13588 Apply Overcomplete Wavelet denoiser.
13590 The filter accepts the following options:
13596 Larger depth values will denoise lower frequency components more, but
13597 slow down filtering.
13599 Must be an int in the range 8-16, default is @code{8}.
13601 @item luma_strength, ls
13604 Must be a double value in the range 0-1000, default is @code{1.0}.
13606 @item chroma_strength, cs
13607 Set chroma strength.
13609 Must be a double value in the range 0-1000, default is @code{1.0}.
13615 Add paddings to the input image, and place the original input at the
13616 provided @var{x}, @var{y} coordinates.
13618 It accepts the following parameters:
13623 Specify an expression for the size of the output image with the
13624 paddings added. If the value for @var{width} or @var{height} is 0, the
13625 corresponding input size is used for the output.
13627 The @var{width} expression can reference the value set by the
13628 @var{height} expression, and vice versa.
13630 The default value of @var{width} and @var{height} is 0.
13634 Specify the offsets to place the input image at within the padded area,
13635 with respect to the top/left border of the output image.
13637 The @var{x} expression can reference the value set by the @var{y}
13638 expression, and vice versa.
13640 The default value of @var{x} and @var{y} is 0.
13642 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
13643 so the input image is centered on the padded area.
13646 Specify the color of the padded area. For the syntax of this option,
13647 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
13648 manual,ffmpeg-utils}.
13650 The default value of @var{color} is "black".
13653 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
13655 It accepts the following values:
13659 Only evaluate expressions once during the filter initialization or when
13660 a command is processed.
13663 Evaluate expressions for each incoming frame.
13667 Default value is @samp{init}.
13670 Pad to aspect instead to a resolution.
13674 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
13675 options are expressions containing the following constants:
13680 The input video width and height.
13684 These are the same as @var{in_w} and @var{in_h}.
13688 The output width and height (the size of the padded area), as
13689 specified by the @var{width} and @var{height} expressions.
13693 These are the same as @var{out_w} and @var{out_h}.
13697 The x and y offsets as specified by the @var{x} and @var{y}
13698 expressions, or NAN if not yet specified.
13701 same as @var{iw} / @var{ih}
13704 input sample aspect ratio
13707 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
13711 The horizontal and vertical chroma subsample values. For example for the
13712 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
13715 @subsection Examples
13719 Add paddings with the color "violet" to the input video. The output video
13720 size is 640x480, and the top-left corner of the input video is placed at
13723 pad=640:480:0:40:violet
13726 The example above is equivalent to the following command:
13728 pad=width=640:height=480:x=0:y=40:color=violet
13732 Pad the input to get an output with dimensions increased by 3/2,
13733 and put the input video at the center of the padded area:
13735 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
13739 Pad the input to get a squared output with size equal to the maximum
13740 value between the input width and height, and put the input video at
13741 the center of the padded area:
13743 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
13747 Pad the input to get a final w/h ratio of 16:9:
13749 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
13753 In case of anamorphic video, in order to set the output display aspect
13754 correctly, it is necessary to use @var{sar} in the expression,
13755 according to the relation:
13757 (ih * X / ih) * sar = output_dar
13758 X = output_dar / sar
13761 Thus the previous example needs to be modified to:
13763 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
13767 Double the output size and put the input video in the bottom-right
13768 corner of the output padded area:
13770 pad="2*iw:2*ih:ow-iw:oh-ih"
13774 @anchor{palettegen}
13775 @section palettegen
13777 Generate one palette for a whole video stream.
13779 It accepts the following options:
13783 Set the maximum number of colors to quantize in the palette.
13784 Note: the palette will still contain 256 colors; the unused palette entries
13787 @item reserve_transparent
13788 Create a palette of 255 colors maximum and reserve the last one for
13789 transparency. Reserving the transparency color is useful for GIF optimization.
13790 If not set, the maximum of colors in the palette will be 256. You probably want
13791 to disable this option for a standalone image.
13794 @item transparency_color
13795 Set the color that will be used as background for transparency.
13798 Set statistics mode.
13800 It accepts the following values:
13803 Compute full frame histograms.
13805 Compute histograms only for the part that differs from previous frame. This
13806 might be relevant to give more importance to the moving part of your input if
13807 the background is static.
13809 Compute new histogram for each frame.
13812 Default value is @var{full}.
13815 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
13816 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
13817 color quantization of the palette. This information is also visible at
13818 @var{info} logging level.
13820 @subsection Examples
13824 Generate a representative palette of a given video using @command{ffmpeg}:
13826 ffmpeg -i input.mkv -vf palettegen palette.png
13830 @section paletteuse
13832 Use a palette to downsample an input video stream.
13834 The filter takes two inputs: one video stream and a palette. The palette must
13835 be a 256 pixels image.
13837 It accepts the following options:
13841 Select dithering mode. Available algorithms are:
13844 Ordered 8x8 bayer dithering (deterministic)
13846 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
13847 Note: this dithering is sometimes considered "wrong" and is included as a
13849 @item floyd_steinberg
13850 Floyd and Steingberg dithering (error diffusion)
13852 Frankie Sierra dithering v2 (error diffusion)
13854 Frankie Sierra dithering v2 "Lite" (error diffusion)
13857 Default is @var{sierra2_4a}.
13860 When @var{bayer} dithering is selected, this option defines the scale of the
13861 pattern (how much the crosshatch pattern is visible). A low value means more
13862 visible pattern for less banding, and higher value means less visible pattern
13863 at the cost of more banding.
13865 The option must be an integer value in the range [0,5]. Default is @var{2}.
13868 If set, define the zone to process
13872 Only the changing rectangle will be reprocessed. This is similar to GIF
13873 cropping/offsetting compression mechanism. This option can be useful for speed
13874 if only a part of the image is changing, and has use cases such as limiting the
13875 scope of the error diffusal @option{dither} to the rectangle that bounds the
13876 moving scene (it leads to more deterministic output if the scene doesn't change
13877 much, and as a result less moving noise and better GIF compression).
13880 Default is @var{none}.
13883 Take new palette for each output frame.
13885 @item alpha_threshold
13886 Sets the alpha threshold for transparency. Alpha values above this threshold
13887 will be treated as completely opaque, and values below this threshold will be
13888 treated as completely transparent.
13890 The option must be an integer value in the range [0,255]. Default is @var{128}.
13893 @subsection Examples
13897 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
13898 using @command{ffmpeg}:
13900 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
13904 @section perspective
13906 Correct perspective of video not recorded perpendicular to the screen.
13908 A description of the accepted parameters follows.
13919 Set coordinates expression for top left, top right, bottom left and bottom right corners.
13920 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
13921 If the @code{sense} option is set to @code{source}, then the specified points will be sent
13922 to the corners of the destination. If the @code{sense} option is set to @code{destination},
13923 then the corners of the source will be sent to the specified coordinates.
13925 The expressions can use the following variables:
13930 the width and height of video frame.
13934 Output frame count.
13937 @item interpolation
13938 Set interpolation for perspective correction.
13940 It accepts the following values:
13946 Default value is @samp{linear}.
13949 Set interpretation of coordinate options.
13951 It accepts the following values:
13955 Send point in the source specified by the given coordinates to
13956 the corners of the destination.
13958 @item 1, destination
13960 Send the corners of the source to the point in the destination specified
13961 by the given coordinates.
13963 Default value is @samp{source}.
13967 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
13969 It accepts the following values:
13972 only evaluate expressions once during the filter initialization or
13973 when a command is processed
13976 evaluate expressions for each incoming frame
13979 Default value is @samp{init}.
13984 Delay interlaced video by one field time so that the field order changes.
13986 The intended use is to fix PAL movies that have been captured with the
13987 opposite field order to the film-to-video transfer.
13989 A description of the accepted parameters follows.
13995 It accepts the following values:
13998 Capture field order top-first, transfer bottom-first.
13999 Filter will delay the bottom field.
14002 Capture field order bottom-first, transfer top-first.
14003 Filter will delay the top field.
14006 Capture and transfer with the same field order. This mode only exists
14007 for the documentation of the other options to refer to, but if you
14008 actually select it, the filter will faithfully do nothing.
14011 Capture field order determined automatically by field flags, transfer
14013 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
14014 basis using field flags. If no field information is available,
14015 then this works just like @samp{u}.
14018 Capture unknown or varying, transfer opposite.
14019 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
14020 analyzing the images and selecting the alternative that produces best
14021 match between the fields.
14024 Capture top-first, transfer unknown or varying.
14025 Filter selects among @samp{t} and @samp{p} using image analysis.
14028 Capture bottom-first, transfer unknown or varying.
14029 Filter selects among @samp{b} and @samp{p} using image analysis.
14032 Capture determined by field flags, transfer unknown or varying.
14033 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
14034 image analysis. If no field information is available, then this works just
14035 like @samp{U}. This is the default mode.
14038 Both capture and transfer unknown or varying.
14039 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
14043 @section pixdesctest
14045 Pixel format descriptor test filter, mainly useful for internal
14046 testing. The output video should be equal to the input video.
14050 format=monow, pixdesctest
14053 can be used to test the monowhite pixel format descriptor definition.
14057 Display sample values of color channels. Mainly useful for checking color
14058 and levels. Minimum supported resolution is 640x480.
14060 The filters accept the following options:
14064 Set scope X position, relative offset on X axis.
14067 Set scope Y position, relative offset on Y axis.
14076 Set window opacity. This window also holds statistics about pixel area.
14079 Set window X position, relative offset on X axis.
14082 Set window Y position, relative offset on Y axis.
14087 Enable the specified chain of postprocessing subfilters using libpostproc. This
14088 library should be automatically selected with a GPL build (@code{--enable-gpl}).
14089 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
14090 Each subfilter and some options have a short and a long name that can be used
14091 interchangeably, i.e. dr/dering are the same.
14093 The filters accept the following options:
14097 Set postprocessing subfilters string.
14100 All subfilters share common options to determine their scope:
14104 Honor the quality commands for this subfilter.
14107 Do chrominance filtering, too (default).
14110 Do luminance filtering only (no chrominance).
14113 Do chrominance filtering only (no luminance).
14116 These options can be appended after the subfilter name, separated by a '|'.
14118 Available subfilters are:
14121 @item hb/hdeblock[|difference[|flatness]]
14122 Horizontal deblocking filter
14125 Difference factor where higher values mean more deblocking (default: @code{32}).
14127 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14130 @item vb/vdeblock[|difference[|flatness]]
14131 Vertical deblocking filter
14134 Difference factor where higher values mean more deblocking (default: @code{32}).
14136 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14139 @item ha/hadeblock[|difference[|flatness]]
14140 Accurate horizontal deblocking filter
14143 Difference factor where higher values mean more deblocking (default: @code{32}).
14145 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14148 @item va/vadeblock[|difference[|flatness]]
14149 Accurate vertical deblocking filter
14152 Difference factor where higher values mean more deblocking (default: @code{32}).
14154 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14158 The horizontal and vertical deblocking filters share the difference and
14159 flatness values so you cannot set different horizontal and vertical
14163 @item h1/x1hdeblock
14164 Experimental horizontal deblocking filter
14166 @item v1/x1vdeblock
14167 Experimental vertical deblocking filter
14172 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
14175 larger -> stronger filtering
14177 larger -> stronger filtering
14179 larger -> stronger filtering
14182 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
14185 Stretch luminance to @code{0-255}.
14188 @item lb/linblenddeint
14189 Linear blend deinterlacing filter that deinterlaces the given block by
14190 filtering all lines with a @code{(1 2 1)} filter.
14192 @item li/linipoldeint
14193 Linear interpolating deinterlacing filter that deinterlaces the given block by
14194 linearly interpolating every second line.
14196 @item ci/cubicipoldeint
14197 Cubic interpolating deinterlacing filter deinterlaces the given block by
14198 cubically interpolating every second line.
14200 @item md/mediandeint
14201 Median deinterlacing filter that deinterlaces the given block by applying a
14202 median filter to every second line.
14204 @item fd/ffmpegdeint
14205 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
14206 second line with a @code{(-1 4 2 4 -1)} filter.
14209 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
14210 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
14212 @item fq/forceQuant[|quantizer]
14213 Overrides the quantizer table from the input with the constant quantizer you
14221 Default pp filter combination (@code{hb|a,vb|a,dr|a})
14224 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
14227 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
14230 @subsection Examples
14234 Apply horizontal and vertical deblocking, deringing and automatic
14235 brightness/contrast:
14241 Apply default filters without brightness/contrast correction:
14247 Apply default filters and temporal denoiser:
14249 pp=default/tmpnoise|1|2|3
14253 Apply deblocking on luminance only, and switch vertical deblocking on or off
14254 automatically depending on available CPU time:
14261 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
14262 similar to spp = 6 with 7 point DCT, where only the center sample is
14265 The filter accepts the following options:
14269 Force a constant quantization parameter. It accepts an integer in range
14270 0 to 63. If not set, the filter will use the QP from the video stream
14274 Set thresholding mode. Available modes are:
14278 Set hard thresholding.
14280 Set soft thresholding (better de-ringing effect, but likely blurrier).
14282 Set medium thresholding (good results, default).
14286 @section premultiply
14287 Apply alpha premultiply effect to input video stream using first plane
14288 of second stream as alpha.
14290 Both streams must have same dimensions and same pixel format.
14292 The filter accepts the following option:
14296 Set which planes will be processed, unprocessed planes will be copied.
14297 By default value 0xf, all planes will be processed.
14300 Do not require 2nd input for processing, instead use alpha plane from input stream.
14304 Apply prewitt operator to input video stream.
14306 The filter accepts the following option:
14310 Set which planes will be processed, unprocessed planes will be copied.
14311 By default value 0xf, all planes will be processed.
14314 Set value which will be multiplied with filtered result.
14317 Set value which will be added to filtered result.
14320 @anchor{program_opencl}
14321 @section program_opencl
14323 Filter video using an OpenCL program.
14328 OpenCL program source file.
14331 Kernel name in program.
14334 Number of inputs to the filter. Defaults to 1.
14337 Size of output frames. Defaults to the same as the first input.
14341 The program source file must contain a kernel function with the given name,
14342 which will be run once for each plane of the output. Each run on a plane
14343 gets enqueued as a separate 2D global NDRange with one work-item for each
14344 pixel to be generated. The global ID offset for each work-item is therefore
14345 the coordinates of a pixel in the destination image.
14347 The kernel function needs to take the following arguments:
14350 Destination image, @var{__write_only image2d_t}.
14352 This image will become the output; the kernel should write all of it.
14354 Frame index, @var{unsigned int}.
14356 This is a counter starting from zero and increasing by one for each frame.
14358 Source images, @var{__read_only image2d_t}.
14360 These are the most recent images on each input. The kernel may read from
14361 them to generate the output, but they can't be written to.
14368 Copy the input to the output (output must be the same size as the input).
14370 __kernel void copy(__write_only image2d_t destination,
14371 unsigned int index,
14372 __read_only image2d_t source)
14374 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
14376 int2 location = (int2)(get_global_id(0), get_global_id(1));
14378 float4 value = read_imagef(source, sampler, location);
14380 write_imagef(destination, location, value);
14385 Apply a simple transformation, rotating the input by an amount increasing
14386 with the index counter. Pixel values are linearly interpolated by the
14387 sampler, and the output need not have the same dimensions as the input.
14389 __kernel void rotate_image(__write_only image2d_t dst,
14390 unsigned int index,
14391 __read_only image2d_t src)
14393 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
14394 CLK_FILTER_LINEAR);
14396 float angle = (float)index / 100.0f;
14398 float2 dst_dim = convert_float2(get_image_dim(dst));
14399 float2 src_dim = convert_float2(get_image_dim(src));
14401 float2 dst_cen = dst_dim / 2.0f;
14402 float2 src_cen = src_dim / 2.0f;
14404 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
14406 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
14408 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
14409 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
14411 src_pos = src_pos * src_dim / dst_dim;
14413 float2 src_loc = src_pos + src_cen;
14415 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
14416 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
14417 write_imagef(dst, dst_loc, 0.5f);
14419 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
14424 Blend two inputs together, with the amount of each input used varying
14425 with the index counter.
14427 __kernel void blend_images(__write_only image2d_t dst,
14428 unsigned int index,
14429 __read_only image2d_t src1,
14430 __read_only image2d_t src2)
14432 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
14433 CLK_FILTER_LINEAR);
14435 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
14437 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
14438 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
14439 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
14441 float4 val1 = read_imagef(src1, sampler, src1_loc);
14442 float4 val2 = read_imagef(src2, sampler, src2_loc);
14444 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
14450 @section pseudocolor
14452 Alter frame colors in video with pseudocolors.
14454 This filter accepts the following options:
14458 set pixel first component expression
14461 set pixel second component expression
14464 set pixel third component expression
14467 set pixel fourth component expression, corresponds to the alpha component
14470 set component to use as base for altering colors
14473 Each of them specifies the expression to use for computing the lookup table for
14474 the corresponding pixel component values.
14476 The expressions can contain the following constants and functions:
14481 The input width and height.
14484 The input value for the pixel component.
14486 @item ymin, umin, vmin, amin
14487 The minimum allowed component value.
14489 @item ymax, umax, vmax, amax
14490 The maximum allowed component value.
14493 All expressions default to "val".
14495 @subsection Examples
14499 Change too high luma values to gradient:
14501 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'"
14507 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
14508 Ratio) between two input videos.
14510 This filter takes in input two input videos, the first input is
14511 considered the "main" source and is passed unchanged to the
14512 output. The second input is used as a "reference" video for computing
14515 Both video inputs must have the same resolution and pixel format for
14516 this filter to work correctly. Also it assumes that both inputs
14517 have the same number of frames, which are compared one by one.
14519 The obtained average PSNR is printed through the logging system.
14521 The filter stores the accumulated MSE (mean squared error) of each
14522 frame, and at the end of the processing it is averaged across all frames
14523 equally, and the following formula is applied to obtain the PSNR:
14526 PSNR = 10*log10(MAX^2/MSE)
14529 Where MAX is the average of the maximum values of each component of the
14532 The description of the accepted parameters follows.
14535 @item stats_file, f
14536 If specified the filter will use the named file to save the PSNR of
14537 each individual frame. When filename equals "-" the data is sent to
14540 @item stats_version
14541 Specifies which version of the stats file format to use. Details of
14542 each format are written below.
14543 Default value is 1.
14545 @item stats_add_max
14546 Determines whether the max value is output to the stats log.
14547 Default value is 0.
14548 Requires stats_version >= 2. If this is set and stats_version < 2,
14549 the filter will return an error.
14552 This filter also supports the @ref{framesync} options.
14554 The file printed if @var{stats_file} is selected, contains a sequence of
14555 key/value pairs of the form @var{key}:@var{value} for each compared
14558 If a @var{stats_version} greater than 1 is specified, a header line precedes
14559 the list of per-frame-pair stats, with key value pairs following the frame
14560 format with the following parameters:
14563 @item psnr_log_version
14564 The version of the log file format. Will match @var{stats_version}.
14567 A comma separated list of the per-frame-pair parameters included in
14571 A description of each shown per-frame-pair parameter follows:
14575 sequential number of the input frame, starting from 1
14578 Mean Square Error pixel-by-pixel average difference of the compared
14579 frames, averaged over all the image components.
14581 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
14582 Mean Square Error pixel-by-pixel average difference of the compared
14583 frames for the component specified by the suffix.
14585 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
14586 Peak Signal to Noise ratio of the compared frames for the component
14587 specified by the suffix.
14589 @item max_avg, max_y, max_u, max_v
14590 Maximum allowed value for each channel, and average over all
14596 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
14597 [main][ref] psnr="stats_file=stats.log" [out]
14600 On this example the input file being processed is compared with the
14601 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
14602 is stored in @file{stats.log}.
14607 Pulldown reversal (inverse telecine) filter, capable of handling mixed
14608 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
14611 The pullup filter is designed to take advantage of future context in making
14612 its decisions. This filter is stateless in the sense that it does not lock
14613 onto a pattern to follow, but it instead looks forward to the following
14614 fields in order to identify matches and rebuild progressive frames.
14616 To produce content with an even framerate, insert the fps filter after
14617 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
14618 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
14620 The filter accepts the following options:
14627 These options set the amount of "junk" to ignore at the left, right, top, and
14628 bottom of the image, respectively. Left and right are in units of 8 pixels,
14629 while top and bottom are in units of 2 lines.
14630 The default is 8 pixels on each side.
14633 Set the strict breaks. Setting this option to 1 will reduce the chances of
14634 filter generating an occasional mismatched frame, but it may also cause an
14635 excessive number of frames to be dropped during high motion sequences.
14636 Conversely, setting it to -1 will make filter match fields more easily.
14637 This may help processing of video where there is slight blurring between
14638 the fields, but may also cause there to be interlaced frames in the output.
14639 Default value is @code{0}.
14642 Set the metric plane to use. It accepts the following values:
14648 Use chroma blue plane.
14651 Use chroma red plane.
14654 This option may be set to use chroma plane instead of the default luma plane
14655 for doing filter's computations. This may improve accuracy on very clean
14656 source material, but more likely will decrease accuracy, especially if there
14657 is chroma noise (rainbow effect) or any grayscale video.
14658 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
14659 load and make pullup usable in realtime on slow machines.
14662 For best results (without duplicated frames in the output file) it is
14663 necessary to change the output frame rate. For example, to inverse
14664 telecine NTSC input:
14666 ffmpeg -i input -vf pullup -r 24000/1001 ...
14671 Change video quantization parameters (QP).
14673 The filter accepts the following option:
14677 Set expression for quantization parameter.
14680 The expression is evaluated through the eval API and can contain, among others,
14681 the following constants:
14685 1 if index is not 129, 0 otherwise.
14688 Sequential index starting from -129 to 128.
14691 @subsection Examples
14695 Some equation like:
14703 Flush video frames from internal cache of frames into a random order.
14704 No frame is discarded.
14705 Inspired by @ref{frei0r} nervous filter.
14709 Set size in number of frames of internal cache, in range from @code{2} to
14710 @code{512}. Default is @code{30}.
14713 Set seed for random number generator, must be an integer included between
14714 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
14715 less than @code{0}, the filter will try to use a good random seed on a
14719 @section readeia608
14721 Read closed captioning (EIA-608) information from the top lines of a video frame.
14723 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
14724 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
14725 with EIA-608 data (starting from 0). A description of each metadata value follows:
14728 @item lavfi.readeia608.X.cc
14729 The two bytes stored as EIA-608 data (printed in hexadecimal).
14731 @item lavfi.readeia608.X.line
14732 The number of the line on which the EIA-608 data was identified and read.
14735 This filter accepts the following options:
14739 Set the line to start scanning for EIA-608 data. Default is @code{0}.
14742 Set the line to end scanning for EIA-608 data. Default is @code{29}.
14745 Set minimal acceptable amplitude change for sync codes detection.
14746 Default is @code{0.2}. Allowed range is @code{[0.001 - 1]}.
14749 Set the ratio of width reserved for sync code detection.
14750 Default is @code{0.27}. Allowed range is @code{[0.01 - 0.7]}.
14753 Set the max peaks height difference for sync code detection.
14754 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
14757 Set max peaks period difference for sync code detection.
14758 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
14761 Set the first two max start code bits differences.
14762 Default is @code{0.02}. Allowed range is @code{[0.0 - 0.5]}.
14765 Set the minimum ratio of bits height compared to 3rd start code bit.
14766 Default is @code{0.75}. Allowed range is @code{[0.01 - 1]}.
14769 Set the white color threshold. Default is @code{0.35}. Allowed range is @code{[0.1 - 1]}.
14772 Set the black color threshold. Default is @code{0.15}. Allowed range is @code{[0.0 - 0.5]}.
14775 Enable checking the parity bit. In the event of a parity error, the filter will output
14776 @code{0x00} for that character. Default is false.
14779 Lowpass lines prior to further processing. Default is disabled.
14782 @subsection Examples
14786 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
14788 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
14794 Read vertical interval timecode (VITC) information from the top lines of a
14797 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
14798 timecode value, if a valid timecode has been detected. Further metadata key
14799 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
14800 timecode data has been found or not.
14802 This filter accepts the following options:
14806 Set the maximum number of lines to scan for VITC data. If the value is set to
14807 @code{-1} the full video frame is scanned. Default is @code{45}.
14810 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
14811 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
14814 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
14815 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
14818 @subsection Examples
14822 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
14823 draw @code{--:--:--:--} as a placeholder:
14825 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
14831 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
14833 Destination pixel at position (X, Y) will be picked from source (x, y) position
14834 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
14835 value for pixel will be used for destination pixel.
14837 Xmap and Ymap input video streams must be of same dimensions. Output video stream
14838 will have Xmap/Ymap video stream dimensions.
14839 Xmap and Ymap input video streams are 16bit depth, single channel.
14843 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
14844 Default is @code{color}.
14847 @section removegrain
14849 The removegrain filter is a spatial denoiser for progressive video.
14853 Set mode for the first plane.
14856 Set mode for the second plane.
14859 Set mode for the third plane.
14862 Set mode for the fourth plane.
14865 Range of mode is from 0 to 24. Description of each mode follows:
14869 Leave input plane unchanged. Default.
14872 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
14875 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
14878 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
14881 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
14882 This is equivalent to a median filter.
14885 Line-sensitive clipping giving the minimal change.
14888 Line-sensitive clipping, intermediate.
14891 Line-sensitive clipping, intermediate.
14894 Line-sensitive clipping, intermediate.
14897 Line-sensitive clipping on a line where the neighbours pixels are the closest.
14900 Replaces the target pixel with the closest neighbour.
14903 [1 2 1] horizontal and vertical kernel blur.
14909 Bob mode, interpolates top field from the line where the neighbours
14910 pixels are the closest.
14913 Bob mode, interpolates bottom field from the line where the neighbours
14914 pixels are the closest.
14917 Bob mode, interpolates top field. Same as 13 but with a more complicated
14918 interpolation formula.
14921 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
14922 interpolation formula.
14925 Clips the pixel with the minimum and maximum of respectively the maximum and
14926 minimum of each pair of opposite neighbour pixels.
14929 Line-sensitive clipping using opposite neighbours whose greatest distance from
14930 the current pixel is minimal.
14933 Replaces the pixel with the average of its 8 neighbours.
14936 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
14939 Clips pixels using the averages of opposite neighbour.
14942 Same as mode 21 but simpler and faster.
14945 Small edge and halo removal, but reputed useless.
14951 @section removelogo
14953 Suppress a TV station logo, using an image file to determine which
14954 pixels comprise the logo. It works by filling in the pixels that
14955 comprise the logo with neighboring pixels.
14957 The filter accepts the following options:
14961 Set the filter bitmap file, which can be any image format supported by
14962 libavformat. The width and height of the image file must match those of the
14963 video stream being processed.
14966 Pixels in the provided bitmap image with a value of zero are not
14967 considered part of the logo, non-zero pixels are considered part of
14968 the logo. If you use white (255) for the logo and black (0) for the
14969 rest, you will be safe. For making the filter bitmap, it is
14970 recommended to take a screen capture of a black frame with the logo
14971 visible, and then using a threshold filter followed by the erode
14972 filter once or twice.
14974 If needed, little splotches can be fixed manually. Remember that if
14975 logo pixels are not covered, the filter quality will be much
14976 reduced. Marking too many pixels as part of the logo does not hurt as
14977 much, but it will increase the amount of blurring needed to cover over
14978 the image and will destroy more information than necessary, and extra
14979 pixels will slow things down on a large logo.
14981 @section repeatfields
14983 This filter uses the repeat_field flag from the Video ES headers and hard repeats
14984 fields based on its value.
14988 Reverse a video clip.
14990 Warning: This filter requires memory to buffer the entire clip, so trimming
14993 @subsection Examples
14997 Take the first 5 seconds of a clip, and reverse it.
15004 Shift R/G/B/A pixels horizontally and/or vertically.
15006 The filter accepts the following options:
15009 Set amount to shift red horizontally.
15011 Set amount to shift red vertically.
15013 Set amount to shift green horizontally.
15015 Set amount to shift green vertically.
15017 Set amount to shift blue horizontally.
15019 Set amount to shift blue vertically.
15021 Set amount to shift alpha horizontally.
15023 Set amount to shift alpha vertically.
15025 Set edge mode, can be @var{smear}, default, or @var{warp}.
15029 Apply roberts cross operator to input video stream.
15031 The filter accepts the following option:
15035 Set which planes will be processed, unprocessed planes will be copied.
15036 By default value 0xf, all planes will be processed.
15039 Set value which will be multiplied with filtered result.
15042 Set value which will be added to filtered result.
15047 Rotate video by an arbitrary angle expressed in radians.
15049 The filter accepts the following options:
15051 A description of the optional parameters follows.
15054 Set an expression for the angle by which to rotate the input video
15055 clockwise, expressed as a number of radians. A negative value will
15056 result in a counter-clockwise rotation. By default it is set to "0".
15058 This expression is evaluated for each frame.
15061 Set the output width expression, default value is "iw".
15062 This expression is evaluated just once during configuration.
15065 Set the output height expression, default value is "ih".
15066 This expression is evaluated just once during configuration.
15069 Enable bilinear interpolation if set to 1, a value of 0 disables
15070 it. Default value is 1.
15073 Set the color used to fill the output area not covered by the rotated
15074 image. For the general syntax of this option, check the
15075 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
15076 If the special value "none" is selected then no
15077 background is printed (useful for example if the background is never shown).
15079 Default value is "black".
15082 The expressions for the angle and the output size can contain the
15083 following constants and functions:
15087 sequential number of the input frame, starting from 0. It is always NAN
15088 before the first frame is filtered.
15091 time in seconds of the input frame, it is set to 0 when the filter is
15092 configured. It is always NAN before the first frame is filtered.
15096 horizontal and vertical chroma subsample values. For example for the
15097 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15101 the input video width and height
15105 the output width and height, that is the size of the padded area as
15106 specified by the @var{width} and @var{height} expressions
15110 the minimal width/height required for completely containing the input
15111 video rotated by @var{a} radians.
15113 These are only available when computing the @option{out_w} and
15114 @option{out_h} expressions.
15117 @subsection Examples
15121 Rotate the input by PI/6 radians clockwise:
15127 Rotate the input by PI/6 radians counter-clockwise:
15133 Rotate the input by 45 degrees clockwise:
15139 Apply a constant rotation with period T, starting from an angle of PI/3:
15141 rotate=PI/3+2*PI*t/T
15145 Make the input video rotation oscillating with a period of T
15146 seconds and an amplitude of A radians:
15148 rotate=A*sin(2*PI/T*t)
15152 Rotate the video, output size is chosen so that the whole rotating
15153 input video is always completely contained in the output:
15155 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
15159 Rotate the video, reduce the output size so that no background is ever
15162 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
15166 @subsection Commands
15168 The filter supports the following commands:
15172 Set the angle expression.
15173 The command accepts the same syntax of the corresponding option.
15175 If the specified expression is not valid, it is kept at its current
15181 Apply Shape Adaptive Blur.
15183 The filter accepts the following options:
15186 @item luma_radius, lr
15187 Set luma blur filter strength, must be a value in range 0.1-4.0, default
15188 value is 1.0. A greater value will result in a more blurred image, and
15189 in slower processing.
15191 @item luma_pre_filter_radius, lpfr
15192 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
15195 @item luma_strength, ls
15196 Set luma maximum difference between pixels to still be considered, must
15197 be a value in the 0.1-100.0 range, default value is 1.0.
15199 @item chroma_radius, cr
15200 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
15201 greater value will result in a more blurred image, and in slower
15204 @item chroma_pre_filter_radius, cpfr
15205 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
15207 @item chroma_strength, cs
15208 Set chroma maximum difference between pixels to still be considered,
15209 must be a value in the -0.9-100.0 range.
15212 Each chroma option value, if not explicitly specified, is set to the
15213 corresponding luma option value.
15218 Scale (resize) the input video, using the libswscale library.
15220 The scale filter forces the output display aspect ratio to be the same
15221 of the input, by changing the output sample aspect ratio.
15223 If the input image format is different from the format requested by
15224 the next filter, the scale filter will convert the input to the
15227 @subsection Options
15228 The filter accepts the following options, or any of the options
15229 supported by the libswscale scaler.
15231 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
15232 the complete list of scaler options.
15237 Set the output video dimension expression. Default value is the input
15240 If the @var{width} or @var{w} value is 0, the input width is used for
15241 the output. If the @var{height} or @var{h} value is 0, the input height
15242 is used for the output.
15244 If one and only one of the values is -n with n >= 1, the scale filter
15245 will use a value that maintains the aspect ratio of the input image,
15246 calculated from the other specified dimension. After that it will,
15247 however, make sure that the calculated dimension is divisible by n and
15248 adjust the value if necessary.
15250 If both values are -n with n >= 1, the behavior will be identical to
15251 both values being set to 0 as previously detailed.
15253 See below for the list of accepted constants for use in the dimension
15257 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
15261 Only evaluate expressions once during the filter initialization or when a command is processed.
15264 Evaluate expressions for each incoming frame.
15268 Default value is @samp{init}.
15272 Set the interlacing mode. It accepts the following values:
15276 Force interlaced aware scaling.
15279 Do not apply interlaced scaling.
15282 Select interlaced aware scaling depending on whether the source frames
15283 are flagged as interlaced or not.
15286 Default value is @samp{0}.
15289 Set libswscale scaling flags. See
15290 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15291 complete list of values. If not explicitly specified the filter applies
15295 @item param0, param1
15296 Set libswscale input parameters for scaling algorithms that need them. See
15297 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15298 complete documentation. If not explicitly specified the filter applies
15304 Set the video size. For the syntax of this option, check the
15305 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15307 @item in_color_matrix
15308 @item out_color_matrix
15309 Set in/output YCbCr color space type.
15311 This allows the autodetected value to be overridden as well as allows forcing
15312 a specific value used for the output and encoder.
15314 If not specified, the color space type depends on the pixel format.
15320 Choose automatically.
15323 Format conforming to International Telecommunication Union (ITU)
15324 Recommendation BT.709.
15327 Set color space conforming to the United States Federal Communications
15328 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
15333 Set color space conforming to:
15337 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
15340 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
15343 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
15348 Set color space conforming to SMPTE ST 240:1999.
15351 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
15356 Set in/output YCbCr sample range.
15358 This allows the autodetected value to be overridden as well as allows forcing
15359 a specific value used for the output and encoder. If not specified, the
15360 range depends on the pixel format. Possible values:
15364 Choose automatically.
15367 Set full range (0-255 in case of 8-bit luma).
15369 @item mpeg/limited/tv
15370 Set "MPEG" range (16-235 in case of 8-bit luma).
15373 @item force_original_aspect_ratio
15374 Enable decreasing or increasing output video width or height if necessary to
15375 keep the original aspect ratio. Possible values:
15379 Scale the video as specified and disable this feature.
15382 The output video dimensions will automatically be decreased if needed.
15385 The output video dimensions will automatically be increased if needed.
15389 One useful instance of this option is that when you know a specific device's
15390 maximum allowed resolution, you can use this to limit the output video to
15391 that, while retaining the aspect ratio. For example, device A allows
15392 1280x720 playback, and your video is 1920x800. Using this option (set it to
15393 decrease) and specifying 1280x720 to the command line makes the output
15396 Please note that this is a different thing than specifying -1 for @option{w}
15397 or @option{h}, you still need to specify the output resolution for this option
15400 @item force_divisible_by Ensures that the output resolution is divisible by the
15401 given integer when used together with @option{force_original_aspect_ratio}. This
15402 works similar to using -n in the @option{w} and @option{h} options.
15404 This option respects the value set for @option{force_original_aspect_ratio},
15405 increasing or decreasing the resolution accordingly. This may slightly modify
15406 the video's aspect ration.
15408 This can be handy, for example, if you want to have a video fit within a defined
15409 resolution using the @option{force_original_aspect_ratio} option but have
15410 encoder restrictions when it comes to width or height.
15414 The values of the @option{w} and @option{h} options are expressions
15415 containing the following constants:
15420 The input width and height
15424 These are the same as @var{in_w} and @var{in_h}.
15428 The output (scaled) width and height
15432 These are the same as @var{out_w} and @var{out_h}
15435 The same as @var{iw} / @var{ih}
15438 input sample aspect ratio
15441 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
15445 horizontal and vertical input chroma subsample values. For example for the
15446 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15450 horizontal and vertical output chroma subsample values. For example for the
15451 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15454 @subsection Examples
15458 Scale the input video to a size of 200x100
15463 This is equivalent to:
15474 Specify a size abbreviation for the output size:
15479 which can also be written as:
15485 Scale the input to 2x:
15487 scale=w=2*iw:h=2*ih
15491 The above is the same as:
15493 scale=2*in_w:2*in_h
15497 Scale the input to 2x with forced interlaced scaling:
15499 scale=2*iw:2*ih:interl=1
15503 Scale the input to half size:
15505 scale=w=iw/2:h=ih/2
15509 Increase the width, and set the height to the same size:
15515 Seek Greek harmony:
15522 Increase the height, and set the width to 3/2 of the height:
15524 scale=w=3/2*oh:h=3/5*ih
15528 Increase the size, making the size a multiple of the chroma
15531 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
15535 Increase the width to a maximum of 500 pixels,
15536 keeping the same aspect ratio as the input:
15538 scale=w='min(500\, iw*3/2):h=-1'
15542 Make pixels square by combining scale and setsar:
15544 scale='trunc(ih*dar):ih',setsar=1/1
15548 Make pixels square by combining scale and setsar,
15549 making sure the resulting resolution is even (required by some codecs):
15551 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
15555 @subsection Commands
15557 This filter supports the following commands:
15561 Set the output video dimension expression.
15562 The command accepts the same syntax of the corresponding option.
15564 If the specified expression is not valid, it is kept at its current
15570 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
15571 format conversion on CUDA video frames. Setting the output width and height
15572 works in the same way as for the @var{scale} filter.
15574 The following additional options are accepted:
15577 The pixel format of the output CUDA frames. If set to the string "same" (the
15578 default), the input format will be kept. Note that automatic format negotiation
15579 and conversion is not yet supported for hardware frames
15582 The interpolation algorithm used for resizing. One of the following:
15589 @item cubic2p_bspline
15590 2-parameter cubic (B=1, C=0)
15592 @item cubic2p_catmullrom
15593 2-parameter cubic (B=0, C=1/2)
15595 @item cubic2p_b05c03
15596 2-parameter cubic (B=1/2, C=3/10)
15608 Scale (resize) the input video, based on a reference video.
15610 See the scale filter for available options, scale2ref supports the same but
15611 uses the reference video instead of the main input as basis. scale2ref also
15612 supports the following additional constants for the @option{w} and
15613 @option{h} options:
15618 The main input video's width and height
15621 The same as @var{main_w} / @var{main_h}
15624 The main input video's sample aspect ratio
15626 @item main_dar, mdar
15627 The main input video's display aspect ratio. Calculated from
15628 @code{(main_w / main_h) * main_sar}.
15632 The main input video's horizontal and vertical chroma subsample values.
15633 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
15637 @subsection Examples
15641 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
15643 'scale2ref[b][a];[a][b]overlay'
15647 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
15649 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
15653 @anchor{selectivecolor}
15654 @section selectivecolor
15656 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
15657 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
15658 by the "purity" of the color (that is, how saturated it already is).
15660 This filter is similar to the Adobe Photoshop Selective Color tool.
15662 The filter accepts the following options:
15665 @item correction_method
15666 Select color correction method.
15668 Available values are:
15671 Specified adjustments are applied "as-is" (added/subtracted to original pixel
15674 Specified adjustments are relative to the original component value.
15676 Default is @code{absolute}.
15678 Adjustments for red pixels (pixels where the red component is the maximum)
15680 Adjustments for yellow pixels (pixels where the blue component is the minimum)
15682 Adjustments for green pixels (pixels where the green component is the maximum)
15684 Adjustments for cyan pixels (pixels where the red component is the minimum)
15686 Adjustments for blue pixels (pixels where the blue component is the maximum)
15688 Adjustments for magenta pixels (pixels where the green component is the minimum)
15690 Adjustments for white pixels (pixels where all components are greater than 128)
15692 Adjustments for all pixels except pure black and pure white
15694 Adjustments for black pixels (pixels where all components are lesser than 128)
15696 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
15699 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
15700 4 space separated floating point adjustment values in the [-1,1] range,
15701 respectively to adjust the amount of cyan, magenta, yellow and black for the
15702 pixels of its range.
15704 @subsection Examples
15708 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
15709 increase magenta by 27% in blue areas:
15711 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
15715 Use a Photoshop selective color preset:
15717 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
15721 @anchor{separatefields}
15722 @section separatefields
15724 The @code{separatefields} takes a frame-based video input and splits
15725 each frame into its components fields, producing a new half height clip
15726 with twice the frame rate and twice the frame count.
15728 This filter use field-dominance information in frame to decide which
15729 of each pair of fields to place first in the output.
15730 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
15732 @section setdar, setsar
15734 The @code{setdar} filter sets the Display Aspect Ratio for the filter
15737 This is done by changing the specified Sample (aka Pixel) Aspect
15738 Ratio, according to the following equation:
15740 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
15743 Keep in mind that the @code{setdar} filter does not modify the pixel
15744 dimensions of the video frame. Also, the display aspect ratio set by
15745 this filter may be changed by later filters in the filterchain,
15746 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
15749 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
15750 the filter output video.
15752 Note that as a consequence of the application of this filter, the
15753 output display aspect ratio will change according to the equation
15756 Keep in mind that the sample aspect ratio set by the @code{setsar}
15757 filter may be changed by later filters in the filterchain, e.g. if
15758 another "setsar" or a "setdar" filter is applied.
15760 It accepts the following parameters:
15763 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
15764 Set the aspect ratio used by the filter.
15766 The parameter can be a floating point number string, an expression, or
15767 a string of the form @var{num}:@var{den}, where @var{num} and
15768 @var{den} are the numerator and denominator of the aspect ratio. If
15769 the parameter is not specified, it is assumed the value "0".
15770 In case the form "@var{num}:@var{den}" is used, the @code{:} character
15774 Set the maximum integer value to use for expressing numerator and
15775 denominator when reducing the expressed aspect ratio to a rational.
15776 Default value is @code{100}.
15780 The parameter @var{sar} is an expression containing
15781 the following constants:
15785 These are approximated values for the mathematical constants e
15786 (Euler's number), pi (Greek pi), and phi (the golden ratio).
15789 The input width and height.
15792 These are the same as @var{w} / @var{h}.
15795 The input sample aspect ratio.
15798 The input display aspect ratio. It is the same as
15799 (@var{w} / @var{h}) * @var{sar}.
15802 Horizontal and vertical chroma subsample values. For example, for the
15803 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15806 @subsection Examples
15811 To change the display aspect ratio to 16:9, specify one of the following:
15818 To change the sample aspect ratio to 10:11, specify:
15824 To set a display aspect ratio of 16:9, and specify a maximum integer value of
15825 1000 in the aspect ratio reduction, use the command:
15827 setdar=ratio=16/9:max=1000
15835 Force field for the output video frame.
15837 The @code{setfield} filter marks the interlace type field for the
15838 output frames. It does not change the input frame, but only sets the
15839 corresponding property, which affects how the frame is treated by
15840 following filters (e.g. @code{fieldorder} or @code{yadif}).
15842 The filter accepts the following options:
15847 Available values are:
15851 Keep the same field property.
15854 Mark the frame as bottom-field-first.
15857 Mark the frame as top-field-first.
15860 Mark the frame as progressive.
15867 Force frame parameter for the output video frame.
15869 The @code{setparams} filter marks interlace and color range for the
15870 output frames. It does not change the input frame, but only sets the
15871 corresponding property, which affects how the frame is treated by
15876 Available values are:
15880 Keep the same field property (default).
15883 Mark the frame as bottom-field-first.
15886 Mark the frame as top-field-first.
15889 Mark the frame as progressive.
15893 Available values are:
15897 Keep the same color range property (default).
15899 @item unspecified, unknown
15900 Mark the frame as unspecified color range.
15902 @item limited, tv, mpeg
15903 Mark the frame as limited range.
15905 @item full, pc, jpeg
15906 Mark the frame as full range.
15909 @item color_primaries
15910 Set the color primaries.
15911 Available values are:
15915 Keep the same color primaries property (default).
15932 Set the color transfer.
15933 Available values are:
15937 Keep the same color trc property (default).
15959 Set the colorspace.
15960 Available values are:
15964 Keep the same colorspace property (default).
15977 @item chroma-derived-nc
15978 @item chroma-derived-c
15985 Show a line containing various information for each input video frame.
15986 The input video is not modified.
15988 This filter supports the following options:
15992 Calculate checksums of each plane. By default enabled.
15995 The shown line contains a sequence of key/value pairs of the form
15996 @var{key}:@var{value}.
15998 The following values are shown in the output:
16002 The (sequential) number of the input frame, starting from 0.
16005 The Presentation TimeStamp of the input frame, expressed as a number of
16006 time base units. The time base unit depends on the filter input pad.
16009 The Presentation TimeStamp of the input frame, expressed as a number of
16013 The position of the frame in the input stream, or -1 if this information is
16014 unavailable and/or meaningless (for example in case of synthetic video).
16017 The pixel format name.
16020 The sample aspect ratio of the input frame, expressed in the form
16021 @var{num}/@var{den}.
16024 The size of the input frame. For the syntax of this option, check the
16025 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16028 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
16029 for bottom field first).
16032 This is 1 if the frame is a key frame, 0 otherwise.
16035 The picture type of the input frame ("I" for an I-frame, "P" for a
16036 P-frame, "B" for a B-frame, or "?" for an unknown type).
16037 Also refer to the documentation of the @code{AVPictureType} enum and of
16038 the @code{av_get_picture_type_char} function defined in
16039 @file{libavutil/avutil.h}.
16042 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
16044 @item plane_checksum
16045 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
16046 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
16049 @section showpalette
16051 Displays the 256 colors palette of each frame. This filter is only relevant for
16052 @var{pal8} pixel format frames.
16054 It accepts the following option:
16058 Set the size of the box used to represent one palette color entry. Default is
16059 @code{30} (for a @code{30x30} pixel box).
16062 @section shuffleframes
16064 Reorder and/or duplicate and/or drop video frames.
16066 It accepts the following parameters:
16070 Set the destination indexes of input frames.
16071 This is space or '|' separated list of indexes that maps input frames to output
16072 frames. Number of indexes also sets maximal value that each index may have.
16073 '-1' index have special meaning and that is to drop frame.
16076 The first frame has the index 0. The default is to keep the input unchanged.
16078 @subsection Examples
16082 Swap second and third frame of every three frames of the input:
16084 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
16088 Swap 10th and 1st frame of every ten frames of the input:
16090 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
16094 @section shuffleplanes
16096 Reorder and/or duplicate video planes.
16098 It accepts the following parameters:
16103 The index of the input plane to be used as the first output plane.
16106 The index of the input plane to be used as the second output plane.
16109 The index of the input plane to be used as the third output plane.
16112 The index of the input plane to be used as the fourth output plane.
16116 The first plane has the index 0. The default is to keep the input unchanged.
16118 @subsection Examples
16122 Swap the second and third planes of the input:
16124 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
16128 @anchor{signalstats}
16129 @section signalstats
16130 Evaluate various visual metrics that assist in determining issues associated
16131 with the digitization of analog video media.
16133 By default the filter will log these metadata values:
16137 Display the minimal Y value contained within the input frame. Expressed in
16141 Display the Y value at the 10% percentile within the input frame. Expressed in
16145 Display the average Y value within the input frame. Expressed in range of
16149 Display the Y value at the 90% percentile within the input frame. Expressed in
16153 Display the maximum Y value contained within the input frame. Expressed in
16157 Display the minimal U value contained within the input frame. Expressed in
16161 Display the U value at the 10% percentile within the input frame. Expressed in
16165 Display the average U value within the input frame. Expressed in range of
16169 Display the U value at the 90% percentile within the input frame. Expressed in
16173 Display the maximum U value contained within the input frame. Expressed in
16177 Display the minimal V value contained within the input frame. Expressed in
16181 Display the V value at the 10% percentile within the input frame. Expressed in
16185 Display the average V value within the input frame. Expressed in range of
16189 Display the V value at the 90% percentile within the input frame. Expressed in
16193 Display the maximum V value contained within the input frame. Expressed in
16197 Display the minimal saturation value contained within the input frame.
16198 Expressed in range of [0-~181.02].
16201 Display the saturation value at the 10% percentile within the input frame.
16202 Expressed in range of [0-~181.02].
16205 Display the average saturation value within the input frame. Expressed in range
16209 Display the saturation value at the 90% percentile within the input frame.
16210 Expressed in range of [0-~181.02].
16213 Display the maximum saturation value contained within the input frame.
16214 Expressed in range of [0-~181.02].
16217 Display the median value for hue within the input frame. Expressed in range of
16221 Display the average value for hue within the input frame. Expressed in range of
16225 Display the average of sample value difference between all values of the Y
16226 plane in the current frame and corresponding values of the previous input frame.
16227 Expressed in range of [0-255].
16230 Display the average of sample value difference between all values of the U
16231 plane in the current frame and corresponding values of the previous input frame.
16232 Expressed in range of [0-255].
16235 Display the average of sample value difference between all values of the V
16236 plane in the current frame and corresponding values of the previous input frame.
16237 Expressed in range of [0-255].
16240 Display bit depth of Y plane in current frame.
16241 Expressed in range of [0-16].
16244 Display bit depth of U plane in current frame.
16245 Expressed in range of [0-16].
16248 Display bit depth of V plane in current frame.
16249 Expressed in range of [0-16].
16252 The filter accepts the following options:
16258 @option{stat} specify an additional form of image analysis.
16259 @option{out} output video with the specified type of pixel highlighted.
16261 Both options accept the following values:
16265 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
16266 unlike the neighboring pixels of the same field. Examples of temporal outliers
16267 include the results of video dropouts, head clogs, or tape tracking issues.
16270 Identify @var{vertical line repetition}. Vertical line repetition includes
16271 similar rows of pixels within a frame. In born-digital video vertical line
16272 repetition is common, but this pattern is uncommon in video digitized from an
16273 analog source. When it occurs in video that results from the digitization of an
16274 analog source it can indicate concealment from a dropout compensator.
16277 Identify pixels that fall outside of legal broadcast range.
16281 Set the highlight color for the @option{out} option. The default color is
16285 @subsection Examples
16289 Output data of various video metrics:
16291 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
16295 Output specific data about the minimum and maximum values of the Y plane per frame:
16297 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
16301 Playback video while highlighting pixels that are outside of broadcast range in red.
16303 ffplay example.mov -vf signalstats="out=brng:color=red"
16307 Playback video with signalstats metadata drawn over the frame.
16309 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
16312 The contents of signalstat_drawtext.txt used in the command are:
16315 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
16316 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
16317 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
16318 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
16326 Calculates the MPEG-7 Video Signature. The filter can handle more than one
16327 input. In this case the matching between the inputs can be calculated additionally.
16328 The filter always passes through the first input. The signature of each stream can
16329 be written into a file.
16331 It accepts the following options:
16335 Enable or disable the matching process.
16337 Available values are:
16341 Disable the calculation of a matching (default).
16343 Calculate the matching for the whole video and output whether the whole video
16344 matches or only parts.
16346 Calculate only until a matching is found or the video ends. Should be faster in
16351 Set the number of inputs. The option value must be a non negative integer.
16352 Default value is 1.
16355 Set the path to which the output is written. If there is more than one input,
16356 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
16357 integer), that will be replaced with the input number. If no filename is
16358 specified, no output will be written. This is the default.
16361 Choose the output format.
16363 Available values are:
16367 Use the specified binary representation (default).
16369 Use the specified xml representation.
16373 Set threshold to detect one word as similar. The option value must be an integer
16374 greater than zero. The default value is 9000.
16377 Set threshold to detect all words as similar. The option value must be an integer
16378 greater than zero. The default value is 60000.
16381 Set threshold to detect frames as similar. The option value must be an integer
16382 greater than zero. The default value is 116.
16385 Set the minimum length of a sequence in frames to recognize it as matching
16386 sequence. The option value must be a non negative integer value.
16387 The default value is 0.
16390 Set the minimum relation, that matching frames to all frames must have.
16391 The option value must be a double value between 0 and 1. The default value is 0.5.
16394 @subsection Examples
16398 To calculate the signature of an input video and store it in signature.bin:
16400 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
16404 To detect whether two videos match and store the signatures in XML format in
16405 signature0.xml and signature1.xml:
16407 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 -
16415 Blur the input video without impacting the outlines.
16417 It accepts the following options:
16420 @item luma_radius, lr
16421 Set the luma radius. The option value must be a float number in
16422 the range [0.1,5.0] that specifies the variance of the gaussian filter
16423 used to blur the image (slower if larger). Default value is 1.0.
16425 @item luma_strength, ls
16426 Set the luma strength. The option value must be a float number
16427 in the range [-1.0,1.0] that configures the blurring. A value included
16428 in [0.0,1.0] will blur the image whereas a value included in
16429 [-1.0,0.0] will sharpen the image. Default value is 1.0.
16431 @item luma_threshold, lt
16432 Set the luma threshold used as a coefficient to determine
16433 whether a pixel should be blurred or not. The option value must be an
16434 integer in the range [-30,30]. A value of 0 will filter all the image,
16435 a value included in [0,30] will filter flat areas and a value included
16436 in [-30,0] will filter edges. Default value is 0.
16438 @item chroma_radius, cr
16439 Set the chroma radius. The option value must be a float number in
16440 the range [0.1,5.0] that specifies the variance of the gaussian filter
16441 used to blur the image (slower if larger). Default value is @option{luma_radius}.
16443 @item chroma_strength, cs
16444 Set the chroma strength. The option value must be a float number
16445 in the range [-1.0,1.0] that configures the blurring. A value included
16446 in [0.0,1.0] will blur the image whereas a value included in
16447 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
16449 @item chroma_threshold, ct
16450 Set the chroma threshold used as a coefficient to determine
16451 whether a pixel should be blurred or not. The option value must be an
16452 integer in the range [-30,30]. A value of 0 will filter all the image,
16453 a value included in [0,30] will filter flat areas and a value included
16454 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
16457 If a chroma option is not explicitly set, the corresponding luma value
16461 Apply sobel operator to input video stream.
16463 The filter accepts the following option:
16467 Set which planes will be processed, unprocessed planes will be copied.
16468 By default value 0xf, all planes will be processed.
16471 Set value which will be multiplied with filtered result.
16474 Set value which will be added to filtered result.
16480 Apply a simple postprocessing filter that compresses and decompresses the image
16481 at several (or - in the case of @option{quality} level @code{6} - all) shifts
16482 and average the results.
16484 The filter accepts the following options:
16488 Set quality. This option defines the number of levels for averaging. It accepts
16489 an integer in the range 0-6. If set to @code{0}, the filter will have no
16490 effect. A value of @code{6} means the higher quality. For each increment of
16491 that value the speed drops by a factor of approximately 2. Default value is
16495 Force a constant quantization parameter. If not set, the filter will use the QP
16496 from the video stream (if available).
16499 Set thresholding mode. Available modes are:
16503 Set hard thresholding (default).
16505 Set soft thresholding (better de-ringing effect, but likely blurrier).
16508 @item use_bframe_qp
16509 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
16510 option may cause flicker since the B-Frames have often larger QP. Default is
16511 @code{0} (not enabled).
16516 Scale the input by applying one of the super-resolution methods based on
16517 convolutional neural networks. Supported models:
16521 Super-Resolution Convolutional Neural Network model (SRCNN).
16522 See @url{https://arxiv.org/abs/1501.00092}.
16525 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
16526 See @url{https://arxiv.org/abs/1609.05158}.
16529 Training scripts as well as scripts for model file (.pb) saving can be found at
16530 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
16531 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
16533 Native model files (.model) can be generated from TensorFlow model
16534 files (.pb) by using tools/python/convert.py
16536 The filter accepts the following options:
16540 Specify which DNN backend to use for model loading and execution. This option accepts
16541 the following values:
16545 Native implementation of DNN loading and execution.
16548 TensorFlow backend. To enable this backend you
16549 need to install the TensorFlow for C library (see
16550 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
16551 @code{--enable-libtensorflow}
16554 Default value is @samp{native}.
16557 Set path to model file specifying network architecture and its parameters.
16558 Note that different backends use different file formats. TensorFlow backend
16559 can load files for both formats, while native backend can load files for only
16563 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
16564 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
16565 input upscaled using bicubic upscaling with proper scale factor.
16570 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
16572 This filter takes in input two input videos, the first input is
16573 considered the "main" source and is passed unchanged to the
16574 output. The second input is used as a "reference" video for computing
16577 Both video inputs must have the same resolution and pixel format for
16578 this filter to work correctly. Also it assumes that both inputs
16579 have the same number of frames, which are compared one by one.
16581 The filter stores the calculated SSIM of each frame.
16583 The description of the accepted parameters follows.
16586 @item stats_file, f
16587 If specified the filter will use the named file to save the SSIM of
16588 each individual frame. When filename equals "-" the data is sent to
16592 The file printed if @var{stats_file} is selected, contains a sequence of
16593 key/value pairs of the form @var{key}:@var{value} for each compared
16596 A description of each shown parameter follows:
16600 sequential number of the input frame, starting from 1
16602 @item Y, U, V, R, G, B
16603 SSIM of the compared frames for the component specified by the suffix.
16606 SSIM of the compared frames for the whole frame.
16609 Same as above but in dB representation.
16612 This filter also supports the @ref{framesync} options.
16616 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16617 [main][ref] ssim="stats_file=stats.log" [out]
16620 On this example the input file being processed is compared with the
16621 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
16622 is stored in @file{stats.log}.
16624 Another example with both psnr and ssim at same time:
16626 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
16631 Convert between different stereoscopic image formats.
16633 The filters accept the following options:
16637 Set stereoscopic image format of input.
16639 Available values for input image formats are:
16642 side by side parallel (left eye left, right eye right)
16645 side by side crosseye (right eye left, left eye right)
16648 side by side parallel with half width resolution
16649 (left eye left, right eye right)
16652 side by side crosseye with half width resolution
16653 (right eye left, left eye right)
16657 above-below (left eye above, right eye below)
16661 above-below (right eye above, left eye below)
16665 above-below with half height resolution
16666 (left eye above, right eye below)
16670 above-below with half height resolution
16671 (right eye above, left eye below)
16674 alternating frames (left eye first, right eye second)
16677 alternating frames (right eye first, left eye second)
16680 interleaved rows (left eye has top row, right eye starts on next row)
16683 interleaved rows (right eye has top row, left eye starts on next row)
16686 interleaved columns, left eye first
16689 interleaved columns, right eye first
16691 Default value is @samp{sbsl}.
16695 Set stereoscopic image format of output.
16699 side by side parallel (left eye left, right eye right)
16702 side by side crosseye (right eye left, left eye right)
16705 side by side parallel with half width resolution
16706 (left eye left, right eye right)
16709 side by side crosseye with half width resolution
16710 (right eye left, left eye right)
16714 above-below (left eye above, right eye below)
16718 above-below (right eye above, left eye below)
16722 above-below with half height resolution
16723 (left eye above, right eye below)
16727 above-below with half height resolution
16728 (right eye above, left eye below)
16731 alternating frames (left eye first, right eye second)
16734 alternating frames (right eye first, left eye second)
16737 interleaved rows (left eye has top row, right eye starts on next row)
16740 interleaved rows (right eye has top row, left eye starts on next row)
16743 anaglyph red/blue gray
16744 (red filter on left eye, blue filter on right eye)
16747 anaglyph red/green gray
16748 (red filter on left eye, green filter on right eye)
16751 anaglyph red/cyan gray
16752 (red filter on left eye, cyan filter on right eye)
16755 anaglyph red/cyan half colored
16756 (red filter on left eye, cyan filter on right eye)
16759 anaglyph red/cyan color
16760 (red filter on left eye, cyan filter on right eye)
16763 anaglyph red/cyan color optimized with the least squares projection of dubois
16764 (red filter on left eye, cyan filter on right eye)
16767 anaglyph green/magenta gray
16768 (green filter on left eye, magenta filter on right eye)
16771 anaglyph green/magenta half colored
16772 (green filter on left eye, magenta filter on right eye)
16775 anaglyph green/magenta colored
16776 (green filter on left eye, magenta filter on right eye)
16779 anaglyph green/magenta color optimized with the least squares projection of dubois
16780 (green filter on left eye, magenta filter on right eye)
16783 anaglyph yellow/blue gray
16784 (yellow filter on left eye, blue filter on right eye)
16787 anaglyph yellow/blue half colored
16788 (yellow filter on left eye, blue filter on right eye)
16791 anaglyph yellow/blue colored
16792 (yellow filter on left eye, blue filter on right eye)
16795 anaglyph yellow/blue color optimized with the least squares projection of dubois
16796 (yellow filter on left eye, blue filter on right eye)
16799 mono output (left eye only)
16802 mono output (right eye only)
16805 checkerboard, left eye first
16808 checkerboard, right eye first
16811 interleaved columns, left eye first
16814 interleaved columns, right eye first
16820 Default value is @samp{arcd}.
16823 @subsection Examples
16827 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
16833 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
16839 @section streamselect, astreamselect
16840 Select video or audio streams.
16842 The filter accepts the following options:
16846 Set number of inputs. Default is 2.
16849 Set input indexes to remap to outputs.
16852 @subsection Commands
16854 The @code{streamselect} and @code{astreamselect} filter supports the following
16859 Set input indexes to remap to outputs.
16862 @subsection Examples
16866 Select first 5 seconds 1st stream and rest of time 2nd stream:
16868 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
16872 Same as above, but for audio:
16874 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
16881 Draw subtitles on top of input video using the libass library.
16883 To enable compilation of this filter you need to configure FFmpeg with
16884 @code{--enable-libass}. This filter also requires a build with libavcodec and
16885 libavformat to convert the passed subtitles file to ASS (Advanced Substation
16886 Alpha) subtitles format.
16888 The filter accepts the following options:
16892 Set the filename of the subtitle file to read. It must be specified.
16894 @item original_size
16895 Specify the size of the original video, the video for which the ASS file
16896 was composed. For the syntax of this option, check the
16897 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16898 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
16899 correctly scale the fonts if the aspect ratio has been changed.
16902 Set a directory path containing fonts that can be used by the filter.
16903 These fonts will be used in addition to whatever the font provider uses.
16906 Process alpha channel, by default alpha channel is untouched.
16909 Set subtitles input character encoding. @code{subtitles} filter only. Only
16910 useful if not UTF-8.
16912 @item stream_index, si
16913 Set subtitles stream index. @code{subtitles} filter only.
16916 Override default style or script info parameters of the subtitles. It accepts a
16917 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
16920 If the first key is not specified, it is assumed that the first value
16921 specifies the @option{filename}.
16923 For example, to render the file @file{sub.srt} on top of the input
16924 video, use the command:
16929 which is equivalent to:
16931 subtitles=filename=sub.srt
16934 To render the default subtitles stream from file @file{video.mkv}, use:
16936 subtitles=video.mkv
16939 To render the second subtitles stream from that file, use:
16941 subtitles=video.mkv:si=1
16944 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
16945 @code{DejaVu Serif}, use:
16947 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
16950 @section super2xsai
16952 Scale the input by 2x and smooth using the Super2xSaI (Scale and
16953 Interpolate) pixel art scaling algorithm.
16955 Useful for enlarging pixel art images without reducing sharpness.
16959 Swap two rectangular objects in video.
16961 This filter accepts the following options:
16971 Set 1st rect x coordinate.
16974 Set 1st rect y coordinate.
16977 Set 2nd rect x coordinate.
16980 Set 2nd rect y coordinate.
16982 All expressions are evaluated once for each frame.
16985 The all options are expressions containing the following constants:
16990 The input width and height.
16993 same as @var{w} / @var{h}
16996 input sample aspect ratio
16999 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
17002 The number of the input frame, starting from 0.
17005 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
17008 the position in the file of the input frame, NAN if unknown
17016 Apply telecine process to the video.
17018 This filter accepts the following options:
17027 The default value is @code{top}.
17031 A string of numbers representing the pulldown pattern you wish to apply.
17032 The default value is @code{23}.
17036 Some typical patterns:
17041 24p: 2332 (preferred)
17048 24p: 222222222223 ("Euro pulldown")
17055 Apply threshold effect to video stream.
17057 This filter needs four video streams to perform thresholding.
17058 First stream is stream we are filtering.
17059 Second stream is holding threshold values, third stream is holding min values,
17060 and last, fourth stream is holding max values.
17062 The filter accepts the following option:
17066 Set which planes will be processed, unprocessed planes will be copied.
17067 By default value 0xf, all planes will be processed.
17070 For example if first stream pixel's component value is less then threshold value
17071 of pixel component from 2nd threshold stream, third stream value will picked,
17072 otherwise fourth stream pixel component value will be picked.
17074 Using color source filter one can perform various types of thresholding:
17076 @subsection Examples
17080 Binary threshold, using gray color as threshold:
17082 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
17086 Inverted binary threshold, using gray color as threshold:
17088 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
17092 Truncate binary threshold, using gray color as threshold:
17094 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
17098 Threshold to zero, using gray color as threshold:
17100 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
17104 Inverted threshold to zero, using gray color as threshold:
17106 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
17111 Select the most representative frame in a given sequence of consecutive frames.
17113 The filter accepts the following options:
17117 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
17118 will pick one of them, and then handle the next batch of @var{n} frames until
17119 the end. Default is @code{100}.
17122 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
17123 value will result in a higher memory usage, so a high value is not recommended.
17125 @subsection Examples
17129 Extract one picture each 50 frames:
17135 Complete example of a thumbnail creation with @command{ffmpeg}:
17137 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
17143 Tile several successive frames together.
17145 The filter accepts the following options:
17150 Set the grid size (i.e. the number of lines and columns). For the syntax of
17151 this option, check the
17152 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17155 Set the maximum number of frames to render in the given area. It must be less
17156 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
17157 the area will be used.
17160 Set the outer border margin in pixels.
17163 Set the inner border thickness (i.e. the number of pixels between frames). For
17164 more advanced padding options (such as having different values for the edges),
17165 refer to the pad video filter.
17168 Specify the color of the unused area. For the syntax of this option, check the
17169 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
17170 The default value of @var{color} is "black".
17173 Set the number of frames to overlap when tiling several successive frames together.
17174 The value must be between @code{0} and @var{nb_frames - 1}.
17177 Set the number of frames to initially be empty before displaying first output frame.
17178 This controls how soon will one get first output frame.
17179 The value must be between @code{0} and @var{nb_frames - 1}.
17182 @subsection Examples
17186 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
17188 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
17190 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
17191 duplicating each output frame to accommodate the originally detected frame
17195 Display @code{5} pictures in an area of @code{3x2} frames,
17196 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
17197 mixed flat and named options:
17199 tile=3x2:nb_frames=5:padding=7:margin=2
17203 @section tinterlace
17205 Perform various types of temporal field interlacing.
17207 Frames are counted starting from 1, so the first input frame is
17210 The filter accepts the following options:
17215 Specify the mode of the interlacing. This option can also be specified
17216 as a value alone. See below for a list of values for this option.
17218 Available values are:
17222 Move odd frames into the upper field, even into the lower field,
17223 generating a double height frame at half frame rate.
17227 Frame 1 Frame 2 Frame 3 Frame 4
17229 11111 22222 33333 44444
17230 11111 22222 33333 44444
17231 11111 22222 33333 44444
17232 11111 22222 33333 44444
17246 Only output odd frames, even frames are dropped, generating a frame with
17247 unchanged height at half frame rate.
17252 Frame 1 Frame 2 Frame 3 Frame 4
17254 11111 22222 33333 44444
17255 11111 22222 33333 44444
17256 11111 22222 33333 44444
17257 11111 22222 33333 44444
17267 Only output even frames, odd frames are dropped, generating a frame with
17268 unchanged height at half frame rate.
17273 Frame 1 Frame 2 Frame 3 Frame 4
17275 11111 22222 33333 44444
17276 11111 22222 33333 44444
17277 11111 22222 33333 44444
17278 11111 22222 33333 44444
17288 Expand each frame to full height, but pad alternate lines with black,
17289 generating a frame with double height at the same input frame rate.
17294 Frame 1 Frame 2 Frame 3 Frame 4
17296 11111 22222 33333 44444
17297 11111 22222 33333 44444
17298 11111 22222 33333 44444
17299 11111 22222 33333 44444
17302 11111 ..... 33333 .....
17303 ..... 22222 ..... 44444
17304 11111 ..... 33333 .....
17305 ..... 22222 ..... 44444
17306 11111 ..... 33333 .....
17307 ..... 22222 ..... 44444
17308 11111 ..... 33333 .....
17309 ..... 22222 ..... 44444
17313 @item interleave_top, 4
17314 Interleave the upper field from odd frames with the lower field from
17315 even frames, generating a frame with unchanged height at half frame rate.
17320 Frame 1 Frame 2 Frame 3 Frame 4
17322 11111<- 22222 33333<- 44444
17323 11111 22222<- 33333 44444<-
17324 11111<- 22222 33333<- 44444
17325 11111 22222<- 33333 44444<-
17335 @item interleave_bottom, 5
17336 Interleave the lower field from odd frames with the upper field from
17337 even frames, generating a frame with unchanged height at half frame rate.
17342 Frame 1 Frame 2 Frame 3 Frame 4
17344 11111 22222<- 33333 44444<-
17345 11111<- 22222 33333<- 44444
17346 11111 22222<- 33333 44444<-
17347 11111<- 22222 33333<- 44444
17357 @item interlacex2, 6
17358 Double frame rate with unchanged height. Frames are inserted each
17359 containing the second temporal field from the previous input frame and
17360 the first temporal field from the next input frame. This mode relies on
17361 the top_field_first flag. Useful for interlaced video displays with no
17362 field synchronisation.
17367 Frame 1 Frame 2 Frame 3 Frame 4
17369 11111 22222 33333 44444
17370 11111 22222 33333 44444
17371 11111 22222 33333 44444
17372 11111 22222 33333 44444
17375 11111 22222 22222 33333 33333 44444 44444
17376 11111 11111 22222 22222 33333 33333 44444
17377 11111 22222 22222 33333 33333 44444 44444
17378 11111 11111 22222 22222 33333 33333 44444
17383 Move odd frames into the upper field, even into the lower field,
17384 generating a double height frame at same frame rate.
17389 Frame 1 Frame 2 Frame 3 Frame 4
17391 11111 22222 33333 44444
17392 11111 22222 33333 44444
17393 11111 22222 33333 44444
17394 11111 22222 33333 44444
17397 11111 33333 33333 55555
17398 22222 22222 44444 44444
17399 11111 33333 33333 55555
17400 22222 22222 44444 44444
17401 11111 33333 33333 55555
17402 22222 22222 44444 44444
17403 11111 33333 33333 55555
17404 22222 22222 44444 44444
17409 Numeric values are deprecated but are accepted for backward
17410 compatibility reasons.
17412 Default mode is @code{merge}.
17415 Specify flags influencing the filter process.
17417 Available value for @var{flags} is:
17420 @item low_pass_filter, vlpf
17421 Enable linear vertical low-pass filtering in the filter.
17422 Vertical low-pass filtering is required when creating an interlaced
17423 destination from a progressive source which contains high-frequency
17424 vertical detail. Filtering will reduce interlace 'twitter' and Moire
17427 @item complex_filter, cvlpf
17428 Enable complex vertical low-pass filtering.
17429 This will slightly less reduce interlace 'twitter' and Moire
17430 patterning but better retain detail and subjective sharpness impression.
17434 Vertical low-pass filtering can only be enabled for @option{mode}
17435 @var{interleave_top} and @var{interleave_bottom}.
17441 Mix successive video frames.
17443 A description of the accepted options follows.
17447 The number of successive frames to mix. If unspecified, it defaults to 3.
17450 Specify weight of each input video frame.
17451 Each weight is separated by space. If number of weights is smaller than
17452 number of @var{frames} last specified weight will be used for all remaining
17456 Specify scale, if it is set it will be multiplied with sum
17457 of each weight multiplied with pixel values to give final destination
17458 pixel value. By default @var{scale} is auto scaled to sum of weights.
17461 @subsection Examples
17465 Average 7 successive frames:
17467 tmix=frames=7:weights="1 1 1 1 1 1 1"
17471 Apply simple temporal convolution:
17473 tmix=frames=3:weights="-1 3 -1"
17477 Similar as above but only showing temporal differences:
17479 tmix=frames=3:weights="-1 2 -1":scale=1
17485 Tone map colors from different dynamic ranges.
17487 This filter expects data in single precision floating point, as it needs to
17488 operate on (and can output) out-of-range values. Another filter, such as
17489 @ref{zscale}, is needed to convert the resulting frame to a usable format.
17491 The tonemapping algorithms implemented only work on linear light, so input
17492 data should be linearized beforehand (and possibly correctly tagged).
17495 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
17498 @subsection Options
17499 The filter accepts the following options.
17503 Set the tone map algorithm to use.
17505 Possible values are:
17508 Do not apply any tone map, only desaturate overbright pixels.
17511 Hard-clip any out-of-range values. Use it for perfect color accuracy for
17512 in-range values, while distorting out-of-range values.
17515 Stretch the entire reference gamut to a linear multiple of the display.
17518 Fit a logarithmic transfer between the tone curves.
17521 Preserve overall image brightness with a simple curve, using nonlinear
17522 contrast, which results in flattening details and degrading color accuracy.
17525 Preserve both dark and bright details better than @var{reinhard}, at the cost
17526 of slightly darkening everything. Use it when detail preservation is more
17527 important than color and brightness accuracy.
17530 Smoothly map out-of-range values, while retaining contrast and colors for
17531 in-range material as much as possible. Use it when color accuracy is more
17532 important than detail preservation.
17538 Tune the tone mapping algorithm.
17540 This affects the following algorithms:
17546 Specifies the scale factor to use while stretching.
17550 Specifies the exponent of the function.
17554 Specify an extra linear coefficient to multiply into the signal before clipping.
17558 Specify the local contrast coefficient at the display peak.
17559 Default to 0.5, which means that in-gamut values will be about half as bright
17566 Specify the transition point from linear to mobius transform. Every value
17567 below this point is guaranteed to be mapped 1:1. The higher the value, the
17568 more accurate the result will be, at the cost of losing bright details.
17569 Default to 0.3, which due to the steep initial slope still preserves in-range
17570 colors fairly accurately.
17574 Apply desaturation for highlights that exceed this level of brightness. The
17575 higher the parameter, the more color information will be preserved. This
17576 setting helps prevent unnaturally blown-out colors for super-highlights, by
17577 (smoothly) turning into white instead. This makes images feel more natural,
17578 at the cost of reducing information about out-of-range colors.
17580 The default of 2.0 is somewhat conservative and will mostly just apply to
17581 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
17583 This option works only if the input frame has a supported color tag.
17586 Override signal/nominal/reference peak with this value. Useful when the
17587 embedded peak information in display metadata is not reliable or when tone
17588 mapping from a lower range to a higher range.
17593 Temporarily pad video frames.
17595 The filter accepts the following options:
17599 Specify number of delay frames before input video stream.
17602 Specify number of padding frames after input video stream.
17603 Set to -1 to pad indefinitely.
17606 Set kind of frames added to beginning of stream.
17607 Can be either @var{add} or @var{clone}.
17608 With @var{add} frames of solid-color are added.
17609 With @var{clone} frames are clones of first frame.
17612 Set kind of frames added to end of stream.
17613 Can be either @var{add} or @var{clone}.
17614 With @var{add} frames of solid-color are added.
17615 With @var{clone} frames are clones of last frame.
17617 @item start_duration, stop_duration
17618 Specify the duration of the start/stop delay. See
17619 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17620 for the accepted syntax.
17621 These options override @var{start} and @var{stop}.
17624 Specify the color of the padded area. For the syntax of this option,
17625 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
17626 manual,ffmpeg-utils}.
17628 The default value of @var{color} is "black".
17634 Transpose rows with columns in the input video and optionally flip it.
17636 It accepts the following parameters:
17641 Specify the transposition direction.
17643 Can assume the following values:
17645 @item 0, 4, cclock_flip
17646 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
17654 Rotate by 90 degrees clockwise, that is:
17662 Rotate by 90 degrees counterclockwise, that is:
17669 @item 3, 7, clock_flip
17670 Rotate by 90 degrees clockwise and vertically flip, that is:
17678 For values between 4-7, the transposition is only done if the input
17679 video geometry is portrait and not landscape. These values are
17680 deprecated, the @code{passthrough} option should be used instead.
17682 Numerical values are deprecated, and should be dropped in favor of
17683 symbolic constants.
17686 Do not apply the transposition if the input geometry matches the one
17687 specified by the specified value. It accepts the following values:
17690 Always apply transposition.
17692 Preserve portrait geometry (when @var{height} >= @var{width}).
17694 Preserve landscape geometry (when @var{width} >= @var{height}).
17697 Default value is @code{none}.
17700 For example to rotate by 90 degrees clockwise and preserve portrait
17703 transpose=dir=1:passthrough=portrait
17706 The command above can also be specified as:
17708 transpose=1:portrait
17711 @section transpose_npp
17713 Transpose rows with columns in the input video and optionally flip it.
17714 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
17716 It accepts the following parameters:
17721 Specify the transposition direction.
17723 Can assume the following values:
17726 Rotate by 90 degrees counterclockwise and vertically flip. (default)
17729 Rotate by 90 degrees clockwise.
17732 Rotate by 90 degrees counterclockwise.
17735 Rotate by 90 degrees clockwise and vertically flip.
17739 Do not apply the transposition if the input geometry matches the one
17740 specified by the specified value. It accepts the following values:
17743 Always apply transposition. (default)
17745 Preserve portrait geometry (when @var{height} >= @var{width}).
17747 Preserve landscape geometry (when @var{width} >= @var{height}).
17753 Trim the input so that the output contains one continuous subpart of the input.
17755 It accepts the following parameters:
17758 Specify the time of the start of the kept section, i.e. the frame with the
17759 timestamp @var{start} will be the first frame in the output.
17762 Specify the time of the first frame that will be dropped, i.e. the frame
17763 immediately preceding the one with the timestamp @var{end} will be the last
17764 frame in the output.
17767 This is the same as @var{start}, except this option sets the start timestamp
17768 in timebase units instead of seconds.
17771 This is the same as @var{end}, except this option sets the end timestamp
17772 in timebase units instead of seconds.
17775 The maximum duration of the output in seconds.
17778 The number of the first frame that should be passed to the output.
17781 The number of the first frame that should be dropped.
17784 @option{start}, @option{end}, and @option{duration} are expressed as time
17785 duration specifications; see
17786 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17787 for the accepted syntax.
17789 Note that the first two sets of the start/end options and the @option{duration}
17790 option look at the frame timestamp, while the _frame variants simply count the
17791 frames that pass through the filter. Also note that this filter does not modify
17792 the timestamps. If you wish for the output timestamps to start at zero, insert a
17793 setpts filter after the trim filter.
17795 If multiple start or end options are set, this filter tries to be greedy and
17796 keep all the frames that match at least one of the specified constraints. To keep
17797 only the part that matches all the constraints at once, chain multiple trim
17800 The defaults are such that all the input is kept. So it is possible to set e.g.
17801 just the end values to keep everything before the specified time.
17806 Drop everything except the second minute of input:
17808 ffmpeg -i INPUT -vf trim=60:120
17812 Keep only the first second:
17814 ffmpeg -i INPUT -vf trim=duration=1
17819 @section unpremultiply
17820 Apply alpha unpremultiply effect to input video stream using first plane
17821 of second stream as alpha.
17823 Both streams must have same dimensions and same pixel format.
17825 The filter accepts the following option:
17829 Set which planes will be processed, unprocessed planes will be copied.
17830 By default value 0xf, all planes will be processed.
17832 If the format has 1 or 2 components, then luma is bit 0.
17833 If the format has 3 or 4 components:
17834 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
17835 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
17836 If present, the alpha channel is always the last bit.
17839 Do not require 2nd input for processing, instead use alpha plane from input stream.
17845 Sharpen or blur the input video.
17847 It accepts the following parameters:
17850 @item luma_msize_x, lx
17851 Set the luma matrix horizontal size. It must be an odd integer between
17852 3 and 23. The default value is 5.
17854 @item luma_msize_y, ly
17855 Set the luma matrix vertical size. It must be an odd integer between 3
17856 and 23. The default value is 5.
17858 @item luma_amount, la
17859 Set the luma effect strength. It must be a floating point number, reasonable
17860 values lay between -1.5 and 1.5.
17862 Negative values will blur the input video, while positive values will
17863 sharpen it, a value of zero will disable the effect.
17865 Default value is 1.0.
17867 @item chroma_msize_x, cx
17868 Set the chroma matrix horizontal size. It must be an odd integer
17869 between 3 and 23. The default value is 5.
17871 @item chroma_msize_y, cy
17872 Set the chroma matrix vertical size. It must be an odd integer
17873 between 3 and 23. The default value is 5.
17875 @item chroma_amount, ca
17876 Set the chroma effect strength. It must be a floating point number, reasonable
17877 values lay between -1.5 and 1.5.
17879 Negative values will blur the input video, while positive values will
17880 sharpen it, a value of zero will disable the effect.
17882 Default value is 0.0.
17886 All parameters are optional and default to the equivalent of the
17887 string '5:5:1.0:5:5:0.0'.
17889 @subsection Examples
17893 Apply strong luma sharpen effect:
17895 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
17899 Apply a strong blur of both luma and chroma parameters:
17901 unsharp=7:7:-2:7:7:-2
17907 Apply ultra slow/simple postprocessing filter that compresses and decompresses
17908 the image at several (or - in the case of @option{quality} level @code{8} - all)
17909 shifts and average the results.
17911 The way this differs from the behavior of spp is that uspp actually encodes &
17912 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
17913 DCT similar to MJPEG.
17915 The filter accepts the following options:
17919 Set quality. This option defines the number of levels for averaging. It accepts
17920 an integer in the range 0-8. If set to @code{0}, the filter will have no
17921 effect. A value of @code{8} means the higher quality. For each increment of
17922 that value the speed drops by a factor of approximately 2. Default value is
17926 Force a constant quantization parameter. If not set, the filter will use the QP
17927 from the video stream (if available).
17932 Convert 360 videos between various formats.
17934 The filter accepts the following options:
17940 Set format of the input/output video.
17948 Equirectangular projection.
17953 Cubemap with 3x2/6x1/1x6 layout.
17955 Format specific options:
17960 Set padding proportion for the input/output cubemap. Values in decimals.
17967 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)
17970 Default value is @b{@samp{0}}.
17974 Set fixed padding for the input/output cubemap. Values in pixels.
17976 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
17980 Set order of faces for the input/output cubemap. Choose one direction for each position.
17982 Designation of directions:
17998 Default value is @b{@samp{rludfb}}.
18002 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
18004 Designation of angles:
18007 0 degrees clockwise
18009 90 degrees clockwise
18011 180 degrees clockwise
18013 270 degrees clockwise
18016 Default value is @b{@samp{000000}}.
18020 Equi-Angular Cubemap.
18025 Regular video. @i{(output only)}
18027 Format specific options:
18032 Set horizontal/vertical/diagonal field of view. Values in degrees.
18034 If diagonal field of view is set it overrides horizontal and vertical field of view.
18040 Format specific options:
18044 Set padding proportion. Values in decimals.
18054 Default value is @b{@samp{0}}.
18059 Facebook's 360 format.
18062 Stereographic format.
18064 Format specific options:
18069 Set horizontal/vertical/diagonal field of view. Values in degrees.
18071 If diagonal field of view is set it overrides horizontal and vertical field of view.
18078 Ball format, gives significant distortion toward the back.
18081 Hammer-Aitoff map projection format.
18084 Sinusoidal map projection format.
18089 Set interpolation method.@*
18090 @i{Note: more complex interpolation methods require much more memory to run.}
18100 Bilinear interpolation.
18103 Bicubic interpolation.
18106 Lanczos interpolation.
18109 Default value is @b{@samp{line}}.
18113 Set the output video resolution.
18115 Default resolution depends on formats.
18119 Set the input/output stereo format.
18130 Default value is @b{@samp{2d}} for input and output format.
18135 Set rotation for the output video. Values in degrees.
18138 Set rotation order for the output video. Choose one item for each position.
18149 Default value is @b{@samp{ypr}}.
18154 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
18158 Set if input video is flipped horizontally/vertically. Boolean values.
18161 Set if input video is transposed. Boolean value, by default disabled.
18164 Set if output video needs to be transposed. Boolean value, by default disabled.
18168 @subsection Examples
18172 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
18174 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
18177 Extract back view of Equi-Angular Cubemap:
18179 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
18182 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
18184 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
18188 @section vaguedenoiser
18190 Apply a wavelet based denoiser.
18192 It transforms each frame from the video input into the wavelet domain,
18193 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
18194 the obtained coefficients. It does an inverse wavelet transform after.
18195 Due to wavelet properties, it should give a nice smoothed result, and
18196 reduced noise, without blurring picture features.
18198 This filter accepts the following options:
18202 The filtering strength. The higher, the more filtered the video will be.
18203 Hard thresholding can use a higher threshold than soft thresholding
18204 before the video looks overfiltered. Default value is 2.
18207 The filtering method the filter will use.
18209 It accepts the following values:
18212 All values under the threshold will be zeroed.
18215 All values under the threshold will be zeroed. All values above will be
18216 reduced by the threshold.
18219 Scales or nullifies coefficients - intermediary between (more) soft and
18220 (less) hard thresholding.
18223 Default is garrote.
18226 Number of times, the wavelet will decompose the picture. Picture can't
18227 be decomposed beyond a particular point (typically, 8 for a 640x480
18228 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
18231 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
18234 A list of the planes to process. By default all planes are processed.
18237 @section vectorscope
18239 Display 2 color component values in the two dimensional graph (which is called
18242 This filter accepts the following options:
18246 Set vectorscope mode.
18248 It accepts the following values:
18251 Gray values are displayed on graph, higher brightness means more pixels have
18252 same component color value on location in graph. This is the default mode.
18255 Gray values are displayed on graph. Surrounding pixels values which are not
18256 present in video frame are drawn in gradient of 2 color components which are
18257 set by option @code{x} and @code{y}. The 3rd color component is static.
18260 Actual color components values present in video frame are displayed on graph.
18263 Similar as color2 but higher frequency of same values @code{x} and @code{y}
18264 on graph increases value of another color component, which is luminance by
18265 default values of @code{x} and @code{y}.
18268 Actual colors present in video frame are displayed on graph. If two different
18269 colors map to same position on graph then color with higher value of component
18270 not present in graph is picked.
18273 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
18274 component picked from radial gradient.
18278 Set which color component will be represented on X-axis. Default is @code{1}.
18281 Set which color component will be represented on Y-axis. Default is @code{2}.
18284 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
18285 of color component which represents frequency of (X, Y) location in graph.
18290 No envelope, this is default.
18293 Instant envelope, even darkest single pixel will be clearly highlighted.
18296 Hold maximum and minimum values presented in graph over time. This way you
18297 can still spot out of range values without constantly looking at vectorscope.
18300 Peak and instant envelope combined together.
18304 Set what kind of graticule to draw.
18312 Set graticule opacity.
18315 Set graticule flags.
18319 Draw graticule for white point.
18322 Draw graticule for black point.
18325 Draw color points short names.
18329 Set background opacity.
18331 @item lthreshold, l
18332 Set low threshold for color component not represented on X or Y axis.
18333 Values lower than this value will be ignored. Default is 0.
18334 Note this value is multiplied with actual max possible value one pixel component
18335 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
18338 @item hthreshold, h
18339 Set high threshold for color component not represented on X or Y axis.
18340 Values higher than this value will be ignored. Default is 1.
18341 Note this value is multiplied with actual max possible value one pixel component
18342 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
18343 is 0.9 * 255 = 230.
18345 @item colorspace, c
18346 Set what kind of colorspace to use when drawing graticule.
18355 @anchor{vidstabdetect}
18356 @section vidstabdetect
18358 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
18359 @ref{vidstabtransform} for pass 2.
18361 This filter generates a file with relative translation and rotation
18362 transform information about subsequent frames, which is then used by
18363 the @ref{vidstabtransform} filter.
18365 To enable compilation of this filter you need to configure FFmpeg with
18366 @code{--enable-libvidstab}.
18368 This filter accepts the following options:
18372 Set the path to the file used to write the transforms information.
18373 Default value is @file{transforms.trf}.
18376 Set how shaky the video is and how quick the camera is. It accepts an
18377 integer in the range 1-10, a value of 1 means little shakiness, a
18378 value of 10 means strong shakiness. Default value is 5.
18381 Set the accuracy of the detection process. It must be a value in the
18382 range 1-15. A value of 1 means low accuracy, a value of 15 means high
18383 accuracy. Default value is 15.
18386 Set stepsize of the search process. The region around minimum is
18387 scanned with 1 pixel resolution. Default value is 6.
18390 Set minimum contrast. Below this value a local measurement field is
18391 discarded. Must be a floating point value in the range 0-1. Default
18395 Set reference frame number for tripod mode.
18397 If enabled, the motion of the frames is compared to a reference frame
18398 in the filtered stream, identified by the specified number. The idea
18399 is to compensate all movements in a more-or-less static scene and keep
18400 the camera view absolutely still.
18402 If set to 0, it is disabled. The frames are counted starting from 1.
18405 Show fields and transforms in the resulting frames. It accepts an
18406 integer in the range 0-2. Default value is 0, which disables any
18410 @subsection Examples
18414 Use default values:
18420 Analyze strongly shaky movie and put the results in file
18421 @file{mytransforms.trf}:
18423 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
18427 Visualize the result of internal transformations in the resulting
18430 vidstabdetect=show=1
18434 Analyze a video with medium shakiness using @command{ffmpeg}:
18436 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
18440 @anchor{vidstabtransform}
18441 @section vidstabtransform
18443 Video stabilization/deshaking: pass 2 of 2,
18444 see @ref{vidstabdetect} for pass 1.
18446 Read a file with transform information for each frame and
18447 apply/compensate them. Together with the @ref{vidstabdetect}
18448 filter this can be used to deshake videos. See also
18449 @url{http://public.hronopik.de/vid.stab}. It is important to also use
18450 the @ref{unsharp} filter, see below.
18452 To enable compilation of this filter you need to configure FFmpeg with
18453 @code{--enable-libvidstab}.
18455 @subsection Options
18459 Set path to the file used to read the transforms. Default value is
18460 @file{transforms.trf}.
18463 Set the number of frames (value*2 + 1) used for lowpass filtering the
18464 camera movements. Default value is 10.
18466 For example a number of 10 means that 21 frames are used (10 in the
18467 past and 10 in the future) to smoothen the motion in the video. A
18468 larger value leads to a smoother video, but limits the acceleration of
18469 the camera (pan/tilt movements). 0 is a special case where a static
18470 camera is simulated.
18473 Set the camera path optimization algorithm.
18475 Accepted values are:
18478 gaussian kernel low-pass filter on camera motion (default)
18480 averaging on transformations
18484 Set maximal number of pixels to translate frames. Default value is -1,
18488 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
18489 value is -1, meaning no limit.
18492 Specify how to deal with borders that may be visible due to movement
18495 Available values are:
18498 keep image information from previous frame (default)
18500 fill the border black
18504 Invert transforms if set to 1. Default value is 0.
18507 Consider transforms as relative to previous frame if set to 1,
18508 absolute if set to 0. Default value is 0.
18511 Set percentage to zoom. A positive value will result in a zoom-in
18512 effect, a negative value in a zoom-out effect. Default value is 0 (no
18516 Set optimal zooming to avoid borders.
18518 Accepted values are:
18523 optimal static zoom value is determined (only very strong movements
18524 will lead to visible borders) (default)
18526 optimal adaptive zoom value is determined (no borders will be
18527 visible), see @option{zoomspeed}
18530 Note that the value given at zoom is added to the one calculated here.
18533 Set percent to zoom maximally each frame (enabled when
18534 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
18538 Specify type of interpolation.
18540 Available values are:
18545 linear only horizontal
18547 linear in both directions (default)
18549 cubic in both directions (slow)
18553 Enable virtual tripod mode if set to 1, which is equivalent to
18554 @code{relative=0:smoothing=0}. Default value is 0.
18556 Use also @code{tripod} option of @ref{vidstabdetect}.
18559 Increase log verbosity if set to 1. Also the detected global motions
18560 are written to the temporary file @file{global_motions.trf}. Default
18564 @subsection Examples
18568 Use @command{ffmpeg} for a typical stabilization with default values:
18570 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
18573 Note the use of the @ref{unsharp} filter which is always recommended.
18576 Zoom in a bit more and load transform data from a given file:
18578 vidstabtransform=zoom=5:input="mytransforms.trf"
18582 Smoothen the video even more:
18584 vidstabtransform=smoothing=30
18590 Flip the input video vertically.
18592 For example, to vertically flip a video with @command{ffmpeg}:
18594 ffmpeg -i in.avi -vf "vflip" out.avi
18599 Detect variable frame rate video.
18601 This filter tries to detect if the input is variable or constant frame rate.
18603 At end it will output number of frames detected as having variable delta pts,
18604 and ones with constant delta pts.
18605 If there was frames with variable delta, than it will also show min and max delta
18610 Boost or alter saturation.
18612 The filter accepts the following options:
18615 Set strength of boost if positive value or strength of alter if negative value.
18616 Default is 0. Allowed range is from -2 to 2.
18619 Set the red balance. Default is 1. Allowed range is from -10 to 10.
18622 Set the green balance. Default is 1. Allowed range is from -10 to 10.
18625 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
18628 Set the red luma coefficient.
18631 Set the green luma coefficient.
18634 Set the blue luma coefficient.
18637 If @code{intensity} is negative and this is set to 1, colors will change,
18638 otherwise colors will be less saturated, more towards gray.
18644 Make or reverse a natural vignetting effect.
18646 The filter accepts the following options:
18650 Set lens angle expression as a number of radians.
18652 The value is clipped in the @code{[0,PI/2]} range.
18654 Default value: @code{"PI/5"}
18658 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
18662 Set forward/backward mode.
18664 Available modes are:
18667 The larger the distance from the central point, the darker the image becomes.
18670 The larger the distance from the central point, the brighter the image becomes.
18671 This can be used to reverse a vignette effect, though there is no automatic
18672 detection to extract the lens @option{angle} and other settings (yet). It can
18673 also be used to create a burning effect.
18676 Default value is @samp{forward}.
18679 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
18681 It accepts the following values:
18684 Evaluate expressions only once during the filter initialization.
18687 Evaluate expressions for each incoming frame. This is way slower than the
18688 @samp{init} mode since it requires all the scalers to be re-computed, but it
18689 allows advanced dynamic expressions.
18692 Default value is @samp{init}.
18695 Set dithering to reduce the circular banding effects. Default is @code{1}
18699 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
18700 Setting this value to the SAR of the input will make a rectangular vignetting
18701 following the dimensions of the video.
18703 Default is @code{1/1}.
18706 @subsection Expressions
18708 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
18709 following parameters.
18714 input width and height
18717 the number of input frame, starting from 0
18720 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
18721 @var{TB} units, NAN if undefined
18724 frame rate of the input video, NAN if the input frame rate is unknown
18727 the PTS (Presentation TimeStamp) of the filtered video frame,
18728 expressed in seconds, NAN if undefined
18731 time base of the input video
18735 @subsection Examples
18739 Apply simple strong vignetting effect:
18745 Make a flickering vignetting:
18747 vignette='PI/4+random(1)*PI/50':eval=frame
18752 @section vmafmotion
18754 Obtain the average vmaf motion score of a video.
18755 It is one of the component filters of VMAF.
18757 The obtained average motion score is printed through the logging system.
18759 In the below example the input file @file{ref.mpg} is being processed and score
18763 ffmpeg -i ref.mpg -lavfi vmafmotion -f null -
18767 Stack input videos vertically.
18769 All streams must be of same pixel format and of same width.
18771 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
18772 to create same output.
18774 The filter accepts the following options:
18778 Set number of input streams. Default is 2.
18781 If set to 1, force the output to terminate when the shortest input
18782 terminates. Default value is 0.
18787 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
18788 Deinterlacing Filter").
18790 Based on the process described by Martin Weston for BBC R&D, and
18791 implemented based on the de-interlace algorithm written by Jim
18792 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
18793 uses filter coefficients calculated by BBC R&D.
18795 This filter uses field-dominance information in frame to decide which
18796 of each pair of fields to place first in the output.
18797 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
18799 There are two sets of filter coefficients, so called "simple"
18800 and "complex". Which set of filter coefficients is used can
18801 be set by passing an optional parameter:
18805 Set the interlacing filter coefficients. Accepts one of the following values:
18809 Simple filter coefficient set.
18811 More-complex filter coefficient set.
18813 Default value is @samp{complex}.
18816 Specify which frames to deinterlace. Accepts one of the following values:
18820 Deinterlace all frames,
18822 Only deinterlace frames marked as interlaced.
18825 Default value is @samp{all}.
18829 Video waveform monitor.
18831 The waveform monitor plots color component intensity. By default luminance
18832 only. Each column of the waveform corresponds to a column of pixels in the
18835 It accepts the following options:
18839 Can be either @code{row}, or @code{column}. Default is @code{column}.
18840 In row mode, the graph on the left side represents color component value 0 and
18841 the right side represents value = 255. In column mode, the top side represents
18842 color component value = 0 and bottom side represents value = 255.
18845 Set intensity. Smaller values are useful to find out how many values of the same
18846 luminance are distributed across input rows/columns.
18847 Default value is @code{0.04}. Allowed range is [0, 1].
18850 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
18851 In mirrored mode, higher values will be represented on the left
18852 side for @code{row} mode and at the top for @code{column} mode. Default is
18853 @code{1} (mirrored).
18857 It accepts the following values:
18860 Presents information identical to that in the @code{parade}, except
18861 that the graphs representing color components are superimposed directly
18864 This display mode makes it easier to spot relative differences or similarities
18865 in overlapping areas of the color components that are supposed to be identical,
18866 such as neutral whites, grays, or blacks.
18869 Display separate graph for the color components side by side in
18870 @code{row} mode or one below the other in @code{column} mode.
18873 Display separate graph for the color components side by side in
18874 @code{column} mode or one below the other in @code{row} mode.
18876 Using this display mode makes it easy to spot color casts in the highlights
18877 and shadows of an image, by comparing the contours of the top and the bottom
18878 graphs of each waveform. Since whites, grays, and blacks are characterized
18879 by exactly equal amounts of red, green, and blue, neutral areas of the picture
18880 should display three waveforms of roughly equal width/height. If not, the
18881 correction is easy to perform by making level adjustments the three waveforms.
18883 Default is @code{stack}.
18885 @item components, c
18886 Set which color components to display. Default is 1, which means only luminance
18887 or red color component if input is in RGB colorspace. If is set for example to
18888 7 it will display all 3 (if) available color components.
18893 No envelope, this is default.
18896 Instant envelope, minimum and maximum values presented in graph will be easily
18897 visible even with small @code{step} value.
18900 Hold minimum and maximum values presented in graph across time. This way you
18901 can still spot out of range values without constantly looking at waveforms.
18904 Peak and instant envelope combined together.
18910 No filtering, this is default.
18913 Luma and chroma combined together.
18916 Similar as above, but shows difference between blue and red chroma.
18919 Similar as above, but use different colors.
18922 Displays only chroma.
18925 Displays actual color value on waveform.
18928 Similar as above, but with luma showing frequency of chroma values.
18932 Set which graticule to display.
18936 Do not display graticule.
18939 Display green graticule showing legal broadcast ranges.
18942 Display orange graticule showing legal broadcast ranges.
18946 Set graticule opacity.
18949 Set graticule flags.
18953 Draw numbers above lines. By default enabled.
18956 Draw dots instead of lines.
18960 Set scale used for displaying graticule.
18967 Default is digital.
18970 Set background opacity.
18973 @section weave, doubleweave
18975 The @code{weave} takes a field-based video input and join
18976 each two sequential fields into single frame, producing a new double
18977 height clip with half the frame rate and half the frame count.
18979 The @code{doubleweave} works same as @code{weave} but without
18980 halving frame rate and frame count.
18982 It accepts the following option:
18986 Set first field. Available values are:
18990 Set the frame as top-field-first.
18993 Set the frame as bottom-field-first.
18997 @subsection Examples
19001 Interlace video using @ref{select} and @ref{separatefields} filter:
19003 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
19008 Apply the xBR high-quality magnification filter which is designed for pixel
19009 art. It follows a set of edge-detection rules, see
19010 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
19012 It accepts the following option:
19016 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
19017 @code{3xBR} and @code{4} for @code{4xBR}.
19018 Default is @code{3}.
19022 Pick median pixels from several input videos.
19024 The filter accepts the following options:
19028 Set number of inputs.
19029 Default is 3. Allowed range is from 3 to 255.
19030 If number of inputs is even number, than result will be mean value between two median values.
19033 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19037 Stack video inputs into custom layout.
19039 All streams must be of same pixel format.
19041 The filter accepts the following options:
19045 Set number of input streams. Default is 2.
19048 Specify layout of inputs.
19049 This option requires the desired layout configuration to be explicitly set by the user.
19050 This sets position of each video input in output. Each input
19051 is separated by '|'.
19052 The first number represents the column, and the second number represents the row.
19053 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
19054 where X is video input from which to take width or height.
19055 Multiple values can be used when separated by '+'. In such
19056 case values are summed together.
19058 Note that if inputs are of different sizes gaps may appear, as not all of
19059 the output video frame will be filled. Similarly, videos can overlap each
19060 other if their position doesn't leave enough space for the full frame of
19063 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
19064 a layout must be set by the user.
19067 If set to 1, force the output to terminate when the shortest input
19068 terminates. Default value is 0.
19071 @subsection Examples
19075 Display 4 inputs into 2x2 grid.
19079 input1(0, 0) | input3(w0, 0)
19080 input2(0, h0) | input4(w0, h0)
19084 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
19087 Note that if inputs are of different sizes, gaps or overlaps may occur.
19090 Display 4 inputs into 1x4 grid.
19097 input4(0, h0+h1+h2)
19101 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
19104 Note that if inputs are of different widths, unused space will appear.
19107 Display 9 inputs into 3x3 grid.
19111 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
19112 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
19113 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
19117 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
19120 Note that if inputs are of different sizes, gaps or overlaps may occur.
19123 Display 16 inputs into 4x4 grid.
19127 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
19128 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
19129 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
19130 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
19134 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|
19135 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
19138 Note that if inputs are of different sizes, gaps or overlaps may occur.
19145 Deinterlace the input video ("yadif" means "yet another deinterlacing
19148 It accepts the following parameters:
19154 The interlacing mode to adopt. It accepts one of the following values:
19157 @item 0, send_frame
19158 Output one frame for each frame.
19159 @item 1, send_field
19160 Output one frame for each field.
19161 @item 2, send_frame_nospatial
19162 Like @code{send_frame}, but it skips the spatial interlacing check.
19163 @item 3, send_field_nospatial
19164 Like @code{send_field}, but it skips the spatial interlacing check.
19167 The default value is @code{send_frame}.
19170 The picture field parity assumed for the input interlaced video. It accepts one
19171 of the following values:
19175 Assume the top field is first.
19177 Assume the bottom field is first.
19179 Enable automatic detection of field parity.
19182 The default value is @code{auto}.
19183 If the interlacing is unknown or the decoder does not export this information,
19184 top field first will be assumed.
19187 Specify which frames to deinterlace. Accepts one of the following
19192 Deinterlace all frames.
19193 @item 1, interlaced
19194 Only deinterlace frames marked as interlaced.
19197 The default value is @code{all}.
19200 @section yadif_cuda
19202 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
19203 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
19206 It accepts the following parameters:
19212 The interlacing mode to adopt. It accepts one of the following values:
19215 @item 0, send_frame
19216 Output one frame for each frame.
19217 @item 1, send_field
19218 Output one frame for each field.
19219 @item 2, send_frame_nospatial
19220 Like @code{send_frame}, but it skips the spatial interlacing check.
19221 @item 3, send_field_nospatial
19222 Like @code{send_field}, but it skips the spatial interlacing check.
19225 The default value is @code{send_frame}.
19228 The picture field parity assumed for the input interlaced video. It accepts one
19229 of the following values:
19233 Assume the top field is first.
19235 Assume the bottom field is first.
19237 Enable automatic detection of field parity.
19240 The default value is @code{auto}.
19241 If the interlacing is unknown or the decoder does not export this information,
19242 top field first will be assumed.
19245 Specify which frames to deinterlace. Accepts one of the following
19250 Deinterlace all frames.
19251 @item 1, interlaced
19252 Only deinterlace frames marked as interlaced.
19255 The default value is @code{all}.
19260 Apply Zoom & Pan effect.
19262 This filter accepts the following options:
19266 Set the zoom expression. Range is 1-10. Default is 1.
19270 Set the x and y expression. Default is 0.
19273 Set the duration expression in number of frames.
19274 This sets for how many number of frames effect will last for
19275 single input image.
19278 Set the output image size, default is 'hd720'.
19281 Set the output frame rate, default is '25'.
19284 Each expression can contain the following constants:
19303 Output frame count.
19307 Last calculated 'x' and 'y' position from 'x' and 'y' expression
19308 for current input frame.
19312 'x' and 'y' of last output frame of previous input frame or 0 when there was
19313 not yet such frame (first input frame).
19316 Last calculated zoom from 'z' expression for current input frame.
19319 Last calculated zoom of last output frame of previous input frame.
19322 Number of output frames for current input frame. Calculated from 'd' expression
19323 for each input frame.
19326 number of output frames created for previous input frame
19329 Rational number: input width / input height
19332 sample aspect ratio
19335 display aspect ratio
19339 @subsection Examples
19343 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
19345 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
19349 Zoom-in up to 1.5 and pan always at center of picture:
19351 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
19355 Same as above but without pausing:
19357 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
19363 Scale (resize) the input video, using the z.lib library:
19364 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
19365 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
19367 The zscale filter forces the output display aspect ratio to be the same
19368 as the input, by changing the output sample aspect ratio.
19370 If the input image format is different from the format requested by
19371 the next filter, the zscale filter will convert the input to the
19374 @subsection Options
19375 The filter accepts the following options.
19380 Set the output video dimension expression. Default value is the input
19383 If the @var{width} or @var{w} value is 0, the input width is used for
19384 the output. If the @var{height} or @var{h} value is 0, the input height
19385 is used for the output.
19387 If one and only one of the values is -n with n >= 1, the zscale filter
19388 will use a value that maintains the aspect ratio of the input image,
19389 calculated from the other specified dimension. After that it will,
19390 however, make sure that the calculated dimension is divisible by n and
19391 adjust the value if necessary.
19393 If both values are -n with n >= 1, the behavior will be identical to
19394 both values being set to 0 as previously detailed.
19396 See below for the list of accepted constants for use in the dimension
19400 Set the video size. For the syntax of this option, check the
19401 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19404 Set the dither type.
19406 Possible values are:
19411 @item error_diffusion
19417 Set the resize filter type.
19419 Possible values are:
19429 Default is bilinear.
19432 Set the color range.
19434 Possible values are:
19441 Default is same as input.
19444 Set the color primaries.
19446 Possible values are:
19456 Default is same as input.
19459 Set the transfer characteristics.
19461 Possible values are:
19475 Default is same as input.
19478 Set the colorspace matrix.
19480 Possible value are:
19491 Default is same as input.
19494 Set the input color range.
19496 Possible values are:
19503 Default is same as input.
19505 @item primariesin, pin
19506 Set the input color primaries.
19508 Possible values are:
19518 Default is same as input.
19520 @item transferin, tin
19521 Set the input transfer characteristics.
19523 Possible values are:
19534 Default is same as input.
19536 @item matrixin, min
19537 Set the input colorspace matrix.
19539 Possible value are:
19551 Set the output chroma location.
19553 Possible values are:
19564 @item chromalin, cin
19565 Set the input chroma location.
19567 Possible values are:
19579 Set the nominal peak luminance.
19582 The values of the @option{w} and @option{h} options are expressions
19583 containing the following constants:
19588 The input width and height
19592 These are the same as @var{in_w} and @var{in_h}.
19596 The output (scaled) width and height
19600 These are the same as @var{out_w} and @var{out_h}
19603 The same as @var{iw} / @var{ih}
19606 input sample aspect ratio
19609 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
19613 horizontal and vertical input chroma subsample values. For example for the
19614 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
19618 horizontal and vertical output chroma subsample values. For example for the
19619 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
19625 @c man end VIDEO FILTERS
19627 @chapter OpenCL Video Filters
19628 @c man begin OPENCL VIDEO FILTERS
19630 Below is a description of the currently available OpenCL video filters.
19632 To enable compilation of these filters you need to configure FFmpeg with
19633 @code{--enable-opencl}.
19635 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
19638 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
19639 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
19640 given device parameters.
19642 @item -filter_hw_device @var{name}
19643 Pass the hardware device called @var{name} to all filters in any filter graph.
19647 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
19651 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
19653 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
19657 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.
19659 @section avgblur_opencl
19661 Apply average blur filter.
19663 The filter accepts the following options:
19667 Set horizontal radius size.
19668 Range is @code{[1, 1024]} and default value is @code{1}.
19671 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19674 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
19677 @subsection Example
19681 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.
19683 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
19687 @section boxblur_opencl
19689 Apply a boxblur algorithm to the input video.
19691 It accepts the following parameters:
19695 @item luma_radius, lr
19696 @item luma_power, lp
19697 @item chroma_radius, cr
19698 @item chroma_power, cp
19699 @item alpha_radius, ar
19700 @item alpha_power, ap
19704 A description of the accepted options follows.
19707 @item luma_radius, lr
19708 @item chroma_radius, cr
19709 @item alpha_radius, ar
19710 Set an expression for the box radius in pixels used for blurring the
19711 corresponding input plane.
19713 The radius value must be a non-negative number, and must not be
19714 greater than the value of the expression @code{min(w,h)/2} for the
19715 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
19718 Default value for @option{luma_radius} is "2". If not specified,
19719 @option{chroma_radius} and @option{alpha_radius} default to the
19720 corresponding value set for @option{luma_radius}.
19722 The expressions can contain the following constants:
19726 The input width and height in pixels.
19730 The input chroma image width and height in pixels.
19734 The horizontal and vertical chroma subsample values. For example, for the
19735 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
19738 @item luma_power, lp
19739 @item chroma_power, cp
19740 @item alpha_power, ap
19741 Specify how many times the boxblur filter is applied to the
19742 corresponding plane.
19744 Default value for @option{luma_power} is 2. If not specified,
19745 @option{chroma_power} and @option{alpha_power} default to the
19746 corresponding value set for @option{luma_power}.
19748 A value of 0 will disable the effect.
19751 @subsection Examples
19753 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.
19757 Apply a boxblur filter with the luma, chroma, and alpha radius
19758 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.
19760 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
19761 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
19765 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.
19767 For the luma plane, a 2x2 box radius will be run once.
19769 For the chroma plane, a 4x4 box radius will be run 5 times.
19771 For the alpha plane, a 3x3 box radius will be run 7 times.
19773 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
19777 @section convolution_opencl
19779 Apply convolution of 3x3, 5x5, 7x7 matrix.
19781 The filter accepts the following options:
19788 Set matrix for each plane.
19789 Matrix is sequence of 9, 25 or 49 signed numbers.
19790 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
19796 Set multiplier for calculated value for each plane.
19797 If unset or 0, it will be sum of all matrix elements.
19798 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
19804 Set bias for each plane. This value is added to the result of the multiplication.
19805 Useful for making the overall image brighter or darker.
19806 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
19810 @subsection Examples
19816 -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
19822 -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
19826 Apply edge enhance:
19828 -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
19834 -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
19838 Apply laplacian edge detector which includes diagonals:
19840 -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
19846 -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
19850 @section dilation_opencl
19852 Apply dilation effect to the video.
19854 This filter replaces the pixel by the local(3x3) maximum.
19856 It accepts the following options:
19863 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
19864 If @code{0}, plane will remain unchanged.
19867 Flag which specifies the pixel to refer to.
19868 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
19870 Flags to local 3x3 coordinates region centered on @code{x}:
19879 @subsection Example
19883 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.
19885 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
19889 @section erosion_opencl
19891 Apply erosion effect to the video.
19893 This filter replaces the pixel by the local(3x3) minimum.
19895 It accepts the following options:
19902 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
19903 If @code{0}, plane will remain unchanged.
19906 Flag which specifies the pixel to refer to.
19907 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
19909 Flags to local 3x3 coordinates region centered on @code{x}:
19918 @subsection Example
19922 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.
19924 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
19928 @section colorkey_opencl
19929 RGB colorspace color keying.
19931 The filter accepts the following options:
19935 The color which will be replaced with transparency.
19938 Similarity percentage with the key color.
19940 0.01 matches only the exact key color, while 1.0 matches everything.
19945 0.0 makes pixels either fully transparent, or not transparent at all.
19947 Higher values result in semi-transparent pixels, with a higher transparency
19948 the more similar the pixels color is to the key color.
19951 @subsection Examples
19955 Make every semi-green pixel in the input transparent with some slight blending:
19957 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
19961 @section deshake_opencl
19962 Feature-point based video stabilization filter.
19964 The filter accepts the following options:
19968 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
19971 Whether or not additional debug info should be displayed, both in the processed output and in the console.
19973 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
19975 Viewing point matches in the output video is only supported for RGB input.
19977 Defaults to @code{0}.
19979 @item adaptive_crop
19980 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
19982 Defaults to @code{1}.
19984 @item refine_features
19985 Whether or not feature points should be refined at a sub-pixel level.
19987 This can be turned off for a slight performance gain at the cost of precision.
19989 Defaults to @code{1}.
19991 @item smooth_strength
19992 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
19994 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
19996 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
19998 Defaults to @code{0.0}.
20000 @item smooth_window_multiplier
20001 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
20003 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
20005 Acceptable values range from @code{0.1} to @code{10.0}.
20007 Larger values increase the amount of motion data available for determining how to smooth the camera path,
20008 potentially improving smoothness, but also increase latency and memory usage.
20010 Defaults to @code{2.0}.
20014 @subsection Examples
20018 Stabilize a video with a fixed, medium smoothing strength:
20020 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
20024 Stabilize a video with debugging (both in console and in rendered video):
20026 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
20030 @section nlmeans_opencl
20032 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
20034 @section overlay_opencl
20036 Overlay one video on top of another.
20038 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
20039 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
20041 The filter accepts the following options:
20046 Set the x coordinate of the overlaid video on the main video.
20047 Default value is @code{0}.
20050 Set the x coordinate of the overlaid video on the main video.
20051 Default value is @code{0}.
20055 @subsection Examples
20059 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
20061 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
20064 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
20066 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
20071 @section prewitt_opencl
20073 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
20075 The filter accepts the following option:
20079 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20082 Set value which will be multiplied with filtered result.
20083 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
20086 Set value which will be added to filtered result.
20087 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
20090 @subsection Example
20094 Apply the Prewitt operator with scale set to 2 and delta set to 10.
20096 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
20100 @section roberts_opencl
20101 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
20103 The filter accepts the following option:
20107 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20110 Set value which will be multiplied with filtered result.
20111 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
20114 Set value which will be added to filtered result.
20115 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
20118 @subsection Example
20122 Apply the Roberts cross operator with scale set to 2 and delta set to 10
20124 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
20128 @section sobel_opencl
20130 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
20132 The filter accepts the following option:
20136 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20139 Set value which will be multiplied with filtered result.
20140 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
20143 Set value which will be added to filtered result.
20144 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
20147 @subsection Example
20151 Apply sobel operator with scale set to 2 and delta set to 10
20153 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
20157 @section tonemap_opencl
20159 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
20161 It accepts the following parameters:
20165 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
20168 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
20171 Apply desaturation for highlights that exceed this level of brightness. The
20172 higher the parameter, the more color information will be preserved. This
20173 setting helps prevent unnaturally blown-out colors for super-highlights, by
20174 (smoothly) turning into white instead. This makes images feel more natural,
20175 at the cost of reducing information about out-of-range colors.
20177 The default value is 0.5, and the algorithm here is a little different from
20178 the cpu version tonemap currently. A setting of 0.0 disables this option.
20181 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
20182 is used to detect whether the scene has changed or not. If the distance between
20183 the current frame average brightness and the current running average exceeds
20184 a threshold value, we would re-calculate scene average and peak brightness.
20185 The default value is 0.2.
20188 Specify the output pixel format.
20190 Currently supported formats are:
20197 Set the output color range.
20199 Possible values are:
20205 Default is same as input.
20208 Set the output color primaries.
20210 Possible values are:
20216 Default is same as input.
20219 Set the output transfer characteristics.
20221 Possible values are:
20230 Set the output colorspace matrix.
20232 Possible value are:
20238 Default is same as input.
20242 @subsection Example
20246 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
20248 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
20252 @section unsharp_opencl
20254 Sharpen or blur the input video.
20256 It accepts the following parameters:
20259 @item luma_msize_x, lx
20260 Set the luma matrix horizontal size.
20261 Range is @code{[1, 23]} and default value is @code{5}.
20263 @item luma_msize_y, ly
20264 Set the luma matrix vertical size.
20265 Range is @code{[1, 23]} and default value is @code{5}.
20267 @item luma_amount, la
20268 Set the luma effect strength.
20269 Range is @code{[-10, 10]} and default value is @code{1.0}.
20271 Negative values will blur the input video, while positive values will
20272 sharpen it, a value of zero will disable the effect.
20274 @item chroma_msize_x, cx
20275 Set the chroma matrix horizontal size.
20276 Range is @code{[1, 23]} and default value is @code{5}.
20278 @item chroma_msize_y, cy
20279 Set the chroma matrix vertical size.
20280 Range is @code{[1, 23]} and default value is @code{5}.
20282 @item chroma_amount, ca
20283 Set the chroma effect strength.
20284 Range is @code{[-10, 10]} and default value is @code{0.0}.
20286 Negative values will blur the input video, while positive values will
20287 sharpen it, a value of zero will disable the effect.
20291 All parameters are optional and default to the equivalent of the
20292 string '5:5:1.0:5:5:0.0'.
20294 @subsection Examples
20298 Apply strong luma sharpen effect:
20300 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
20304 Apply a strong blur of both luma and chroma parameters:
20306 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
20310 @c man end OPENCL VIDEO FILTERS
20312 @chapter Video Sources
20313 @c man begin VIDEO SOURCES
20315 Below is a description of the currently available video sources.
20319 Buffer video frames, and make them available to the filter chain.
20321 This source is mainly intended for a programmatic use, in particular
20322 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
20324 It accepts the following parameters:
20329 Specify the size (width and height) of the buffered video frames. For the
20330 syntax of this option, check the
20331 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20334 The input video width.
20337 The input video height.
20340 A string representing the pixel format of the buffered video frames.
20341 It may be a number corresponding to a pixel format, or a pixel format
20345 Specify the timebase assumed by the timestamps of the buffered frames.
20348 Specify the frame rate expected for the video stream.
20350 @item pixel_aspect, sar
20351 The sample (pixel) aspect ratio of the input video.
20354 Specify the optional parameters to be used for the scale filter which
20355 is automatically inserted when an input change is detected in the
20356 input size or format.
20358 @item hw_frames_ctx
20359 When using a hardware pixel format, this should be a reference to an
20360 AVHWFramesContext describing input frames.
20365 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
20368 will instruct the source to accept video frames with size 320x240 and
20369 with format "yuv410p", assuming 1/24 as the timestamps timebase and
20370 square pixels (1:1 sample aspect ratio).
20371 Since the pixel format with name "yuv410p" corresponds to the number 6
20372 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
20373 this example corresponds to:
20375 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
20378 Alternatively, the options can be specified as a flat string, but this
20379 syntax is deprecated:
20381 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}[:@var{sws_param}]
20385 Create a pattern generated by an elementary cellular automaton.
20387 The initial state of the cellular automaton can be defined through the
20388 @option{filename} and @option{pattern} options. If such options are
20389 not specified an initial state is created randomly.
20391 At each new frame a new row in the video is filled with the result of
20392 the cellular automaton next generation. The behavior when the whole
20393 frame is filled is defined by the @option{scroll} option.
20395 This source accepts the following options:
20399 Read the initial cellular automaton state, i.e. the starting row, from
20400 the specified file.
20401 In the file, each non-whitespace character is considered an alive
20402 cell, a newline will terminate the row, and further characters in the
20403 file will be ignored.
20406 Read the initial cellular automaton state, i.e. the starting row, from
20407 the specified string.
20409 Each non-whitespace character in the string is considered an alive
20410 cell, a newline will terminate the row, and further characters in the
20411 string will be ignored.
20414 Set the video rate, that is the number of frames generated per second.
20417 @item random_fill_ratio, ratio
20418 Set the random fill ratio for the initial cellular automaton row. It
20419 is a floating point number value ranging from 0 to 1, defaults to
20422 This option is ignored when a file or a pattern is specified.
20424 @item random_seed, seed
20425 Set the seed for filling randomly the initial row, must be an integer
20426 included between 0 and UINT32_MAX. If not specified, or if explicitly
20427 set to -1, the filter will try to use a good random seed on a best
20431 Set the cellular automaton rule, it is a number ranging from 0 to 255.
20432 Default value is 110.
20435 Set the size of the output video. For the syntax of this option, check the
20436 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20438 If @option{filename} or @option{pattern} is specified, the size is set
20439 by default to the width of the specified initial state row, and the
20440 height is set to @var{width} * PHI.
20442 If @option{size} is set, it must contain the width of the specified
20443 pattern string, and the specified pattern will be centered in the
20446 If a filename or a pattern string is not specified, the size value
20447 defaults to "320x518" (used for a randomly generated initial state).
20450 If set to 1, scroll the output upward when all the rows in the output
20451 have been already filled. If set to 0, the new generated row will be
20452 written over the top row just after the bottom row is filled.
20455 @item start_full, full
20456 If set to 1, completely fill the output with generated rows before
20457 outputting the first frame.
20458 This is the default behavior, for disabling set the value to 0.
20461 If set to 1, stitch the left and right row edges together.
20462 This is the default behavior, for disabling set the value to 0.
20465 @subsection Examples
20469 Read the initial state from @file{pattern}, and specify an output of
20472 cellauto=f=pattern:s=200x400
20476 Generate a random initial row with a width of 200 cells, with a fill
20479 cellauto=ratio=2/3:s=200x200
20483 Create a pattern generated by rule 18 starting by a single alive cell
20484 centered on an initial row with width 100:
20486 cellauto=p=@@:s=100x400:full=0:rule=18
20490 Specify a more elaborated initial pattern:
20492 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
20497 @anchor{coreimagesrc}
20498 @section coreimagesrc
20499 Video source generated on GPU using Apple's CoreImage API on OSX.
20501 This video source is a specialized version of the @ref{coreimage} video filter.
20502 Use a core image generator at the beginning of the applied filterchain to
20503 generate the content.
20505 The coreimagesrc video source accepts the following options:
20507 @item list_generators
20508 List all available generators along with all their respective options as well as
20509 possible minimum and maximum values along with the default values.
20511 list_generators=true
20515 Specify the size of the sourced video. For the syntax of this option, check the
20516 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20517 The default value is @code{320x240}.
20520 Specify the frame rate of the sourced video, as the number of frames
20521 generated per second. It has to be a string in the format
20522 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20523 number or a valid video frame rate abbreviation. The default value is
20527 Set the sample aspect ratio of the sourced video.
20530 Set the duration of the sourced video. See
20531 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20532 for the accepted syntax.
20534 If not specified, or the expressed duration is negative, the video is
20535 supposed to be generated forever.
20538 Additionally, all options of the @ref{coreimage} video filter are accepted.
20539 A complete filterchain can be used for further processing of the
20540 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
20541 and examples for details.
20543 @subsection Examples
20548 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
20549 given as complete and escaped command-line for Apple's standard bash shell:
20551 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
20553 This example is equivalent to the QRCode example of @ref{coreimage} without the
20554 need for a nullsrc video source.
20558 @section mandelbrot
20560 Generate a Mandelbrot set fractal, and progressively zoom towards the
20561 point specified with @var{start_x} and @var{start_y}.
20563 This source accepts the following options:
20568 Set the terminal pts value. Default value is 400.
20571 Set the terminal scale value.
20572 Must be a floating point value. Default value is 0.3.
20575 Set the inner coloring mode, that is the algorithm used to draw the
20576 Mandelbrot fractal internal region.
20578 It shall assume one of the following values:
20583 Show time until convergence.
20585 Set color based on point closest to the origin of the iterations.
20590 Default value is @var{mincol}.
20593 Set the bailout value. Default value is 10.0.
20596 Set the maximum of iterations performed by the rendering
20597 algorithm. Default value is 7189.
20600 Set outer coloring mode.
20601 It shall assume one of following values:
20603 @item iteration_count
20604 Set iteration count mode.
20605 @item normalized_iteration_count
20606 set normalized iteration count mode.
20608 Default value is @var{normalized_iteration_count}.
20611 Set frame rate, expressed as number of frames per second. Default
20615 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
20616 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
20619 Set the initial scale value. Default value is 3.0.
20622 Set the initial x position. Must be a floating point value between
20623 -100 and 100. Default value is -0.743643887037158704752191506114774.
20626 Set the initial y position. Must be a floating point value between
20627 -100 and 100. Default value is -0.131825904205311970493132056385139.
20632 Generate various test patterns, as generated by the MPlayer test filter.
20634 The size of the generated video is fixed, and is 256x256.
20635 This source is useful in particular for testing encoding features.
20637 This source accepts the following options:
20642 Specify the frame rate of the sourced video, as the number of frames
20643 generated per second. It has to be a string in the format
20644 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20645 number or a valid video frame rate abbreviation. The default value is
20649 Set the duration of the sourced video. See
20650 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20651 for the accepted syntax.
20653 If not specified, or the expressed duration is negative, the video is
20654 supposed to be generated forever.
20658 Set the number or the name of the test to perform. Supported tests are:
20674 Default value is "all", which will cycle through the list of all tests.
20679 mptestsrc=t=dc_luma
20682 will generate a "dc_luma" test pattern.
20684 @section frei0r_src
20686 Provide a frei0r source.
20688 To enable compilation of this filter you need to install the frei0r
20689 header and configure FFmpeg with @code{--enable-frei0r}.
20691 This source accepts the following parameters:
20696 The size of the video to generate. For the syntax of this option, check the
20697 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20700 The framerate of the generated video. It may be a string of the form
20701 @var{num}/@var{den} or a frame rate abbreviation.
20704 The name to the frei0r source to load. For more information regarding frei0r and
20705 how to set the parameters, read the @ref{frei0r} section in the video filters
20708 @item filter_params
20709 A '|'-separated list of parameters to pass to the frei0r source.
20713 For example, to generate a frei0r partik0l source with size 200x200
20714 and frame rate 10 which is overlaid on the overlay filter main input:
20716 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
20721 Generate a life pattern.
20723 This source is based on a generalization of John Conway's life game.
20725 The sourced input represents a life grid, each pixel represents a cell
20726 which can be in one of two possible states, alive or dead. Every cell
20727 interacts with its eight neighbours, which are the cells that are
20728 horizontally, vertically, or diagonally adjacent.
20730 At each interaction the grid evolves according to the adopted rule,
20731 which specifies the number of neighbor alive cells which will make a
20732 cell stay alive or born. The @option{rule} option allows one to specify
20735 This source accepts the following options:
20739 Set the file from which to read the initial grid state. In the file,
20740 each non-whitespace character is considered an alive cell, and newline
20741 is used to delimit the end of each row.
20743 If this option is not specified, the initial grid is generated
20747 Set the video rate, that is the number of frames generated per second.
20750 @item random_fill_ratio, ratio
20751 Set the random fill ratio for the initial random grid. It is a
20752 floating point number value ranging from 0 to 1, defaults to 1/PHI.
20753 It is ignored when a file is specified.
20755 @item random_seed, seed
20756 Set the seed for filling the initial random grid, must be an integer
20757 included between 0 and UINT32_MAX. If not specified, or if explicitly
20758 set to -1, the filter will try to use a good random seed on a best
20764 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
20765 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
20766 @var{NS} specifies the number of alive neighbor cells which make a
20767 live cell stay alive, and @var{NB} the number of alive neighbor cells
20768 which make a dead cell to become alive (i.e. to "born").
20769 "s" and "b" can be used in place of "S" and "B", respectively.
20771 Alternatively a rule can be specified by an 18-bits integer. The 9
20772 high order bits are used to encode the next cell state if it is alive
20773 for each number of neighbor alive cells, the low order bits specify
20774 the rule for "borning" new cells. Higher order bits encode for an
20775 higher number of neighbor cells.
20776 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
20777 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
20779 Default value is "S23/B3", which is the original Conway's game of life
20780 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
20781 cells, and will born a new cell if there are three alive cells around
20785 Set the size of the output video. For the syntax of this option, check the
20786 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20788 If @option{filename} is specified, the size is set by default to the
20789 same size of the input file. If @option{size} is set, it must contain
20790 the size specified in the input file, and the initial grid defined in
20791 that file is centered in the larger resulting area.
20793 If a filename is not specified, the size value defaults to "320x240"
20794 (used for a randomly generated initial grid).
20797 If set to 1, stitch the left and right grid edges together, and the
20798 top and bottom edges also. Defaults to 1.
20801 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
20802 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
20803 value from 0 to 255.
20806 Set the color of living (or new born) cells.
20809 Set the color of dead cells. If @option{mold} is set, this is the first color
20810 used to represent a dead cell.
20813 Set mold color, for definitely dead and moldy cells.
20815 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
20816 ffmpeg-utils manual,ffmpeg-utils}.
20819 @subsection Examples
20823 Read a grid from @file{pattern}, and center it on a grid of size
20826 life=f=pattern:s=300x300
20830 Generate a random grid of size 200x200, with a fill ratio of 2/3:
20832 life=ratio=2/3:s=200x200
20836 Specify a custom rule for evolving a randomly generated grid:
20842 Full example with slow death effect (mold) using @command{ffplay}:
20844 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
20851 @anchor{haldclutsrc}
20854 @anchor{pal100bars}
20855 @anchor{rgbtestsrc}
20857 @anchor{smptehdbars}
20860 @anchor{yuvtestsrc}
20861 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
20863 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
20865 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
20867 The @code{color} source provides an uniformly colored input.
20869 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
20870 @ref{haldclut} filter.
20872 The @code{nullsrc} source returns unprocessed video frames. It is
20873 mainly useful to be employed in analysis / debugging tools, or as the
20874 source for filters which ignore the input data.
20876 The @code{pal75bars} source generates a color bars pattern, based on
20877 EBU PAL recommendations with 75% color levels.
20879 The @code{pal100bars} source generates a color bars pattern, based on
20880 EBU PAL recommendations with 100% color levels.
20882 The @code{rgbtestsrc} source generates an RGB test pattern useful for
20883 detecting RGB vs BGR issues. You should see a red, green and blue
20884 stripe from top to bottom.
20886 The @code{smptebars} source generates a color bars pattern, based on
20887 the SMPTE Engineering Guideline EG 1-1990.
20889 The @code{smptehdbars} source generates a color bars pattern, based on
20890 the SMPTE RP 219-2002.
20892 The @code{testsrc} source generates a test video pattern, showing a
20893 color pattern, a scrolling gradient and a timestamp. This is mainly
20894 intended for testing purposes.
20896 The @code{testsrc2} source is similar to testsrc, but supports more
20897 pixel formats instead of just @code{rgb24}. This allows using it as an
20898 input for other tests without requiring a format conversion.
20900 The @code{yuvtestsrc} source generates an YUV test pattern. You should
20901 see a y, cb and cr stripe from top to bottom.
20903 The sources accept the following parameters:
20908 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
20909 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
20910 pixels to be used as identity matrix for 3D lookup tables. Each component is
20911 coded on a @code{1/(N*N)} scale.
20914 Specify the color of the source, only available in the @code{color}
20915 source. For the syntax of this option, check the
20916 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
20919 Specify the size of the sourced video. For the syntax of this option, check the
20920 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20921 The default value is @code{320x240}.
20923 This option is not available with the @code{allrgb}, @code{allyuv}, and
20924 @code{haldclutsrc} filters.
20927 Specify the frame rate of the sourced video, as the number of frames
20928 generated per second. It has to be a string in the format
20929 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20930 number or a valid video frame rate abbreviation. The default value is
20934 Set the duration of the sourced video. See
20935 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20936 for the accepted syntax.
20938 If not specified, or the expressed duration is negative, the video is
20939 supposed to be generated forever.
20942 Set the sample aspect ratio of the sourced video.
20945 Specify the alpha (opacity) of the background, only available in the
20946 @code{testsrc2} source. The value must be between 0 (fully transparent) and
20947 255 (fully opaque, the default).
20950 Set the number of decimals to show in the timestamp, only available in the
20951 @code{testsrc} source.
20953 The displayed timestamp value will correspond to the original
20954 timestamp value multiplied by the power of 10 of the specified
20955 value. Default value is 0.
20958 @subsection Examples
20962 Generate a video with a duration of 5.3 seconds, with size
20963 176x144 and a frame rate of 10 frames per second:
20965 testsrc=duration=5.3:size=qcif:rate=10
20969 The following graph description will generate a red source
20970 with an opacity of 0.2, with size "qcif" and a frame rate of 10
20973 color=c=red@@0.2:s=qcif:r=10
20977 If the input content is to be ignored, @code{nullsrc} can be used. The
20978 following command generates noise in the luminance plane by employing
20979 the @code{geq} filter:
20981 nullsrc=s=256x256, geq=random(1)*255:128:128
20985 @subsection Commands
20987 The @code{color} source supports the following commands:
20991 Set the color of the created image. Accepts the same syntax of the
20992 corresponding @option{color} option.
20997 Generate video using an OpenCL program.
21002 OpenCL program source file.
21005 Kernel name in program.
21008 Size of frames to generate. This must be set.
21011 Pixel format to use for the generated frames. This must be set.
21014 Number of frames generated every second. Default value is '25'.
21018 For details of how the program loading works, see the @ref{program_opencl}
21025 Generate a colour ramp by setting pixel values from the position of the pixel
21026 in the output image. (Note that this will work with all pixel formats, but
21027 the generated output will not be the same.)
21029 __kernel void ramp(__write_only image2d_t dst,
21030 unsigned int index)
21032 int2 loc = (int2)(get_global_id(0), get_global_id(1));
21035 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
21037 write_imagef(dst, loc, val);
21042 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
21044 __kernel void sierpinski_carpet(__write_only image2d_t dst,
21045 unsigned int index)
21047 int2 loc = (int2)(get_global_id(0), get_global_id(1));
21049 float4 value = 0.0f;
21050 int x = loc.x + index;
21051 int y = loc.y + index;
21052 while (x > 0 || y > 0) {
21053 if (x % 3 == 1 && y % 3 == 1) {
21061 write_imagef(dst, loc, value);
21067 @section sierpinski
21069 Generate a Sierpinksi carpet fractal, and randomly pan around.
21071 This source accepts the following options:
21075 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
21076 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
21079 Set frame rate, expressed as number of frames per second. Default
21083 Set seed which is used for random panning.
21086 Set max jump for single pan destination. Allowed range is from 1 to 10000.
21089 @c man end VIDEO SOURCES
21091 @chapter Video Sinks
21092 @c man begin VIDEO SINKS
21094 Below is a description of the currently available video sinks.
21096 @section buffersink
21098 Buffer video frames, and make them available to the end of the filter
21101 This sink is mainly intended for programmatic use, in particular
21102 through the interface defined in @file{libavfilter/buffersink.h}
21103 or the options system.
21105 It accepts a pointer to an AVBufferSinkContext structure, which
21106 defines the incoming buffers' formats, to be passed as the opaque
21107 parameter to @code{avfilter_init_filter} for initialization.
21111 Null video sink: do absolutely nothing with the input video. It is
21112 mainly useful as a template and for use in analysis / debugging
21115 @c man end VIDEO SINKS
21117 @chapter Multimedia Filters
21118 @c man begin MULTIMEDIA FILTERS
21120 Below is a description of the currently available multimedia filters.
21124 Convert input audio to a video output, displaying the audio bit scope.
21126 The filter accepts the following options:
21130 Set frame rate, expressed as number of frames per second. Default
21134 Specify the video size for the output. For the syntax of this option, check the
21135 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21136 Default value is @code{1024x256}.
21139 Specify list of colors separated by space or by '|' which will be used to
21140 draw channels. Unrecognized or missing colors will be replaced
21144 @section ahistogram
21146 Convert input audio to a video output, displaying the volume histogram.
21148 The filter accepts the following options:
21152 Specify how histogram is calculated.
21154 It accepts the following values:
21157 Use single histogram for all channels.
21159 Use separate histogram for each channel.
21161 Default is @code{single}.
21164 Set frame rate, expressed as number of frames per second. Default
21168 Specify the video size for the output. For the syntax of this option, check the
21169 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21170 Default value is @code{hd720}.
21175 It accepts the following values:
21186 reverse logarithmic
21188 Default is @code{log}.
21191 Set amplitude scale.
21193 It accepts the following values:
21200 Default is @code{log}.
21203 Set how much frames to accumulate in histogram.
21204 Default is 1. Setting this to -1 accumulates all frames.
21207 Set histogram ratio of window height.
21210 Set sonogram sliding.
21212 It accepts the following values:
21215 replace old rows with new ones.
21217 scroll from top to bottom.
21219 Default is @code{replace}.
21222 @section aphasemeter
21224 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
21225 representing mean phase of current audio frame. A video output can also be produced and is
21226 enabled by default. The audio is passed through as first output.
21228 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
21229 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
21230 and @code{1} means channels are in phase.
21232 The filter accepts the following options, all related to its video output:
21236 Set the output frame rate. Default value is @code{25}.
21239 Set the video size for the output. For the syntax of this option, check the
21240 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21241 Default value is @code{800x400}.
21246 Specify the red, green, blue contrast. Default values are @code{2},
21247 @code{7} and @code{1}.
21248 Allowed range is @code{[0, 255]}.
21251 Set color which will be used for drawing median phase. If color is
21252 @code{none} which is default, no median phase value will be drawn.
21255 Enable video output. Default is enabled.
21258 @section avectorscope
21260 Convert input audio to a video output, representing the audio vector
21263 The filter is used to measure the difference between channels of stereo
21264 audio stream. A monoaural signal, consisting of identical left and right
21265 signal, results in straight vertical line. Any stereo separation is visible
21266 as a deviation from this line, creating a Lissajous figure.
21267 If the straight (or deviation from it) but horizontal line appears this
21268 indicates that the left and right channels are out of phase.
21270 The filter accepts the following options:
21274 Set the vectorscope mode.
21276 Available values are:
21279 Lissajous rotated by 45 degrees.
21282 Same as above but not rotated.
21285 Shape resembling half of circle.
21288 Default value is @samp{lissajous}.
21291 Set the video size for the output. For the syntax of this option, check the
21292 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21293 Default value is @code{400x400}.
21296 Set the output frame rate. Default value is @code{25}.
21302 Specify the red, green, blue and alpha contrast. Default values are @code{40},
21303 @code{160}, @code{80} and @code{255}.
21304 Allowed range is @code{[0, 255]}.
21310 Specify the red, green, blue and alpha fade. Default values are @code{15},
21311 @code{10}, @code{5} and @code{5}.
21312 Allowed range is @code{[0, 255]}.
21315 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
21316 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
21319 Set the vectorscope drawing mode.
21321 Available values are:
21324 Draw dot for each sample.
21327 Draw line between previous and current sample.
21330 Default value is @samp{dot}.
21333 Specify amplitude scale of audio samples.
21335 Available values are:
21351 Swap left channel axis with right channel axis.
21361 Mirror only x axis.
21364 Mirror only y axis.
21372 @subsection Examples
21376 Complete example using @command{ffplay}:
21378 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
21379 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
21383 @section bench, abench
21385 Benchmark part of a filtergraph.
21387 The filter accepts the following options:
21391 Start or stop a timer.
21393 Available values are:
21396 Get the current time, set it as frame metadata (using the key
21397 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
21400 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
21401 the input frame metadata to get the time difference. Time difference, average,
21402 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
21403 @code{min}) are then printed. The timestamps are expressed in seconds.
21407 @subsection Examples
21411 Benchmark @ref{selectivecolor} filter:
21413 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
21419 Concatenate audio and video streams, joining them together one after the
21422 The filter works on segments of synchronized video and audio streams. All
21423 segments must have the same number of streams of each type, and that will
21424 also be the number of streams at output.
21426 The filter accepts the following options:
21431 Set the number of segments. Default is 2.
21434 Set the number of output video streams, that is also the number of video
21435 streams in each segment. Default is 1.
21438 Set the number of output audio streams, that is also the number of audio
21439 streams in each segment. Default is 0.
21442 Activate unsafe mode: do not fail if segments have a different format.
21446 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
21447 @var{a} audio outputs.
21449 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
21450 segment, in the same order as the outputs, then the inputs for the second
21453 Related streams do not always have exactly the same duration, for various
21454 reasons including codec frame size or sloppy authoring. For that reason,
21455 related synchronized streams (e.g. a video and its audio track) should be
21456 concatenated at once. The concat filter will use the duration of the longest
21457 stream in each segment (except the last one), and if necessary pad shorter
21458 audio streams with silence.
21460 For this filter to work correctly, all segments must start at timestamp 0.
21462 All corresponding streams must have the same parameters in all segments; the
21463 filtering system will automatically select a common pixel format for video
21464 streams, and a common sample format, sample rate and channel layout for
21465 audio streams, but other settings, such as resolution, must be converted
21466 explicitly by the user.
21468 Different frame rates are acceptable but will result in variable frame rate
21469 at output; be sure to configure the output file to handle it.
21471 @subsection Examples
21475 Concatenate an opening, an episode and an ending, all in bilingual version
21476 (video in stream 0, audio in streams 1 and 2):
21478 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
21479 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
21480 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
21481 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
21485 Concatenate two parts, handling audio and video separately, using the
21486 (a)movie sources, and adjusting the resolution:
21488 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
21489 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
21490 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
21492 Note that a desync will happen at the stitch if the audio and video streams
21493 do not have exactly the same duration in the first file.
21497 @subsection Commands
21499 This filter supports the following commands:
21502 Close the current segment and step to the next one
21505 @section drawgraph, adrawgraph
21507 Draw a graph using input video or audio metadata.
21509 It accepts the following parameters:
21513 Set 1st frame metadata key from which metadata values will be used to draw a graph.
21516 Set 1st foreground color expression.
21519 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
21522 Set 2nd foreground color expression.
21525 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
21528 Set 3rd foreground color expression.
21531 Set 4th frame metadata key from which metadata values will be used to draw a graph.
21534 Set 4th foreground color expression.
21537 Set minimal value of metadata value.
21540 Set maximal value of metadata value.
21543 Set graph background color. Default is white.
21548 Available values for mode is:
21555 Default is @code{line}.
21560 Available values for slide is:
21563 Draw new frame when right border is reached.
21566 Replace old columns with new ones.
21569 Scroll from right to left.
21572 Scroll from left to right.
21575 Draw single picture.
21578 Default is @code{frame}.
21581 Set size of graph video. For the syntax of this option, check the
21582 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21583 The default value is @code{900x256}.
21585 The foreground color expressions can use the following variables:
21588 Minimal value of metadata value.
21591 Maximal value of metadata value.
21594 Current metadata key value.
21597 The color is defined as 0xAABBGGRR.
21600 Example using metadata from @ref{signalstats} filter:
21602 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
21605 Example using metadata from @ref{ebur128} filter:
21607 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
21613 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
21614 level. By default, it logs a message at a frequency of 10Hz with the
21615 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
21616 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
21618 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
21619 sample format is double-precision floating point. The input stream will be converted to
21620 this specification, if needed. Users may need to insert aformat and/or aresample filters
21621 after this filter to obtain the original parameters.
21623 The filter also has a video output (see the @var{video} option) with a real
21624 time graph to observe the loudness evolution. The graphic contains the logged
21625 message mentioned above, so it is not printed anymore when this option is set,
21626 unless the verbose logging is set. The main graphing area contains the
21627 short-term loudness (3 seconds of analysis), and the gauge on the right is for
21628 the momentary loudness (400 milliseconds), but can optionally be configured
21629 to instead display short-term loudness (see @var{gauge}).
21631 The green area marks a +/- 1LU target range around the target loudness
21632 (-23LUFS by default, unless modified through @var{target}).
21634 More information about the Loudness Recommendation EBU R128 on
21635 @url{http://tech.ebu.ch/loudness}.
21637 The filter accepts the following options:
21642 Activate the video output. The audio stream is passed unchanged whether this
21643 option is set or no. The video stream will be the first output stream if
21644 activated. Default is @code{0}.
21647 Set the video size. This option is for video only. For the syntax of this
21649 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21650 Default and minimum resolution is @code{640x480}.
21653 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
21654 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
21655 other integer value between this range is allowed.
21658 Set metadata injection. If set to @code{1}, the audio input will be segmented
21659 into 100ms output frames, each of them containing various loudness information
21660 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
21662 Default is @code{0}.
21665 Force the frame logging level.
21667 Available values are:
21670 information logging level
21672 verbose logging level
21675 By default, the logging level is set to @var{info}. If the @option{video} or
21676 the @option{metadata} options are set, it switches to @var{verbose}.
21681 Available modes can be cumulated (the option is a @code{flag} type). Possible
21685 Disable any peak mode (default).
21687 Enable sample-peak mode.
21689 Simple peak mode looking for the higher sample value. It logs a message
21690 for sample-peak (identified by @code{SPK}).
21692 Enable true-peak mode.
21694 If enabled, the peak lookup is done on an over-sampled version of the input
21695 stream for better peak accuracy. It logs a message for true-peak.
21696 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
21697 This mode requires a build with @code{libswresample}.
21701 Treat mono input files as "dual mono". If a mono file is intended for playback
21702 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
21703 If set to @code{true}, this option will compensate for this effect.
21704 Multi-channel input files are not affected by this option.
21707 Set a specific pan law to be used for the measurement of dual mono files.
21708 This parameter is optional, and has a default value of -3.01dB.
21711 Set a specific target level (in LUFS) used as relative zero in the visualization.
21712 This parameter is optional and has a default value of -23LUFS as specified
21713 by EBU R128. However, material published online may prefer a level of -16LUFS
21714 (e.g. for use with podcasts or video platforms).
21717 Set the value displayed by the gauge. Valid values are @code{momentary} and s
21718 @code{shortterm}. By default the momentary value will be used, but in certain
21719 scenarios it may be more useful to observe the short term value instead (e.g.
21723 Sets the display scale for the loudness. Valid parameters are @code{absolute}
21724 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
21725 video output, not the summary or continuous log output.
21728 @subsection Examples
21732 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
21734 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
21738 Run an analysis with @command{ffmpeg}:
21740 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
21744 @section interleave, ainterleave
21746 Temporally interleave frames from several inputs.
21748 @code{interleave} works with video inputs, @code{ainterleave} with audio.
21750 These filters read frames from several inputs and send the oldest
21751 queued frame to the output.
21753 Input streams must have well defined, monotonically increasing frame
21756 In order to submit one frame to output, these filters need to enqueue
21757 at least one frame for each input, so they cannot work in case one
21758 input is not yet terminated and will not receive incoming frames.
21760 For example consider the case when one input is a @code{select} filter
21761 which always drops input frames. The @code{interleave} filter will keep
21762 reading from that input, but it will never be able to send new frames
21763 to output until the input sends an end-of-stream signal.
21765 Also, depending on inputs synchronization, the filters will drop
21766 frames in case one input receives more frames than the other ones, and
21767 the queue is already filled.
21769 These filters accept the following options:
21773 Set the number of different inputs, it is 2 by default.
21776 @subsection Examples
21780 Interleave frames belonging to different streams using @command{ffmpeg}:
21782 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
21786 Add flickering blur effect:
21788 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
21792 @section metadata, ametadata
21794 Manipulate frame metadata.
21796 This filter accepts the following options:
21800 Set mode of operation of the filter.
21802 Can be one of the following:
21806 If both @code{value} and @code{key} is set, select frames
21807 which have such metadata. If only @code{key} is set, select
21808 every frame that has such key in metadata.
21811 Add new metadata @code{key} and @code{value}. If key is already available
21815 Modify value of already present key.
21818 If @code{value} is set, delete only keys that have such value.
21819 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
21823 Print key and its value if metadata was found. If @code{key} is not set print all
21824 metadata values available in frame.
21828 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
21831 Set metadata value which will be used. This option is mandatory for
21832 @code{modify} and @code{add} mode.
21835 Which function to use when comparing metadata value and @code{value}.
21837 Can be one of following:
21841 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
21844 Values are interpreted as strings, returns true if metadata value starts with
21845 the @code{value} option string.
21848 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
21851 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
21854 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
21857 Values are interpreted as floats, returns true if expression from option @code{expr}
21861 Values are interpreted as strings, returns true if metadata value ends with
21862 the @code{value} option string.
21866 Set expression which is used when @code{function} is set to @code{expr}.
21867 The expression is evaluated through the eval API and can contain the following
21872 Float representation of @code{value} from metadata key.
21875 Float representation of @code{value} as supplied by user in @code{value} option.
21879 If specified in @code{print} mode, output is written to the named file. Instead of
21880 plain filename any writable url can be specified. Filename ``-'' is a shorthand
21881 for standard output. If @code{file} option is not set, output is written to the log
21882 with AV_LOG_INFO loglevel.
21886 @subsection Examples
21890 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
21893 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
21896 Print silencedetect output to file @file{metadata.txt}.
21898 silencedetect,ametadata=mode=print:file=metadata.txt
21901 Direct all metadata to a pipe with file descriptor 4.
21903 metadata=mode=print:file='pipe\:4'
21907 @section perms, aperms
21909 Set read/write permissions for the output frames.
21911 These filters are mainly aimed at developers to test direct path in the
21912 following filter in the filtergraph.
21914 The filters accept the following options:
21918 Select the permissions mode.
21920 It accepts the following values:
21923 Do nothing. This is the default.
21925 Set all the output frames read-only.
21927 Set all the output frames directly writable.
21929 Make the frame read-only if writable, and writable if read-only.
21931 Set each output frame read-only or writable randomly.
21935 Set the seed for the @var{random} mode, must be an integer included between
21936 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
21937 @code{-1}, the filter will try to use a good random seed on a best effort
21941 Note: in case of auto-inserted filter between the permission filter and the
21942 following one, the permission might not be received as expected in that
21943 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
21944 perms/aperms filter can avoid this problem.
21946 @section realtime, arealtime
21948 Slow down filtering to match real time approximately.
21950 These filters will pause the filtering for a variable amount of time to
21951 match the output rate with the input timestamps.
21952 They are similar to the @option{re} option to @code{ffmpeg}.
21954 They accept the following options:
21958 Time limit for the pauses. Any pause longer than that will be considered
21959 a timestamp discontinuity and reset the timer. Default is 2 seconds.
21961 Speed factor for processing. The value must be a float larger than zero.
21962 Values larger than 1.0 will result in faster than realtime processing,
21963 smaller will slow processing down. The @var{limit} is automatically adapted
21964 accordingly. Default is 1.0.
21966 A processing speed faster than what is possible without these filters cannot
21971 @section select, aselect
21973 Select frames to pass in output.
21975 This filter accepts the following options:
21980 Set expression, which is evaluated for each input frame.
21982 If the expression is evaluated to zero, the frame is discarded.
21984 If the evaluation result is negative or NaN, the frame is sent to the
21985 first output; otherwise it is sent to the output with index
21986 @code{ceil(val)-1}, assuming that the input index starts from 0.
21988 For example a value of @code{1.2} corresponds to the output with index
21989 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
21992 Set the number of outputs. The output to which to send the selected
21993 frame is based on the result of the evaluation. Default value is 1.
21996 The expression can contain the following constants:
22000 The (sequential) number of the filtered frame, starting from 0.
22003 The (sequential) number of the selected frame, starting from 0.
22005 @item prev_selected_n
22006 The sequential number of the last selected frame. It's NAN if undefined.
22009 The timebase of the input timestamps.
22012 The PTS (Presentation TimeStamp) of the filtered video frame,
22013 expressed in @var{TB} units. It's NAN if undefined.
22016 The PTS of the filtered video frame,
22017 expressed in seconds. It's NAN if undefined.
22020 The PTS of the previously filtered video frame. It's NAN if undefined.
22022 @item prev_selected_pts
22023 The PTS of the last previously filtered video frame. It's NAN if undefined.
22025 @item prev_selected_t
22026 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
22029 The PTS of the first video frame in the video. It's NAN if undefined.
22032 The time of the first video frame in the video. It's NAN if undefined.
22034 @item pict_type @emph{(video only)}
22035 The type of the filtered frame. It can assume one of the following
22047 @item interlace_type @emph{(video only)}
22048 The frame interlace type. It can assume one of the following values:
22051 The frame is progressive (not interlaced).
22053 The frame is top-field-first.
22055 The frame is bottom-field-first.
22058 @item consumed_sample_n @emph{(audio only)}
22059 the number of selected samples before the current frame
22061 @item samples_n @emph{(audio only)}
22062 the number of samples in the current frame
22064 @item sample_rate @emph{(audio only)}
22065 the input sample rate
22068 This is 1 if the filtered frame is a key-frame, 0 otherwise.
22071 the position in the file of the filtered frame, -1 if the information
22072 is not available (e.g. for synthetic video)
22074 @item scene @emph{(video only)}
22075 value between 0 and 1 to indicate a new scene; a low value reflects a low
22076 probability for the current frame to introduce a new scene, while a higher
22077 value means the current frame is more likely to be one (see the example below)
22079 @item concatdec_select
22080 The concat demuxer can select only part of a concat input file by setting an
22081 inpoint and an outpoint, but the output packets may not be entirely contained
22082 in the selected interval. By using this variable, it is possible to skip frames
22083 generated by the concat demuxer which are not exactly contained in the selected
22086 This works by comparing the frame pts against the @var{lavf.concat.start_time}
22087 and the @var{lavf.concat.duration} packet metadata values which are also
22088 present in the decoded frames.
22090 The @var{concatdec_select} variable is -1 if the frame pts is at least
22091 start_time and either the duration metadata is missing or the frame pts is less
22092 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
22095 That basically means that an input frame is selected if its pts is within the
22096 interval set by the concat demuxer.
22100 The default value of the select expression is "1".
22102 @subsection Examples
22106 Select all frames in input:
22111 The example above is the same as:
22123 Select only I-frames:
22125 select='eq(pict_type\,I)'
22129 Select one frame every 100:
22131 select='not(mod(n\,100))'
22135 Select only frames contained in the 10-20 time interval:
22137 select=between(t\,10\,20)
22141 Select only I-frames contained in the 10-20 time interval:
22143 select=between(t\,10\,20)*eq(pict_type\,I)
22147 Select frames with a minimum distance of 10 seconds:
22149 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
22153 Use aselect to select only audio frames with samples number > 100:
22155 aselect='gt(samples_n\,100)'
22159 Create a mosaic of the first scenes:
22161 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
22164 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
22168 Send even and odd frames to separate outputs, and compose them:
22170 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
22174 Select useful frames from an ffconcat file which is using inpoints and
22175 outpoints but where the source files are not intra frame only.
22177 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
22181 @section sendcmd, asendcmd
22183 Send commands to filters in the filtergraph.
22185 These filters read commands to be sent to other filters in the
22188 @code{sendcmd} must be inserted between two video filters,
22189 @code{asendcmd} must be inserted between two audio filters, but apart
22190 from that they act the same way.
22192 The specification of commands can be provided in the filter arguments
22193 with the @var{commands} option, or in a file specified by the
22194 @var{filename} option.
22196 These filters accept the following options:
22199 Set the commands to be read and sent to the other filters.
22201 Set the filename of the commands to be read and sent to the other
22205 @subsection Commands syntax
22207 A commands description consists of a sequence of interval
22208 specifications, comprising a list of commands to be executed when a
22209 particular event related to that interval occurs. The occurring event
22210 is typically the current frame time entering or leaving a given time
22213 An interval is specified by the following syntax:
22215 @var{START}[-@var{END}] @var{COMMANDS};
22218 The time interval is specified by the @var{START} and @var{END} times.
22219 @var{END} is optional and defaults to the maximum time.
22221 The current frame time is considered within the specified interval if
22222 it is included in the interval [@var{START}, @var{END}), that is when
22223 the time is greater or equal to @var{START} and is lesser than
22226 @var{COMMANDS} consists of a sequence of one or more command
22227 specifications, separated by ",", relating to that interval. The
22228 syntax of a command specification is given by:
22230 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
22233 @var{FLAGS} is optional and specifies the type of events relating to
22234 the time interval which enable sending the specified command, and must
22235 be a non-null sequence of identifier flags separated by "+" or "|" and
22236 enclosed between "[" and "]".
22238 The following flags are recognized:
22241 The command is sent when the current frame timestamp enters the
22242 specified interval. In other words, the command is sent when the
22243 previous frame timestamp was not in the given interval, and the
22247 The command is sent when the current frame timestamp leaves the
22248 specified interval. In other words, the command is sent when the
22249 previous frame timestamp was in the given interval, and the
22253 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
22256 @var{TARGET} specifies the target of the command, usually the name of
22257 the filter class or a specific filter instance name.
22259 @var{COMMAND} specifies the name of the command for the target filter.
22261 @var{ARG} is optional and specifies the optional list of argument for
22262 the given @var{COMMAND}.
22264 Between one interval specification and another, whitespaces, or
22265 sequences of characters starting with @code{#} until the end of line,
22266 are ignored and can be used to annotate comments.
22268 A simplified BNF description of the commands specification syntax
22271 @var{COMMAND_FLAG} ::= "enter" | "leave"
22272 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
22273 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
22274 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
22275 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
22276 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
22279 @subsection Examples
22283 Specify audio tempo change at second 4:
22285 asendcmd=c='4.0 atempo tempo 1.5',atempo
22289 Target a specific filter instance:
22291 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
22295 Specify a list of drawtext and hue commands in a file.
22297 # show text in the interval 5-10
22298 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
22299 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
22301 # desaturate the image in the interval 15-20
22302 15.0-20.0 [enter] hue s 0,
22303 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
22305 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
22307 # apply an exponential saturation fade-out effect, starting from time 25
22308 25 [enter] hue s exp(25-t)
22311 A filtergraph allowing to read and process the above command list
22312 stored in a file @file{test.cmd}, can be specified with:
22314 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
22319 @section setpts, asetpts
22321 Change the PTS (presentation timestamp) of the input frames.
22323 @code{setpts} works on video frames, @code{asetpts} on audio frames.
22325 This filter accepts the following options:
22330 The expression which is evaluated for each frame to construct its timestamp.
22334 The expression is evaluated through the eval API and can contain the following
22338 @item FRAME_RATE, FR
22339 frame rate, only defined for constant frame-rate video
22342 The presentation timestamp in input
22345 The count of the input frame for video or the number of consumed samples,
22346 not including the current frame for audio, starting from 0.
22348 @item NB_CONSUMED_SAMPLES
22349 The number of consumed samples, not including the current frame (only
22352 @item NB_SAMPLES, S
22353 The number of samples in the current frame (only audio)
22355 @item SAMPLE_RATE, SR
22356 The audio sample rate.
22359 The PTS of the first frame.
22362 the time in seconds of the first frame
22365 State whether the current frame is interlaced.
22368 the time in seconds of the current frame
22371 original position in the file of the frame, or undefined if undefined
22372 for the current frame
22375 The previous input PTS.
22378 previous input time in seconds
22381 The previous output PTS.
22384 previous output time in seconds
22387 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
22391 The wallclock (RTC) time at the start of the movie in microseconds.
22394 The timebase of the input timestamps.
22398 @subsection Examples
22402 Start counting PTS from zero
22404 setpts=PTS-STARTPTS
22408 Apply fast motion effect:
22414 Apply slow motion effect:
22420 Set fixed rate of 25 frames per second:
22426 Set fixed rate 25 fps with some jitter:
22428 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
22432 Apply an offset of 10 seconds to the input PTS:
22438 Generate timestamps from a "live source" and rebase onto the current timebase:
22440 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
22444 Generate timestamps by counting samples:
22453 Force color range for the output video frame.
22455 The @code{setrange} filter marks the color range property for the
22456 output frames. It does not change the input frame, but only sets the
22457 corresponding property, which affects how the frame is treated by
22460 The filter accepts the following options:
22465 Available values are:
22469 Keep the same color range property.
22471 @item unspecified, unknown
22472 Set the color range as unspecified.
22474 @item limited, tv, mpeg
22475 Set the color range as limited.
22477 @item full, pc, jpeg
22478 Set the color range as full.
22482 @section settb, asettb
22484 Set the timebase to use for the output frames timestamps.
22485 It is mainly useful for testing timebase configuration.
22487 It accepts the following parameters:
22492 The expression which is evaluated into the output timebase.
22496 The value for @option{tb} is an arithmetic expression representing a
22497 rational. The expression can contain the constants "AVTB" (the default
22498 timebase), "intb" (the input timebase) and "sr" (the sample rate,
22499 audio only). Default value is "intb".
22501 @subsection Examples
22505 Set the timebase to 1/25:
22511 Set the timebase to 1/10:
22517 Set the timebase to 1001/1000:
22523 Set the timebase to 2*intb:
22529 Set the default timebase value:
22536 Convert input audio to a video output representing frequency spectrum
22537 logarithmically using Brown-Puckette constant Q transform algorithm with
22538 direct frequency domain coefficient calculation (but the transform itself
22539 is not really constant Q, instead the Q factor is actually variable/clamped),
22540 with musical tone scale, from E0 to D#10.
22542 The filter accepts the following options:
22546 Specify the video size for the output. It must be even. For the syntax of this option,
22547 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22548 Default value is @code{1920x1080}.
22551 Set the output frame rate. Default value is @code{25}.
22554 Set the bargraph height. It must be even. Default value is @code{-1} which
22555 computes the bargraph height automatically.
22558 Set the axis height. It must be even. Default value is @code{-1} which computes
22559 the axis height automatically.
22562 Set the sonogram height. It must be even. Default value is @code{-1} which
22563 computes the sonogram height automatically.
22566 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
22567 instead. Default value is @code{1}.
22569 @item sono_v, volume
22570 Specify the sonogram volume expression. It can contain variables:
22573 the @var{bar_v} evaluated expression
22574 @item frequency, freq, f
22575 the frequency where it is evaluated
22576 @item timeclamp, tc
22577 the value of @var{timeclamp} option
22581 @item a_weighting(f)
22582 A-weighting of equal loudness
22583 @item b_weighting(f)
22584 B-weighting of equal loudness
22585 @item c_weighting(f)
22586 C-weighting of equal loudness.
22588 Default value is @code{16}.
22590 @item bar_v, volume2
22591 Specify the bargraph volume expression. It can contain variables:
22594 the @var{sono_v} evaluated expression
22595 @item frequency, freq, f
22596 the frequency where it is evaluated
22597 @item timeclamp, tc
22598 the value of @var{timeclamp} option
22602 @item a_weighting(f)
22603 A-weighting of equal loudness
22604 @item b_weighting(f)
22605 B-weighting of equal loudness
22606 @item c_weighting(f)
22607 C-weighting of equal loudness.
22609 Default value is @code{sono_v}.
22611 @item sono_g, gamma
22612 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
22613 higher gamma makes the spectrum having more range. Default value is @code{3}.
22614 Acceptable range is @code{[1, 7]}.
22616 @item bar_g, gamma2
22617 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
22621 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
22622 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
22624 @item timeclamp, tc
22625 Specify the transform timeclamp. At low frequency, there is trade-off between
22626 accuracy in time domain and frequency domain. If timeclamp is lower,
22627 event in time domain is represented more accurately (such as fast bass drum),
22628 otherwise event in frequency domain is represented more accurately
22629 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
22632 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
22633 limits future samples by applying asymmetric windowing in time domain, useful
22634 when low latency is required. Accepted range is @code{[0, 1]}.
22637 Specify the transform base frequency. Default value is @code{20.01523126408007475},
22638 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
22641 Specify the transform end frequency. Default value is @code{20495.59681441799654},
22642 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
22645 This option is deprecated and ignored.
22648 Specify the transform length in time domain. Use this option to control accuracy
22649 trade-off between time domain and frequency domain at every frequency sample.
22650 It can contain variables:
22652 @item frequency, freq, f
22653 the frequency where it is evaluated
22654 @item timeclamp, tc
22655 the value of @var{timeclamp} option.
22657 Default value is @code{384*tc/(384+tc*f)}.
22660 Specify the transform count for every video frame. Default value is @code{6}.
22661 Acceptable range is @code{[1, 30]}.
22664 Specify the transform count for every single pixel. Default value is @code{0},
22665 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
22668 Specify font file for use with freetype to draw the axis. If not specified,
22669 use embedded font. Note that drawing with font file or embedded font is not
22670 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
22674 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
22675 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
22679 Specify font color expression. This is arithmetic expression that should return
22680 integer value 0xRRGGBB. It can contain variables:
22682 @item frequency, freq, f
22683 the frequency where it is evaluated
22684 @item timeclamp, tc
22685 the value of @var{timeclamp} option
22690 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
22691 @item r(x), g(x), b(x)
22692 red, green, and blue value of intensity x.
22694 Default value is @code{st(0, (midi(f)-59.5)/12);
22695 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
22696 r(1-ld(1)) + b(ld(1))}.
22699 Specify image file to draw the axis. This option override @var{fontfile} and
22700 @var{fontcolor} option.
22703 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
22704 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
22705 Default value is @code{1}.
22708 Set colorspace. The accepted values are:
22711 Unspecified (default)
22720 BT.470BG or BT.601-6 625
22723 SMPTE-170M or BT.601-6 525
22729 BT.2020 with non-constant luminance
22734 Set spectrogram color scheme. This is list of floating point values with format
22735 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
22736 The default is @code{1|0.5|0|0|0.5|1}.
22740 @subsection Examples
22744 Playing audio while showing the spectrum:
22746 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
22750 Same as above, but with frame rate 30 fps:
22752 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
22756 Playing at 1280x720:
22758 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
22762 Disable sonogram display:
22768 A1 and its harmonics: A1, A2, (near)E3, A3:
22770 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),
22771 asplit[a][out1]; [a] showcqt [out0]'
22775 Same as above, but with more accuracy in frequency domain:
22777 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),
22778 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
22784 bar_v=10:sono_v=bar_v*a_weighting(f)
22788 Custom gamma, now spectrum is linear to the amplitude.
22794 Custom tlength equation:
22796 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)))'
22800 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
22802 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
22806 Custom font using fontconfig:
22808 font='Courier New,Monospace,mono|bold'
22812 Custom frequency range with custom axis using image file:
22814 axisfile=myaxis.png:basefreq=40:endfreq=10000
22820 Convert input audio to video output representing the audio power spectrum.
22821 Audio amplitude is on Y-axis while frequency is on X-axis.
22823 The filter accepts the following options:
22827 Specify size of video. For the syntax of this option, check the
22828 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22829 Default is @code{1024x512}.
22833 This set how each frequency bin will be represented.
22835 It accepts the following values:
22841 Default is @code{bar}.
22844 Set amplitude scale.
22846 It accepts the following values:
22860 Default is @code{log}.
22863 Set frequency scale.
22865 It accepts the following values:
22874 Reverse logarithmic scale.
22876 Default is @code{lin}.
22879 Set window size. Allowed range is from 16 to 65536.
22881 Default is @code{2048}
22884 Set windowing function.
22886 It accepts the following values:
22909 Default is @code{hanning}.
22912 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
22913 which means optimal overlap for selected window function will be picked.
22916 Set time averaging. Setting this to 0 will display current maximal peaks.
22917 Default is @code{1}, which means time averaging is disabled.
22920 Specify list of colors separated by space or by '|' which will be used to
22921 draw channel frequencies. Unrecognized or missing colors will be replaced
22925 Set channel display mode.
22927 It accepts the following values:
22932 Default is @code{combined}.
22935 Set minimum amplitude used in @code{log} amplitude scaler.
22939 @section showspatial
22941 Convert stereo input audio to a video output, representing the spatial relationship
22942 between two channels.
22944 The filter accepts the following options:
22948 Specify the video size for the output. For the syntax of this option, check the
22949 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22950 Default value is @code{512x512}.
22953 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
22956 Set window function.
22958 It accepts the following values:
22983 Default value is @code{hann}.
22986 Set ratio of overlap window. Default value is @code{0.5}.
22987 When value is @code{1} overlap is set to recommended size for specific
22988 window function currently used.
22991 @anchor{showspectrum}
22992 @section showspectrum
22994 Convert input audio to a video output, representing the audio frequency
22997 The filter accepts the following options:
23001 Specify the video size for the output. For the syntax of this option, check the
23002 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23003 Default value is @code{640x512}.
23006 Specify how the spectrum should slide along the window.
23008 It accepts the following values:
23011 the samples start again on the left when they reach the right
23013 the samples scroll from right to left
23015 frames are only produced when the samples reach the right
23017 the samples scroll from left to right
23020 Default value is @code{replace}.
23023 Specify display mode.
23025 It accepts the following values:
23028 all channels are displayed in the same row
23030 all channels are displayed in separate rows
23033 Default value is @samp{combined}.
23036 Specify display color mode.
23038 It accepts the following values:
23041 each channel is displayed in a separate color
23043 each channel is displayed using the same color scheme
23045 each channel is displayed using the rainbow color scheme
23047 each channel is displayed using the moreland color scheme
23049 each channel is displayed using the nebulae color scheme
23051 each channel is displayed using the fire color scheme
23053 each channel is displayed using the fiery color scheme
23055 each channel is displayed using the fruit color scheme
23057 each channel is displayed using the cool color scheme
23059 each channel is displayed using the magma color scheme
23061 each channel is displayed using the green color scheme
23063 each channel is displayed using the viridis color scheme
23065 each channel is displayed using the plasma color scheme
23067 each channel is displayed using the cividis color scheme
23069 each channel is displayed using the terrain color scheme
23072 Default value is @samp{channel}.
23075 Specify scale used for calculating intensity color values.
23077 It accepts the following values:
23082 square root, default
23093 Default value is @samp{sqrt}.
23096 Specify frequency scale.
23098 It accepts the following values:
23106 Default value is @samp{lin}.
23109 Set saturation modifier for displayed colors. Negative values provide
23110 alternative color scheme. @code{0} is no saturation at all.
23111 Saturation must be in [-10.0, 10.0] range.
23112 Default value is @code{1}.
23115 Set window function.
23117 It accepts the following values:
23142 Default value is @code{hann}.
23145 Set orientation of time vs frequency axis. Can be @code{vertical} or
23146 @code{horizontal}. Default is @code{vertical}.
23149 Set ratio of overlap window. Default value is @code{0}.
23150 When value is @code{1} overlap is set to recommended size for specific
23151 window function currently used.
23154 Set scale gain for calculating intensity color values.
23155 Default value is @code{1}.
23158 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
23161 Set color rotation, must be in [-1.0, 1.0] range.
23162 Default value is @code{0}.
23165 Set start frequency from which to display spectrogram. Default is @code{0}.
23168 Set stop frequency to which to display spectrogram. Default is @code{0}.
23171 Set upper frame rate limit. Default is @code{auto}, unlimited.
23174 Draw time and frequency axes and legends. Default is disabled.
23177 The usage is very similar to the showwaves filter; see the examples in that
23180 @subsection Examples
23184 Large window with logarithmic color scaling:
23186 showspectrum=s=1280x480:scale=log
23190 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
23192 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
23193 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
23197 @section showspectrumpic
23199 Convert input audio to a single video frame, representing the audio frequency
23202 The filter accepts the following options:
23206 Specify the video size for the output. For the syntax of this option, check the
23207 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23208 Default value is @code{4096x2048}.
23211 Specify display mode.
23213 It accepts the following values:
23216 all channels are displayed in the same row
23218 all channels are displayed in separate rows
23220 Default value is @samp{combined}.
23223 Specify display color mode.
23225 It accepts the following values:
23228 each channel is displayed in a separate color
23230 each channel is displayed using the same color scheme
23232 each channel is displayed using the rainbow color scheme
23234 each channel is displayed using the moreland color scheme
23236 each channel is displayed using the nebulae color scheme
23238 each channel is displayed using the fire color scheme
23240 each channel is displayed using the fiery color scheme
23242 each channel is displayed using the fruit color scheme
23244 each channel is displayed using the cool color scheme
23246 each channel is displayed using the magma color scheme
23248 each channel is displayed using the green color scheme
23250 each channel is displayed using the viridis color scheme
23252 each channel is displayed using the plasma color scheme
23254 each channel is displayed using the cividis color scheme
23256 each channel is displayed using the terrain color scheme
23258 Default value is @samp{intensity}.
23261 Specify scale used for calculating intensity color values.
23263 It accepts the following values:
23268 square root, default
23278 Default value is @samp{log}.
23281 Specify frequency scale.
23283 It accepts the following values:
23291 Default value is @samp{lin}.
23294 Set saturation modifier for displayed colors. Negative values provide
23295 alternative color scheme. @code{0} is no saturation at all.
23296 Saturation must be in [-10.0, 10.0] range.
23297 Default value is @code{1}.
23300 Set window function.
23302 It accepts the following values:
23326 Default value is @code{hann}.
23329 Set orientation of time vs frequency axis. Can be @code{vertical} or
23330 @code{horizontal}. Default is @code{vertical}.
23333 Set scale gain for calculating intensity color values.
23334 Default value is @code{1}.
23337 Draw time and frequency axes and legends. Default is enabled.
23340 Set color rotation, must be in [-1.0, 1.0] range.
23341 Default value is @code{0}.
23344 Set start frequency from which to display spectrogram. Default is @code{0}.
23347 Set stop frequency to which to display spectrogram. Default is @code{0}.
23350 @subsection Examples
23354 Extract an audio spectrogram of a whole audio track
23355 in a 1024x1024 picture using @command{ffmpeg}:
23357 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
23361 @section showvolume
23363 Convert input audio volume to a video output.
23365 The filter accepts the following options:
23372 Set border width, allowed range is [0, 5]. Default is 1.
23375 Set channel width, allowed range is [80, 8192]. Default is 400.
23378 Set channel height, allowed range is [1, 900]. Default is 20.
23381 Set fade, allowed range is [0, 1]. Default is 0.95.
23384 Set volume color expression.
23386 The expression can use the following variables:
23390 Current max volume of channel in dB.
23396 Current channel number, starting from 0.
23400 If set, displays channel names. Default is enabled.
23403 If set, displays volume values. Default is enabled.
23406 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
23407 default is @code{h}.
23410 Set step size, allowed range is [0, 5]. Default is 0, which means
23414 Set background opacity, allowed range is [0, 1]. Default is 0.
23417 Set metering mode, can be peak: @code{p} or rms: @code{r},
23418 default is @code{p}.
23421 Set display scale, can be linear: @code{lin} or log: @code{log},
23422 default is @code{lin}.
23426 If set to > 0., display a line for the max level
23427 in the previous seconds.
23428 default is disabled: @code{0.}
23431 The color of the max line. Use when @code{dm} option is set to > 0.
23432 default is: @code{orange}
23437 Convert input audio to a video output, representing the samples waves.
23439 The filter accepts the following options:
23443 Specify the video size for the output. For the syntax of this option, check the
23444 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23445 Default value is @code{600x240}.
23450 Available values are:
23453 Draw a point for each sample.
23456 Draw a vertical line for each sample.
23459 Draw a point for each sample and a line between them.
23462 Draw a centered vertical line for each sample.
23465 Default value is @code{point}.
23468 Set the number of samples which are printed on the same column. A
23469 larger value will decrease the frame rate. Must be a positive
23470 integer. This option can be set only if the value for @var{rate}
23471 is not explicitly specified.
23474 Set the (approximate) output frame rate. This is done by setting the
23475 option @var{n}. Default value is "25".
23477 @item split_channels
23478 Set if channels should be drawn separately or overlap. Default value is 0.
23481 Set colors separated by '|' which are going to be used for drawing of each channel.
23484 Set amplitude scale.
23486 Available values are:
23504 Set the draw mode. This is mostly useful to set for high @var{n}.
23506 Available values are:
23509 Scale pixel values for each drawn sample.
23512 Draw every sample directly.
23515 Default value is @code{scale}.
23518 @subsection Examples
23522 Output the input file audio and the corresponding video representation
23525 amovie=a.mp3,asplit[out0],showwaves[out1]
23529 Create a synthetic signal and show it with showwaves, forcing a
23530 frame rate of 30 frames per second:
23532 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
23536 @section showwavespic
23538 Convert input audio to a single video frame, representing the samples waves.
23540 The filter accepts the following options:
23544 Specify the video size for the output. For the syntax of this option, check the
23545 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23546 Default value is @code{600x240}.
23548 @item split_channels
23549 Set if channels should be drawn separately or overlap. Default value is 0.
23552 Set colors separated by '|' which are going to be used for drawing of each channel.
23555 Set amplitude scale.
23557 Available values are:
23577 Available values are:
23580 Scale pixel values for each drawn sample.
23583 Draw every sample directly.
23586 Default value is @code{scale}.
23589 @subsection Examples
23593 Extract a channel split representation of the wave form of a whole audio track
23594 in a 1024x800 picture using @command{ffmpeg}:
23596 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
23600 @section sidedata, asidedata
23602 Delete frame side data, or select frames based on it.
23604 This filter accepts the following options:
23608 Set mode of operation of the filter.
23610 Can be one of the following:
23614 Select every frame with side data of @code{type}.
23617 Delete side data of @code{type}. If @code{type} is not set, delete all side
23623 Set side data type used with all modes. Must be set for @code{select} mode. For
23624 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
23625 in @file{libavutil/frame.h}. For example, to choose
23626 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
23630 @section spectrumsynth
23632 Sythesize audio from 2 input video spectrums, first input stream represents
23633 magnitude across time and second represents phase across time.
23634 The filter will transform from frequency domain as displayed in videos back
23635 to time domain as presented in audio output.
23637 This filter is primarily created for reversing processed @ref{showspectrum}
23638 filter outputs, but can synthesize sound from other spectrograms too.
23639 But in such case results are going to be poor if the phase data is not
23640 available, because in such cases phase data need to be recreated, usually
23641 it's just recreated from random noise.
23642 For best results use gray only output (@code{channel} color mode in
23643 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
23644 @code{lin} scale for phase video. To produce phase, for 2nd video, use
23645 @code{data} option. Inputs videos should generally use @code{fullframe}
23646 slide mode as that saves resources needed for decoding video.
23648 The filter accepts the following options:
23652 Specify sample rate of output audio, the sample rate of audio from which
23653 spectrum was generated may differ.
23656 Set number of channels represented in input video spectrums.
23659 Set scale which was used when generating magnitude input spectrum.
23660 Can be @code{lin} or @code{log}. Default is @code{log}.
23663 Set slide which was used when generating inputs spectrums.
23664 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
23665 Default is @code{fullframe}.
23668 Set window function used for resynthesis.
23671 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
23672 which means optimal overlap for selected window function will be picked.
23675 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
23676 Default is @code{vertical}.
23679 @subsection Examples
23683 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
23684 then resynthesize videos back to audio with spectrumsynth:
23686 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
23687 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
23688 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
23692 @section split, asplit
23694 Split input into several identical outputs.
23696 @code{asplit} works with audio input, @code{split} with video.
23698 The filter accepts a single parameter which specifies the number of outputs. If
23699 unspecified, it defaults to 2.
23701 @subsection Examples
23705 Create two separate outputs from the same input:
23707 [in] split [out0][out1]
23711 To create 3 or more outputs, you need to specify the number of
23714 [in] asplit=3 [out0][out1][out2]
23718 Create two separate outputs from the same input, one cropped and
23721 [in] split [splitout1][splitout2];
23722 [splitout1] crop=100:100:0:0 [cropout];
23723 [splitout2] pad=200:200:100:100 [padout];
23727 Create 5 copies of the input audio with @command{ffmpeg}:
23729 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
23735 Receive commands sent through a libzmq client, and forward them to
23736 filters in the filtergraph.
23738 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
23739 must be inserted between two video filters, @code{azmq} between two
23740 audio filters. Both are capable to send messages to any filter type.
23742 To enable these filters you need to install the libzmq library and
23743 headers and configure FFmpeg with @code{--enable-libzmq}.
23745 For more information about libzmq see:
23746 @url{http://www.zeromq.org/}
23748 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
23749 receives messages sent through a network interface defined by the
23750 @option{bind_address} (or the abbreviation "@option{b}") option.
23751 Default value of this option is @file{tcp://localhost:5555}. You may
23752 want to alter this value to your needs, but do not forget to escape any
23753 ':' signs (see @ref{filtergraph escaping}).
23755 The received message must be in the form:
23757 @var{TARGET} @var{COMMAND} [@var{ARG}]
23760 @var{TARGET} specifies the target of the command, usually the name of
23761 the filter class or a specific filter instance name. The default
23762 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
23763 but you can override this by using the @samp{filter_name@@id} syntax
23764 (see @ref{Filtergraph syntax}).
23766 @var{COMMAND} specifies the name of the command for the target filter.
23768 @var{ARG} is optional and specifies the optional argument list for the
23769 given @var{COMMAND}.
23771 Upon reception, the message is processed and the corresponding command
23772 is injected into the filtergraph. Depending on the result, the filter
23773 will send a reply to the client, adopting the format:
23775 @var{ERROR_CODE} @var{ERROR_REASON}
23779 @var{MESSAGE} is optional.
23781 @subsection Examples
23783 Look at @file{tools/zmqsend} for an example of a zmq client which can
23784 be used to send commands processed by these filters.
23786 Consider the following filtergraph generated by @command{ffplay}.
23787 In this example the last overlay filter has an instance name. All other
23788 filters will have default instance names.
23791 ffplay -dumpgraph 1 -f lavfi "
23792 color=s=100x100:c=red [l];
23793 color=s=100x100:c=blue [r];
23794 nullsrc=s=200x100, zmq [bg];
23795 [bg][l] overlay [bg+l];
23796 [bg+l][r] overlay@@my=x=100 "
23799 To change the color of the left side of the video, the following
23800 command can be used:
23802 echo Parsed_color_0 c yellow | tools/zmqsend
23805 To change the right side:
23807 echo Parsed_color_1 c pink | tools/zmqsend
23810 To change the position of the right side:
23812 echo overlay@@my x 150 | tools/zmqsend
23816 @c man end MULTIMEDIA FILTERS
23818 @chapter Multimedia Sources
23819 @c man begin MULTIMEDIA SOURCES
23821 Below is a description of the currently available multimedia sources.
23825 This is the same as @ref{movie} source, except it selects an audio
23831 Read audio and/or video stream(s) from a movie container.
23833 It accepts the following parameters:
23837 The name of the resource to read (not necessarily a file; it can also be a
23838 device or a stream accessed through some protocol).
23840 @item format_name, f
23841 Specifies the format assumed for the movie to read, and can be either
23842 the name of a container or an input device. If not specified, the
23843 format is guessed from @var{movie_name} or by probing.
23845 @item seek_point, sp
23846 Specifies the seek point in seconds. The frames will be output
23847 starting from this seek point. The parameter is evaluated with
23848 @code{av_strtod}, so the numerical value may be suffixed by an IS
23849 postfix. The default value is "0".
23852 Specifies the streams to read. Several streams can be specified,
23853 separated by "+". The source will then have as many outputs, in the
23854 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
23855 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
23856 respectively the default (best suited) video and audio stream. Default
23857 is "dv", or "da" if the filter is called as "amovie".
23859 @item stream_index, si
23860 Specifies the index of the video stream to read. If the value is -1,
23861 the most suitable video stream will be automatically selected. The default
23862 value is "-1". Deprecated. If the filter is called "amovie", it will select
23863 audio instead of video.
23866 Specifies how many times to read the stream in sequence.
23867 If the value is 0, the stream will be looped infinitely.
23868 Default value is "1".
23870 Note that when the movie is looped the source timestamps are not
23871 changed, so it will generate non monotonically increasing timestamps.
23873 @item discontinuity
23874 Specifies the time difference between frames above which the point is
23875 considered a timestamp discontinuity which is removed by adjusting the later
23879 It allows overlaying a second video on top of the main input of
23880 a filtergraph, as shown in this graph:
23882 input -----------> deltapts0 --> overlay --> output
23885 movie --> scale--> deltapts1 -------+
23887 @subsection Examples
23891 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
23892 on top of the input labelled "in":
23894 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
23895 [in] setpts=PTS-STARTPTS [main];
23896 [main][over] overlay=16:16 [out]
23900 Read from a video4linux2 device, and overlay it on top of the input
23903 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
23904 [in] setpts=PTS-STARTPTS [main];
23905 [main][over] overlay=16:16 [out]
23909 Read the first video stream and the audio stream with id 0x81 from
23910 dvd.vob; the video is connected to the pad named "video" and the audio is
23911 connected to the pad named "audio":
23913 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
23917 @subsection Commands
23919 Both movie and amovie support the following commands:
23922 Perform seek using "av_seek_frame".
23923 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
23926 @var{stream_index}: If stream_index is -1, a default
23927 stream is selected, and @var{timestamp} is automatically converted
23928 from AV_TIME_BASE units to the stream specific time_base.
23930 @var{timestamp}: Timestamp in AVStream.time_base units
23931 or, if no stream is specified, in AV_TIME_BASE units.
23933 @var{flags}: Flags which select direction and seeking mode.
23937 Get movie duration in AV_TIME_BASE units.
23941 @c man end MULTIMEDIA SOURCES