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 sound 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})
1129 It accepts the following values:
1145 Default is @code{w4096}
1148 Set window function. Default is @code{hann}.
1151 Set window overlap. If set to 1, the recommended overlap for selected
1152 window function will be picked. Default is @code{0.75}.
1155 @subsection Examples
1159 Leave almost only low frequencies in audio:
1161 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1168 Apply an arbitrary Frequency Impulse Response filter.
1170 This filter is designed for applying long FIR filters,
1171 up to 60 seconds long.
1173 It can be used as component for digital crossover filters,
1174 room equalization, cross talk cancellation, wavefield synthesis,
1175 auralization, ambiophonics, ambisonics and spatialization.
1177 This filter uses second stream as FIR coefficients.
1178 If second stream holds single channel, it will be used
1179 for all input channels in first stream, otherwise
1180 number of channels in second stream must be same as
1181 number of channels in first stream.
1183 It accepts the following parameters:
1187 Set dry gain. This sets input gain.
1190 Set wet gain. This sets final output gain.
1193 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1196 Enable applying gain measured from power of IR.
1198 Set which approach to use for auto gain measurement.
1202 Do not apply any gain.
1205 select peak gain, very conservative approach. This is default value.
1208 select DC gain, limited application.
1211 select gain to noise approach, this is most popular one.
1215 Set gain to be applied to IR coefficients before filtering.
1216 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1219 Set format of IR stream. Can be @code{mono} or @code{input}.
1220 Default is @code{input}.
1223 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1224 Allowed range is 0.1 to 60 seconds.
1227 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1228 By default it is disabled.
1231 Set for which IR channel to display frequency response. By default is first channel
1232 displayed. This option is used only when @var{response} is enabled.
1235 Set video stream size. This option is used only when @var{response} is enabled.
1238 Set video stream frame rate. This option is used only when @var{response} is enabled.
1241 Set minimal partition size used for convolution. Default is @var{8192}.
1242 Allowed range is from @var{8} to @var{32768}.
1243 Lower values decreases latency at cost of higher CPU usage.
1246 Set maximal partition size used for convolution. Default is @var{8192}.
1247 Allowed range is from @var{8} to @var{32768}.
1248 Lower values may increase CPU usage.
1251 @subsection Examples
1255 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1257 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1264 Set output format constraints for the input audio. The framework will
1265 negotiate the most appropriate format to minimize conversions.
1267 It accepts the following parameters:
1271 A '|'-separated list of requested sample formats.
1274 A '|'-separated list of requested sample rates.
1276 @item channel_layouts
1277 A '|'-separated list of requested channel layouts.
1279 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1280 for the required syntax.
1283 If a parameter is omitted, all values are allowed.
1285 Force the output to either unsigned 8-bit or signed 16-bit stereo
1287 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1292 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1293 processing reduces disturbing noise between useful signals.
1295 Gating is done by detecting the volume below a chosen level @var{threshold}
1296 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1297 floor is set via @var{range}. Because an exact manipulation of the signal
1298 would cause distortion of the waveform the reduction can be levelled over
1299 time. This is done by setting @var{attack} and @var{release}.
1301 @var{attack} determines how long the signal has to fall below the threshold
1302 before any reduction will occur and @var{release} sets the time the signal
1303 has to rise above the threshold to reduce the reduction again.
1304 Shorter signals than the chosen attack time will be left untouched.
1308 Set input level before filtering.
1309 Default is 1. Allowed range is from 0.015625 to 64.
1312 Set the mode of operation. Can be @code{upward} or @code{downward}.
1313 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1314 will be amplified, expanding dynamic range in upward direction.
1315 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1318 Set the level of gain reduction when the signal is below the threshold.
1319 Default is 0.06125. Allowed range is from 0 to 1.
1320 Setting this to 0 disables reduction and then filter behaves like expander.
1323 If a signal rises above this level the gain reduction is released.
1324 Default is 0.125. Allowed range is from 0 to 1.
1327 Set a ratio by which the signal is reduced.
1328 Default is 2. Allowed range is from 1 to 9000.
1331 Amount of milliseconds the signal has to rise above the threshold before gain
1333 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1336 Amount of milliseconds the signal has to fall below the threshold before the
1337 reduction is increased again. Default is 250 milliseconds.
1338 Allowed range is from 0.01 to 9000.
1341 Set amount of amplification of signal after processing.
1342 Default is 1. Allowed range is from 1 to 64.
1345 Curve the sharp knee around the threshold to enter gain reduction more softly.
1346 Default is 2.828427125. Allowed range is from 1 to 8.
1349 Choose if exact signal should be taken for detection or an RMS like one.
1350 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1353 Choose if the average level between all channels or the louder channel affects
1355 Default is @code{average}. Can be @code{average} or @code{maximum}.
1360 Apply an arbitrary Infinite Impulse Response filter.
1362 It accepts the following parameters:
1366 Set numerator/zeros coefficients.
1369 Set denominator/poles coefficients.
1381 Set coefficients format.
1387 Z-plane zeros/poles, cartesian (default)
1389 Z-plane zeros/poles, polar radians
1391 Z-plane zeros/poles, polar degrees
1395 Set kind of processing.
1396 Can be @code{d} - direct or @code{s} - serial cascading. Default is @code{s}.
1399 Set filtering precision.
1403 double-precision floating-point (default)
1405 single-precision floating-point
1413 Show IR frequency response, magnitude and phase in additional video stream.
1414 By default it is disabled.
1417 Set for which IR channel to display frequency response. By default is first channel
1418 displayed. This option is used only when @var{response} is enabled.
1421 Set video stream size. This option is used only when @var{response} is enabled.
1424 Coefficients in @code{tf} format are separated by spaces and are in ascending
1427 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1428 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1431 Different coefficients and gains can be provided for every channel, in such case
1432 use '|' to separate coefficients or gains. Last provided coefficients will be
1433 used for all remaining channels.
1435 @subsection Examples
1439 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1441 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
1445 Same as above but in @code{zp} format:
1447 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
1453 The limiter prevents an input signal from rising over a desired threshold.
1454 This limiter uses lookahead technology to prevent your signal from distorting.
1455 It means that there is a small delay after the signal is processed. Keep in mind
1456 that the delay it produces is the attack time you set.
1458 The filter accepts the following options:
1462 Set input gain. Default is 1.
1465 Set output gain. Default is 1.
1468 Don't let signals above this level pass the limiter. Default is 1.
1471 The limiter will reach its attenuation level in this amount of time in
1472 milliseconds. Default is 5 milliseconds.
1475 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1476 Default is 50 milliseconds.
1479 When gain reduction is always needed ASC takes care of releasing to an
1480 average reduction level rather than reaching a reduction of 0 in the release
1484 Select how much the release time is affected by ASC, 0 means nearly no changes
1485 in release time while 1 produces higher release times.
1488 Auto level output signal. Default is enabled.
1489 This normalizes audio back to 0dB if enabled.
1492 Depending on picked setting it is recommended to upsample input 2x or 4x times
1493 with @ref{aresample} before applying this filter.
1497 Apply a two-pole all-pass filter with central frequency (in Hz)
1498 @var{frequency}, and filter-width @var{width}.
1499 An all-pass filter changes the audio's frequency to phase relationship
1500 without changing its frequency to amplitude relationship.
1502 The filter accepts the following options:
1506 Set frequency in Hz.
1509 Set method to specify band-width of filter.
1524 Specify the band-width of a filter in width_type units.
1527 Specify which channels to filter, by default all available are filtered.
1530 @subsection Commands
1532 This filter supports the following commands:
1535 Change allpass frequency.
1536 Syntax for the command is : "@var{frequency}"
1539 Change allpass width_type.
1540 Syntax for the command is : "@var{width_type}"
1543 Change allpass width.
1544 Syntax for the command is : "@var{width}"
1551 The filter accepts the following options:
1555 Set the number of loops. Setting this value to -1 will result in infinite loops.
1559 Set maximal number of samples. Default is 0.
1562 Set first sample of loop. Default is 0.
1568 Merge two or more audio streams into a single multi-channel stream.
1570 The filter accepts the following options:
1575 Set the number of inputs. Default is 2.
1579 If the channel layouts of the inputs are disjoint, and therefore compatible,
1580 the channel layout of the output will be set accordingly and the channels
1581 will be reordered as necessary. If the channel layouts of the inputs are not
1582 disjoint, the output will have all the channels of the first input then all
1583 the channels of the second input, in that order, and the channel layout of
1584 the output will be the default value corresponding to the total number of
1587 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1588 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1589 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1590 first input, b1 is the first channel of the second input).
1592 On the other hand, if both input are in stereo, the output channels will be
1593 in the default order: a1, a2, b1, b2, and the channel layout will be
1594 arbitrarily set to 4.0, which may or may not be the expected value.
1596 All inputs must have the same sample rate, and format.
1598 If inputs do not have the same duration, the output will stop with the
1601 @subsection Examples
1605 Merge two mono files into a stereo stream:
1607 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1611 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1613 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
1619 Mixes multiple audio inputs into a single output.
1621 Note that this filter only supports float samples (the @var{amerge}
1622 and @var{pan} audio filters support many formats). If the @var{amix}
1623 input has integer samples then @ref{aresample} will be automatically
1624 inserted to perform the conversion to float samples.
1628 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1630 will mix 3 input audio streams to a single output with the same duration as the
1631 first input and a dropout transition time of 3 seconds.
1633 It accepts the following parameters:
1637 The number of inputs. If unspecified, it defaults to 2.
1640 How to determine the end-of-stream.
1644 The duration of the longest input. (default)
1647 The duration of the shortest input.
1650 The duration of the first input.
1654 @item dropout_transition
1655 The transition time, in seconds, for volume renormalization when an input
1656 stream ends. The default value is 2 seconds.
1659 Specify weight of each input audio stream as sequence.
1660 Each weight is separated by space. By default all inputs have same weight.
1665 Multiply first audio stream with second audio stream and store result
1666 in output audio stream. Multiplication is done by multiplying each
1667 sample from first stream with sample at same position from second stream.
1669 With this element-wise multiplication one can create amplitude fades and
1670 amplitude modulations.
1672 @section anequalizer
1674 High-order parametric multiband equalizer for each channel.
1676 It accepts the following parameters:
1680 This option string is in format:
1681 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1682 Each equalizer band is separated by '|'.
1686 Set channel number to which equalization will be applied.
1687 If input doesn't have that channel the entry is ignored.
1690 Set central frequency for band.
1691 If input doesn't have that frequency the entry is ignored.
1694 Set band width in hertz.
1697 Set band gain in dB.
1700 Set filter type for band, optional, can be:
1704 Butterworth, this is default.
1715 With this option activated frequency response of anequalizer is displayed
1719 Set video stream size. Only useful if curves option is activated.
1722 Set max gain that will be displayed. Only useful if curves option is activated.
1723 Setting this to a reasonable value makes it possible to display gain which is derived from
1724 neighbour bands which are too close to each other and thus produce higher gain
1725 when both are activated.
1728 Set frequency scale used to draw frequency response in video output.
1729 Can be linear or logarithmic. Default is logarithmic.
1732 Set color for each channel curve which is going to be displayed in video stream.
1733 This is list of color names separated by space or by '|'.
1734 Unrecognised or missing colors will be replaced by white color.
1737 @subsection Examples
1741 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1742 for first 2 channels using Chebyshev type 1 filter:
1744 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1748 @subsection Commands
1750 This filter supports the following commands:
1753 Alter existing filter parameters.
1754 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1756 @var{fN} is existing filter number, starting from 0, if no such filter is available
1758 @var{freq} set new frequency parameter.
1759 @var{width} set new width parameter in herz.
1760 @var{gain} set new gain parameter in dB.
1762 Full filter invocation with asendcmd may look like this:
1763 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1768 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1770 Each sample is adjusted by looking for other samples with similar contexts. This
1771 context similarity is defined by comparing their surrounding patches of size
1772 @option{p}. Patches are searched in an area of @option{r} around the sample.
1774 The filter accepts the following options.
1778 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1781 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1782 Default value is 2 milliseconds.
1785 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1786 Default value is 6 milliseconds.
1789 Set the output mode.
1791 It accepts the following values:
1794 Pass input unchanged.
1797 Pass noise filtered out.
1802 Default value is @var{o}.
1808 Pass the audio source unchanged to the output.
1812 Pad the end of an audio stream with silence.
1814 This can be used together with @command{ffmpeg} @option{-shortest} to
1815 extend audio streams to the same length as the video stream.
1817 A description of the accepted options follows.
1821 Set silence packet size. Default value is 4096.
1824 Set the number of samples of silence to add to the end. After the
1825 value is reached, the stream is terminated. This option is mutually
1826 exclusive with @option{whole_len}.
1829 Set the minimum total number of samples in the output audio stream. If
1830 the value is longer than the input audio length, silence is added to
1831 the end, until the value is reached. This option is mutually exclusive
1832 with @option{pad_len}.
1835 Specify the duration of samples of silence to add. See
1836 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1837 for the accepted syntax. Used only if set to non-zero value.
1840 Specify the minimum total duration in the output audio stream. See
1841 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1842 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
1843 the input audio length, silence is added to the end, until the value is reached.
1844 This option is mutually exclusive with @option{pad_dur}
1847 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
1848 nor @option{whole_dur} option is set, the filter will add silence to the end of
1849 the input stream indefinitely.
1851 @subsection Examples
1855 Add 1024 samples of silence to the end of the input:
1861 Make sure the audio output will contain at least 10000 samples, pad
1862 the input with silence if required:
1864 apad=whole_len=10000
1868 Use @command{ffmpeg} to pad the audio input with silence, so that the
1869 video stream will always result the shortest and will be converted
1870 until the end in the output file when using the @option{shortest}
1873 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
1878 Add a phasing effect to the input audio.
1880 A phaser filter creates series of peaks and troughs in the frequency spectrum.
1881 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
1883 A description of the accepted parameters follows.
1887 Set input gain. Default is 0.4.
1890 Set output gain. Default is 0.74
1893 Set delay in milliseconds. Default is 3.0.
1896 Set decay. Default is 0.4.
1899 Set modulation speed in Hz. Default is 0.5.
1902 Set modulation type. Default is triangular.
1904 It accepts the following values:
1913 Audio pulsator is something between an autopanner and a tremolo.
1914 But it can produce funny stereo effects as well. Pulsator changes the volume
1915 of the left and right channel based on a LFO (low frequency oscillator) with
1916 different waveforms and shifted phases.
1917 This filter have the ability to define an offset between left and right
1918 channel. An offset of 0 means that both LFO shapes match each other.
1919 The left and right channel are altered equally - a conventional tremolo.
1920 An offset of 50% means that the shape of the right channel is exactly shifted
1921 in phase (or moved backwards about half of the frequency) - pulsator acts as
1922 an autopanner. At 1 both curves match again. Every setting in between moves the
1923 phase shift gapless between all stages and produces some "bypassing" sounds with
1924 sine and triangle waveforms. The more you set the offset near 1 (starting from
1925 the 0.5) the faster the signal passes from the left to the right speaker.
1927 The filter accepts the following options:
1931 Set input gain. By default it is 1. Range is [0.015625 - 64].
1934 Set output gain. By default it is 1. Range is [0.015625 - 64].
1937 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
1938 sawup or sawdown. Default is sine.
1941 Set modulation. Define how much of original signal is affected by the LFO.
1944 Set left channel offset. Default is 0. Allowed range is [0 - 1].
1947 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
1950 Set pulse width. Default is 1. Allowed range is [0 - 2].
1953 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
1956 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
1960 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
1964 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
1965 if timing is set to hz.
1971 Resample the input audio to the specified parameters, using the
1972 libswresample library. If none are specified then the filter will
1973 automatically convert between its input and output.
1975 This filter is also able to stretch/squeeze the audio data to make it match
1976 the timestamps or to inject silence / cut out audio to make it match the
1977 timestamps, do a combination of both or do neither.
1979 The filter accepts the syntax
1980 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
1981 expresses a sample rate and @var{resampler_options} is a list of
1982 @var{key}=@var{value} pairs, separated by ":". See the
1983 @ref{Resampler Options,,"Resampler Options" section in the
1984 ffmpeg-resampler(1) manual,ffmpeg-resampler}
1985 for the complete list of supported options.
1987 @subsection Examples
1991 Resample the input audio to 44100Hz:
1997 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
1998 samples per second compensation:
2000 aresample=async=1000
2006 Reverse an audio clip.
2008 Warning: This filter requires memory to buffer the entire clip, so trimming
2011 @subsection Examples
2015 Take the first 5 seconds of a clip, and reverse it.
2017 atrim=end=5,areverse
2021 @section asetnsamples
2023 Set the number of samples per each output audio frame.
2025 The last output packet may contain a different number of samples, as
2026 the filter will flush all the remaining samples when the input audio
2029 The filter accepts the following options:
2033 @item nb_out_samples, n
2034 Set the number of frames per each output audio frame. The number is
2035 intended as the number of samples @emph{per each channel}.
2036 Default value is 1024.
2039 If set to 1, the filter will pad the last audio frame with zeroes, so
2040 that the last frame will contain the same number of samples as the
2041 previous ones. Default value is 1.
2044 For example, to set the number of per-frame samples to 1234 and
2045 disable padding for the last frame, use:
2047 asetnsamples=n=1234:p=0
2052 Set the sample rate without altering the PCM data.
2053 This will result in a change of speed and pitch.
2055 The filter accepts the following options:
2058 @item sample_rate, r
2059 Set the output sample rate. Default is 44100 Hz.
2064 Show a line containing various information for each input audio frame.
2065 The input audio is not modified.
2067 The shown line contains a sequence of key/value pairs of the form
2068 @var{key}:@var{value}.
2070 The following values are shown in the output:
2074 The (sequential) number of the input frame, starting from 0.
2077 The presentation timestamp of the input frame, in time base units; the time base
2078 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2081 The presentation timestamp of the input frame in seconds.
2084 position of the frame in the input stream, -1 if this information in
2085 unavailable and/or meaningless (for example in case of synthetic audio)
2094 The sample rate for the audio frame.
2097 The number of samples (per channel) in the frame.
2100 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2101 audio, the data is treated as if all the planes were concatenated.
2103 @item plane_checksums
2104 A list of Adler-32 checksums for each data plane.
2110 Display time domain statistical information about the audio channels.
2111 Statistics are calculated and displayed for each audio channel and,
2112 where applicable, an overall figure is also given.
2114 It accepts the following option:
2117 Short window length in seconds, used for peak and trough RMS measurement.
2118 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2122 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2123 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2126 Available keys for each channel are:
2162 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2163 this @code{lavfi.astats.Overall.Peak_count}.
2165 For description what each key means read below.
2168 Set number of frame after which stats are going to be recalculated.
2169 Default is disabled.
2171 @item measure_perchannel
2172 Select the entries which need to be measured per channel. The metadata keys can
2173 be used as flags, default is @option{all} which measures everything.
2174 @option{none} disables all per channel measurement.
2176 @item measure_overall
2177 Select the entries which need to be measured overall. The metadata keys can
2178 be used as flags, default is @option{all} which measures everything.
2179 @option{none} disables all overall measurement.
2183 A description of each shown parameter follows:
2187 Mean amplitude displacement from zero.
2190 Minimal sample level.
2193 Maximal sample level.
2195 @item Min difference
2196 Minimal difference between two consecutive samples.
2198 @item Max difference
2199 Maximal difference between two consecutive samples.
2201 @item Mean difference
2202 Mean difference between two consecutive samples.
2203 The average of each difference between two consecutive samples.
2205 @item RMS difference
2206 Root Mean Square difference between two consecutive samples.
2210 Standard peak and RMS level measured in dBFS.
2214 Peak and trough values for RMS level measured over a short window.
2217 Standard ratio of peak to RMS level (note: not in dB).
2220 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2221 (i.e. either @var{Min level} or @var{Max level}).
2224 Number of occasions (not the number of samples) that the signal attained either
2225 @var{Min level} or @var{Max level}.
2228 Overall bit depth of audio. Number of bits used for each sample.
2231 Measured dynamic range of audio in dB.
2233 @item Zero crossings
2234 Number of points where the waveform crosses the zero level axis.
2236 @item Zero crossings rate
2237 Rate of Zero crossings and number of audio samples.
2244 The filter accepts exactly one parameter, the audio tempo. If not
2245 specified then the filter will assume nominal 1.0 tempo. Tempo must
2246 be in the [0.5, 100.0] range.
2248 Note that tempo greater than 2 will skip some samples rather than
2249 blend them in. If for any reason this is a concern it is always
2250 possible to daisy-chain several instances of atempo to achieve the
2251 desired product tempo.
2253 @subsection Examples
2257 Slow down audio to 80% tempo:
2263 To speed up audio to 300% tempo:
2269 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2271 atempo=sqrt(3),atempo=sqrt(3)
2277 Trim the input so that the output contains one continuous subpart of the input.
2279 It accepts the following parameters:
2282 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2283 sample with the timestamp @var{start} will be the first sample in the output.
2286 Specify time of the first audio sample that will be dropped, i.e. the
2287 audio sample immediately preceding the one with the timestamp @var{end} will be
2288 the last sample in the output.
2291 Same as @var{start}, except this option sets the start timestamp in samples
2295 Same as @var{end}, except this option sets the end timestamp in samples instead
2299 The maximum duration of the output in seconds.
2302 The number of the first sample that should be output.
2305 The number of the first sample that should be dropped.
2308 @option{start}, @option{end}, and @option{duration} are expressed as time
2309 duration specifications; see
2310 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2312 Note that the first two sets of the start/end options and the @option{duration}
2313 option look at the frame timestamp, while the _sample options simply count the
2314 samples that pass through the filter. So start/end_pts and start/end_sample will
2315 give different results when the timestamps are wrong, inexact or do not start at
2316 zero. Also note that this filter does not modify the timestamps. If you wish
2317 to have the output timestamps start at zero, insert the asetpts filter after the
2320 If multiple start or end options are set, this filter tries to be greedy and
2321 keep all samples that match at least one of the specified constraints. To keep
2322 only the part that matches all the constraints at once, chain multiple atrim
2325 The defaults are such that all the input is kept. So it is possible to set e.g.
2326 just the end values to keep everything before the specified time.
2331 Drop everything except the second minute of input:
2333 ffmpeg -i INPUT -af atrim=60:120
2337 Keep only the first 1000 samples:
2339 ffmpeg -i INPUT -af atrim=end_sample=1000
2346 Apply a two-pole Butterworth band-pass filter with central
2347 frequency @var{frequency}, and (3dB-point) band-width width.
2348 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2349 instead of the default: constant 0dB peak gain.
2350 The filter roll off at 6dB per octave (20dB per decade).
2352 The filter accepts the following options:
2356 Set the filter's central frequency. Default is @code{3000}.
2359 Constant skirt gain if set to 1. Defaults to 0.
2362 Set method to specify band-width of filter.
2377 Specify the band-width of a filter in width_type units.
2380 Specify which channels to filter, by default all available are filtered.
2383 @subsection Commands
2385 This filter supports the following commands:
2388 Change bandpass frequency.
2389 Syntax for the command is : "@var{frequency}"
2392 Change bandpass width_type.
2393 Syntax for the command is : "@var{width_type}"
2396 Change bandpass width.
2397 Syntax for the command is : "@var{width}"
2402 Apply a two-pole Butterworth band-reject filter with central
2403 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2404 The filter roll off at 6dB per octave (20dB per decade).
2406 The filter accepts the following options:
2410 Set the filter's central frequency. Default is @code{3000}.
2413 Set method to specify band-width of filter.
2428 Specify the band-width of a filter in width_type units.
2431 Specify which channels to filter, by default all available are filtered.
2434 @subsection Commands
2436 This filter supports the following commands:
2439 Change bandreject frequency.
2440 Syntax for the command is : "@var{frequency}"
2443 Change bandreject width_type.
2444 Syntax for the command is : "@var{width_type}"
2447 Change bandreject width.
2448 Syntax for the command is : "@var{width}"
2451 @section bass, lowshelf
2453 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2454 shelving filter with a response similar to that of a standard
2455 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2457 The filter accepts the following options:
2461 Give the gain at 0 Hz. Its useful range is about -20
2462 (for a large cut) to +20 (for a large boost).
2463 Beware of clipping when using a positive gain.
2466 Set the filter's central frequency and so can be used
2467 to extend or reduce the frequency range to be boosted or cut.
2468 The default value is @code{100} Hz.
2471 Set method to specify band-width of filter.
2486 Determine how steep is the filter's shelf transition.
2489 Specify which channels to filter, by default all available are filtered.
2492 @subsection Commands
2494 This filter supports the following commands:
2497 Change bass frequency.
2498 Syntax for the command is : "@var{frequency}"
2501 Change bass width_type.
2502 Syntax for the command is : "@var{width_type}"
2506 Syntax for the command is : "@var{width}"
2510 Syntax for the command is : "@var{gain}"
2515 Apply a biquad IIR filter with the given coefficients.
2516 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2517 are the numerator and denominator coefficients respectively.
2518 and @var{channels}, @var{c} specify which channels to filter, by default all
2519 available are filtered.
2521 @subsection Commands
2523 This filter supports the following commands:
2531 Change biquad parameter.
2532 Syntax for the command is : "@var{value}"
2536 Bauer stereo to binaural transformation, which improves headphone listening of
2537 stereo audio records.
2539 To enable compilation of this filter you need to configure FFmpeg with
2540 @code{--enable-libbs2b}.
2542 It accepts the following parameters:
2546 Pre-defined crossfeed level.
2550 Default level (fcut=700, feed=50).
2553 Chu Moy circuit (fcut=700, feed=60).
2556 Jan Meier circuit (fcut=650, feed=95).
2561 Cut frequency (in Hz).
2570 Remap input channels to new locations.
2572 It accepts the following parameters:
2575 Map channels from input to output. The argument is a '|'-separated list of
2576 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2577 @var{in_channel} form. @var{in_channel} can be either the name of the input
2578 channel (e.g. FL for front left) or its index in the input channel layout.
2579 @var{out_channel} is the name of the output channel or its index in the output
2580 channel layout. If @var{out_channel} is not given then it is implicitly an
2581 index, starting with zero and increasing by one for each mapping.
2583 @item channel_layout
2584 The channel layout of the output stream.
2587 If no mapping is present, the filter will implicitly map input channels to
2588 output channels, preserving indices.
2590 @subsection Examples
2594 For example, assuming a 5.1+downmix input MOV file,
2596 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2598 will create an output WAV file tagged as stereo from the downmix channels of
2602 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2604 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2608 @section channelsplit
2610 Split each channel from an input audio stream into a separate output stream.
2612 It accepts the following parameters:
2614 @item channel_layout
2615 The channel layout of the input stream. The default is "stereo".
2617 A channel layout describing the channels to be extracted as separate output streams
2618 or "all" to extract each input channel as a separate stream. The default is "all".
2620 Choosing channels not present in channel layout in the input will result in an error.
2623 @subsection Examples
2627 For example, assuming a stereo input MP3 file,
2629 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2631 will create an output Matroska file with two audio streams, one containing only
2632 the left channel and the other the right channel.
2635 Split a 5.1 WAV file into per-channel files:
2637 ffmpeg -i in.wav -filter_complex
2638 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2639 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2640 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2645 Extract only LFE from a 5.1 WAV file:
2647 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
2648 -map '[LFE]' lfe.wav
2653 Add a chorus effect to the audio.
2655 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2657 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2658 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2659 The modulation depth defines the range the modulated delay is played before or after
2660 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2661 sound tuned around the original one, like in a chorus where some vocals are slightly
2664 It accepts the following parameters:
2667 Set input gain. Default is 0.4.
2670 Set output gain. Default is 0.4.
2673 Set delays. A typical delay is around 40ms to 60ms.
2685 @subsection Examples
2691 chorus=0.7:0.9:55:0.4:0.25:2
2697 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2701 Fuller sounding chorus with three delays:
2703 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
2708 Compress or expand the audio's dynamic range.
2710 It accepts the following parameters:
2716 A list of times in seconds for each channel over which the instantaneous level
2717 of the input signal is averaged to determine its volume. @var{attacks} refers to
2718 increase of volume and @var{decays} refers to decrease of volume. For most
2719 situations, the attack time (response to the audio getting louder) should be
2720 shorter than the decay time, because the human ear is more sensitive to sudden
2721 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
2722 a typical value for decay is 0.8 seconds.
2723 If specified number of attacks & decays is lower than number of channels, the last
2724 set attack/decay will be used for all remaining channels.
2727 A list of points for the transfer function, specified in dB relative to the
2728 maximum possible signal amplitude. Each key points list must be defined using
2729 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
2730 @code{x0/y0 x1/y1 x2/y2 ....}
2732 The input values must be in strictly increasing order but the transfer function
2733 does not have to be monotonically rising. The point @code{0/0} is assumed but
2734 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
2735 function are @code{-70/-70|-60/-20|1/0}.
2738 Set the curve radius in dB for all joints. It defaults to 0.01.
2741 Set the additional gain in dB to be applied at all points on the transfer
2742 function. This allows for easy adjustment of the overall gain.
2746 Set an initial volume, in dB, to be assumed for each channel when filtering
2747 starts. This permits the user to supply a nominal level initially, so that, for
2748 example, a very large gain is not applied to initial signal levels before the
2749 companding has begun to operate. A typical value for audio which is initially
2750 quiet is -90 dB. It defaults to 0.
2753 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
2754 delayed before being fed to the volume adjuster. Specifying a delay
2755 approximately equal to the attack/decay times allows the filter to effectively
2756 operate in predictive rather than reactive mode. It defaults to 0.
2760 @subsection Examples
2764 Make music with both quiet and loud passages suitable for listening to in a
2767 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
2770 Another example for audio with whisper and explosion parts:
2772 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
2776 A noise gate for when the noise is at a lower level than the signal:
2778 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
2782 Here is another noise gate, this time for when the noise is at a higher level
2783 than the signal (making it, in some ways, similar to squelch):
2785 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
2789 2:1 compression starting at -6dB:
2791 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
2795 2:1 compression starting at -9dB:
2797 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
2801 2:1 compression starting at -12dB:
2803 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
2807 2:1 compression starting at -18dB:
2809 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
2813 3:1 compression starting at -15dB:
2815 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
2821 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
2827 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
2831 Hard limiter at -6dB:
2833 compand=attacks=0:points=-80/-80|-6/-6|20/-6
2837 Hard limiter at -12dB:
2839 compand=attacks=0:points=-80/-80|-12/-12|20/-12
2843 Hard noise gate at -35 dB:
2845 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
2851 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
2855 @section compensationdelay
2857 Compensation Delay Line is a metric based delay to compensate differing
2858 positions of microphones or speakers.
2860 For example, you have recorded guitar with two microphones placed in
2861 different location. Because the front of sound wave has fixed speed in
2862 normal conditions, the phasing of microphones can vary and depends on
2863 their location and interposition. The best sound mix can be achieved when
2864 these microphones are in phase (synchronized). Note that distance of
2865 ~30 cm between microphones makes one microphone to capture signal in
2866 antiphase to another microphone. That makes the final mix sounding moody.
2867 This filter helps to solve phasing problems by adding different delays
2868 to each microphone track and make them synchronized.
2870 The best result can be reached when you take one track as base and
2871 synchronize other tracks one by one with it.
2872 Remember that synchronization/delay tolerance depends on sample rate, too.
2873 Higher sample rates will give more tolerance.
2875 It accepts the following parameters:
2879 Set millimeters distance. This is compensation distance for fine tuning.
2883 Set cm distance. This is compensation distance for tightening distance setup.
2887 Set meters distance. This is compensation distance for hard distance setup.
2891 Set dry amount. Amount of unprocessed (dry) signal.
2895 Set wet amount. Amount of processed (wet) signal.
2899 Set temperature degree in Celsius. This is the temperature of the environment.
2904 Apply headphone crossfeed filter.
2906 Crossfeed is the process of blending the left and right channels of stereo
2908 It is mainly used to reduce extreme stereo separation of low frequencies.
2910 The intent is to produce more speaker like sound to the listener.
2912 The filter accepts the following options:
2916 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
2917 This sets gain of low shelf filter for side part of stereo image.
2918 Default is -6dB. Max allowed is -30db when strength is set to 1.
2921 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
2922 This sets cut off frequency of low shelf filter. Default is cut off near
2923 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
2926 Set input gain. Default is 0.9.
2929 Set output gain. Default is 1.
2932 @section crystalizer
2933 Simple algorithm to expand audio dynamic range.
2935 The filter accepts the following options:
2939 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
2940 (unchanged sound) to 10.0 (maximum effect).
2943 Enable clipping. By default is enabled.
2947 Apply a DC shift to the audio.
2949 This can be useful to remove a DC offset (caused perhaps by a hardware problem
2950 in the recording chain) from the audio. The effect of a DC offset is reduced
2951 headroom and hence volume. The @ref{astats} filter can be used to determine if
2952 a signal has a DC offset.
2956 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
2960 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
2961 used to prevent clipping.
2965 Measure audio dynamic range.
2967 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
2968 is found in transition material. And anything less that 8 have very poor dynamics
2969 and is very compressed.
2971 The filter accepts the following options:
2975 Set window length in seconds used to split audio into segments of equal length.
2976 Default is 3 seconds.
2980 Dynamic Audio Normalizer.
2982 This filter applies a certain amount of gain to the input audio in order
2983 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
2984 contrast to more "simple" normalization algorithms, the Dynamic Audio
2985 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
2986 This allows for applying extra gain to the "quiet" sections of the audio
2987 while avoiding distortions or clipping the "loud" sections. In other words:
2988 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
2989 sections, in the sense that the volume of each section is brought to the
2990 same target level. Note, however, that the Dynamic Audio Normalizer achieves
2991 this goal *without* applying "dynamic range compressing". It will retain 100%
2992 of the dynamic range *within* each section of the audio file.
2996 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
2997 Default is 500 milliseconds.
2998 The Dynamic Audio Normalizer processes the input audio in small chunks,
2999 referred to as frames. This is required, because a peak magnitude has no
3000 meaning for just a single sample value. Instead, we need to determine the
3001 peak magnitude for a contiguous sequence of sample values. While a "standard"
3002 normalizer would simply use the peak magnitude of the complete file, the
3003 Dynamic Audio Normalizer determines the peak magnitude individually for each
3004 frame. The length of a frame is specified in milliseconds. By default, the
3005 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3006 been found to give good results with most files.
3007 Note that the exact frame length, in number of samples, will be determined
3008 automatically, based on the sampling rate of the individual input audio file.
3011 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3012 number. Default is 31.
3013 Probably the most important parameter of the Dynamic Audio Normalizer is the
3014 @code{window size} of the Gaussian smoothing filter. The filter's window size
3015 is specified in frames, centered around the current frame. For the sake of
3016 simplicity, this must be an odd number. Consequently, the default value of 31
3017 takes into account the current frame, as well as the 15 preceding frames and
3018 the 15 subsequent frames. Using a larger window results in a stronger
3019 smoothing effect and thus in less gain variation, i.e. slower gain
3020 adaptation. Conversely, using a smaller window results in a weaker smoothing
3021 effect and thus in more gain variation, i.e. faster gain adaptation.
3022 In other words, the more you increase this value, the more the Dynamic Audio
3023 Normalizer will behave like a "traditional" normalization filter. On the
3024 contrary, the more you decrease this value, the more the Dynamic Audio
3025 Normalizer will behave like a dynamic range compressor.
3028 Set the target peak value. This specifies the highest permissible magnitude
3029 level for the normalized audio input. This filter will try to approach the
3030 target peak magnitude as closely as possible, but at the same time it also
3031 makes sure that the normalized signal will never exceed the peak magnitude.
3032 A frame's maximum local gain factor is imposed directly by the target peak
3033 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3034 It is not recommended to go above this value.
3037 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3038 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3039 factor for each input frame, i.e. the maximum gain factor that does not
3040 result in clipping or distortion. The maximum gain factor is determined by
3041 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3042 additionally bounds the frame's maximum gain factor by a predetermined
3043 (global) maximum gain factor. This is done in order to avoid excessive gain
3044 factors in "silent" or almost silent frames. By default, the maximum gain
3045 factor is 10.0, For most inputs the default value should be sufficient and
3046 it usually is not recommended to increase this value. Though, for input
3047 with an extremely low overall volume level, it may be necessary to allow even
3048 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3049 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3050 Instead, a "sigmoid" threshold function will be applied. This way, the
3051 gain factors will smoothly approach the threshold value, but never exceed that
3055 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3056 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3057 This means that the maximum local gain factor for each frame is defined
3058 (only) by the frame's highest magnitude sample. This way, the samples can
3059 be amplified as much as possible without exceeding the maximum signal
3060 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3061 Normalizer can also take into account the frame's root mean square,
3062 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3063 determine the power of a time-varying signal. It is therefore considered
3064 that the RMS is a better approximation of the "perceived loudness" than
3065 just looking at the signal's peak magnitude. Consequently, by adjusting all
3066 frames to a constant RMS value, a uniform "perceived loudness" can be
3067 established. If a target RMS value has been specified, a frame's local gain
3068 factor is defined as the factor that would result in exactly that RMS value.
3069 Note, however, that the maximum local gain factor is still restricted by the
3070 frame's highest magnitude sample, in order to prevent clipping.
3073 Enable channels coupling. By default is enabled.
3074 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3075 amount. This means the same gain factor will be applied to all channels, i.e.
3076 the maximum possible gain factor is determined by the "loudest" channel.
3077 However, in some recordings, it may happen that the volume of the different
3078 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3079 In this case, this option can be used to disable the channel coupling. This way,
3080 the gain factor will be determined independently for each channel, depending
3081 only on the individual channel's highest magnitude sample. This allows for
3082 harmonizing the volume of the different channels.
3085 Enable DC bias correction. By default is disabled.
3086 An audio signal (in the time domain) is a sequence of sample values.
3087 In the Dynamic Audio Normalizer these sample values are represented in the
3088 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3089 audio signal, or "waveform", should be centered around the zero point.
3090 That means if we calculate the mean value of all samples in a file, or in a
3091 single frame, then the result should be 0.0 or at least very close to that
3092 value. If, however, there is a significant deviation of the mean value from
3093 0.0, in either positive or negative direction, this is referred to as a
3094 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3095 Audio Normalizer provides optional DC bias correction.
3096 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3097 the mean value, or "DC correction" offset, of each input frame and subtract
3098 that value from all of the frame's sample values which ensures those samples
3099 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3100 boundaries, the DC correction offset values will be interpolated smoothly
3101 between neighbouring frames.
3104 Enable alternative boundary mode. By default is disabled.
3105 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3106 around each frame. This includes the preceding frames as well as the
3107 subsequent frames. However, for the "boundary" frames, located at the very
3108 beginning and at the very end of the audio file, not all neighbouring
3109 frames are available. In particular, for the first few frames in the audio
3110 file, the preceding frames are not known. And, similarly, for the last few
3111 frames in the audio file, the subsequent frames are not known. Thus, the
3112 question arises which gain factors should be assumed for the missing frames
3113 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3114 to deal with this situation. The default boundary mode assumes a gain factor
3115 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3116 "fade out" at the beginning and at the end of the input, respectively.
3119 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3120 By default, the Dynamic Audio Normalizer does not apply "traditional"
3121 compression. This means that signal peaks will not be pruned and thus the
3122 full dynamic range will be retained within each local neighbourhood. However,
3123 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3124 normalization algorithm with a more "traditional" compression.
3125 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3126 (thresholding) function. If (and only if) the compression feature is enabled,
3127 all input frames will be processed by a soft knee thresholding function prior
3128 to the actual normalization process. Put simply, the thresholding function is
3129 going to prune all samples whose magnitude exceeds a certain threshold value.
3130 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3131 value. Instead, the threshold value will be adjusted for each individual
3133 In general, smaller parameters result in stronger compression, and vice versa.
3134 Values below 3.0 are not recommended, because audible distortion may appear.
3139 Make audio easier to listen to on headphones.
3141 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3142 so that when listened to on headphones the stereo image is moved from
3143 inside your head (standard for headphones) to outside and in front of
3144 the listener (standard for speakers).
3150 Apply a two-pole peaking equalisation (EQ) filter. With this
3151 filter, the signal-level at and around a selected frequency can
3152 be increased or decreased, whilst (unlike bandpass and bandreject
3153 filters) that at all other frequencies is unchanged.
3155 In order to produce complex equalisation curves, this filter can
3156 be given several times, each with a different central frequency.
3158 The filter accepts the following options:
3162 Set the filter's central frequency in Hz.
3165 Set method to specify band-width of filter.
3180 Specify the band-width of a filter in width_type units.
3183 Set the required gain or attenuation in dB.
3184 Beware of clipping when using a positive gain.
3187 Specify which channels to filter, by default all available are filtered.
3190 @subsection Examples
3193 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3195 equalizer=f=1000:t=h:width=200:g=-10
3199 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3201 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3205 @subsection Commands
3207 This filter supports the following commands:
3210 Change equalizer frequency.
3211 Syntax for the command is : "@var{frequency}"
3214 Change equalizer width_type.
3215 Syntax for the command is : "@var{width_type}"
3218 Change equalizer width.
3219 Syntax for the command is : "@var{width}"
3222 Change equalizer gain.
3223 Syntax for the command is : "@var{gain}"
3226 @section extrastereo
3228 Linearly increases the difference between left and right channels which
3229 adds some sort of "live" effect to playback.
3231 The filter accepts the following options:
3235 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3236 (average of both channels), with 1.0 sound will be unchanged, with
3237 -1.0 left and right channels will be swapped.
3240 Enable clipping. By default is enabled.
3243 @section firequalizer
3244 Apply FIR Equalization using arbitrary frequency response.
3246 The filter accepts the following option:
3250 Set gain curve equation (in dB). The expression can contain variables:
3253 the evaluated frequency
3257 channel number, set to 0 when multichannels evaluation is disabled
3259 channel id, see libavutil/channel_layout.h, set to the first channel id when
3260 multichannels evaluation is disabled
3264 channel_layout, see libavutil/channel_layout.h
3269 @item gain_interpolate(f)
3270 interpolate gain on frequency f based on gain_entry
3271 @item cubic_interpolate(f)
3272 same as gain_interpolate, but smoother
3274 This option is also available as command. Default is @code{gain_interpolate(f)}.
3277 Set gain entry for gain_interpolate function. The expression can
3281 store gain entry at frequency f with value g
3283 This option is also available as command.
3286 Set filter delay in seconds. Higher value means more accurate.
3287 Default is @code{0.01}.
3290 Set filter accuracy in Hz. Lower value means more accurate.
3291 Default is @code{5}.
3294 Set window function. Acceptable values are:
3297 rectangular window, useful when gain curve is already smooth
3299 hann window (default)
3305 3-terms continuous 1st derivative nuttall window
3307 minimum 3-terms discontinuous nuttall window
3309 4-terms continuous 1st derivative nuttall window
3311 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3313 blackman-harris window
3319 If enabled, use fixed number of audio samples. This improves speed when
3320 filtering with large delay. Default is disabled.
3323 Enable multichannels evaluation on gain. Default is disabled.
3326 Enable zero phase mode by subtracting timestamp to compensate delay.
3327 Default is disabled.
3330 Set scale used by gain. Acceptable values are:
3333 linear frequency, linear gain
3335 linear frequency, logarithmic (in dB) gain (default)
3337 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3339 logarithmic frequency, logarithmic gain
3343 Set file for dumping, suitable for gnuplot.
3346 Set scale for dumpfile. Acceptable values are same with scale option.
3350 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3351 Default is disabled.
3354 Enable minimum phase impulse response. Default is disabled.
3357 @subsection Examples
3362 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3365 lowpass at 1000 Hz with gain_entry:
3367 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3370 custom equalization:
3372 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3375 higher delay with zero phase to compensate delay:
3377 firequalizer=delay=0.1:fixed=on:zero_phase=on
3380 lowpass on left channel, highpass on right channel:
3382 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3383 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3388 Apply a flanging effect to the audio.
3390 The filter accepts the following options:
3394 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3397 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3400 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3404 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3405 Default value is 71.
3408 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3411 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3412 Default value is @var{sinusoidal}.
3415 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3416 Default value is 25.
3419 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3420 Default is @var{linear}.
3424 Apply Haas effect to audio.
3426 Note that this makes most sense to apply on mono signals.
3427 With this filter applied to mono signals it give some directionality and
3428 stretches its stereo image.
3430 The filter accepts the following options:
3434 Set input level. By default is @var{1}, or 0dB
3437 Set output level. By default is @var{1}, or 0dB.
3440 Set gain applied to side part of signal. By default is @var{1}.
3443 Set kind of middle source. Can be one of the following:
3453 Pick middle part signal of stereo image.
3456 Pick side part signal of stereo image.
3460 Change middle phase. By default is disabled.
3463 Set left channel delay. By default is @var{2.05} milliseconds.
3466 Set left channel balance. By default is @var{-1}.
3469 Set left channel gain. By default is @var{1}.
3472 Change left phase. By default is disabled.
3475 Set right channel delay. By defaults is @var{2.12} milliseconds.
3478 Set right channel balance. By default is @var{1}.
3481 Set right channel gain. By default is @var{1}.
3484 Change right phase. By default is enabled.
3489 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3490 embedded HDCD codes is expanded into a 20-bit PCM stream.
3492 The filter supports the Peak Extend and Low-level Gain Adjustment features
3493 of HDCD, and detects the Transient Filter flag.
3496 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3499 When using the filter with wav, note the default encoding for wav is 16-bit,
3500 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3501 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3503 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3504 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3507 The filter accepts the following options:
3510 @item disable_autoconvert
3511 Disable any automatic format conversion or resampling in the filter graph.
3513 @item process_stereo
3514 Process the stereo channels together. If target_gain does not match between
3515 channels, consider it invalid and use the last valid target_gain.
3518 Set the code detect timer period in ms.
3521 Always extend peaks above -3dBFS even if PE isn't signaled.
3524 Replace audio with a solid tone and adjust the amplitude to signal some
3525 specific aspect of the decoding process. The output file can be loaded in
3526 an audio editor alongside the original to aid analysis.
3528 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3535 Gain adjustment level at each sample
3537 Samples where peak extend occurs
3539 Samples where the code detect timer is active
3541 Samples where the target gain does not match between channels
3547 Apply head-related transfer functions (HRTFs) to create virtual
3548 loudspeakers around the user for binaural listening via headphones.
3549 The HRIRs are provided via additional streams, for each channel
3550 one stereo input stream is needed.
3552 The filter accepts the following options:
3556 Set mapping of input streams for convolution.
3557 The argument is a '|'-separated list of channel names in order as they
3558 are given as additional stream inputs for filter.
3559 This also specify number of input streams. Number of input streams
3560 must be not less than number of channels in first stream plus one.
3563 Set gain applied to audio. Value is in dB. Default is 0.
3566 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3567 processing audio in time domain which is slow.
3568 @var{freq} is processing audio in frequency domain which is fast.
3569 Default is @var{freq}.
3572 Set custom gain for LFE channels. Value is in dB. Default is 0.
3575 Set size of frame in number of samples which will be processed at once.
3576 Default value is @var{1024}. Allowed range is from 1024 to 96000.
3579 Set format of hrir stream.
3580 Default value is @var{stereo}. Alternative value is @var{multich}.
3581 If value is set to @var{stereo}, number of additional streams should
3582 be greater or equal to number of input channels in first input stream.
3583 Also each additional stream should have stereo number of channels.
3584 If value is set to @var{multich}, number of additional streams should
3585 be exactly one. Also number of input channels of additional stream
3586 should be equal or greater than twice number of channels of first input
3590 @subsection Examples
3594 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3595 each amovie filter use stereo file with IR coefficients as input.
3596 The files give coefficients for each position of virtual loudspeaker:
3599 -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"
3604 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3605 but now in @var{multich} @var{hrir} format.
3607 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"
3614 Apply a high-pass filter with 3dB point frequency.
3615 The filter can be either single-pole, or double-pole (the default).
3616 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3618 The filter accepts the following options:
3622 Set frequency in Hz. Default is 3000.
3625 Set number of poles. Default is 2.
3628 Set method to specify band-width of filter.
3643 Specify the band-width of a filter in width_type units.
3644 Applies only to double-pole filter.
3645 The default is 0.707q and gives a Butterworth response.
3648 Specify which channels to filter, by default all available are filtered.
3651 @subsection Commands
3653 This filter supports the following commands:
3656 Change highpass frequency.
3657 Syntax for the command is : "@var{frequency}"
3660 Change highpass width_type.
3661 Syntax for the command is : "@var{width_type}"
3664 Change highpass width.
3665 Syntax for the command is : "@var{width}"
3670 Join multiple input streams into one multi-channel stream.
3672 It accepts the following parameters:
3676 The number of input streams. It defaults to 2.
3678 @item channel_layout
3679 The desired output channel layout. It defaults to stereo.
3682 Map channels from inputs to output. The argument is a '|'-separated list of
3683 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
3684 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
3685 can be either the name of the input channel (e.g. FL for front left) or its
3686 index in the specified input stream. @var{out_channel} is the name of the output
3690 The filter will attempt to guess the mappings when they are not specified
3691 explicitly. It does so by first trying to find an unused matching input channel
3692 and if that fails it picks the first unused input channel.
3694 Join 3 inputs (with properly set channel layouts):
3696 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
3699 Build a 5.1 output from 6 single-channel streams:
3701 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
3702 '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'
3708 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
3710 To enable compilation of this filter you need to configure FFmpeg with
3711 @code{--enable-ladspa}.
3715 Specifies the name of LADSPA plugin library to load. If the environment
3716 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
3717 each one of the directories specified by the colon separated list in
3718 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
3719 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
3720 @file{/usr/lib/ladspa/}.
3723 Specifies the plugin within the library. Some libraries contain only
3724 one plugin, but others contain many of them. If this is not set filter
3725 will list all available plugins within the specified library.
3728 Set the '|' separated list of controls which are zero or more floating point
3729 values that determine the behavior of the loaded plugin (for example delay,
3731 Controls need to be defined using the following syntax:
3732 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
3733 @var{valuei} is the value set on the @var{i}-th control.
3734 Alternatively they can be also defined using the following syntax:
3735 @var{value0}|@var{value1}|@var{value2}|..., where
3736 @var{valuei} is the value set on the @var{i}-th control.
3737 If @option{controls} is set to @code{help}, all available controls and
3738 their valid ranges are printed.
3740 @item sample_rate, s
3741 Specify the sample rate, default to 44100. Only used if plugin have
3745 Set the number of samples per channel per each output frame, default
3746 is 1024. Only used if plugin have zero inputs.
3749 Set the minimum duration of the sourced audio. See
3750 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
3751 for the accepted syntax.
3752 Note that the resulting duration may be greater than the specified duration,
3753 as the generated audio is always cut at the end of a complete frame.
3754 If not specified, or the expressed duration is negative, the audio is
3755 supposed to be generated forever.
3756 Only used if plugin have zero inputs.
3760 @subsection Examples
3764 List all available plugins within amp (LADSPA example plugin) library:
3770 List all available controls and their valid ranges for @code{vcf_notch}
3771 plugin from @code{VCF} library:
3773 ladspa=f=vcf:p=vcf_notch:c=help
3777 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
3780 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
3784 Add reverberation to the audio using TAP-plugins
3785 (Tom's Audio Processing plugins):
3787 ladspa=file=tap_reverb:tap_reverb
3791 Generate white noise, with 0.2 amplitude:
3793 ladspa=file=cmt:noise_source_white:c=c0=.2
3797 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
3798 @code{C* Audio Plugin Suite} (CAPS) library:
3800 ladspa=file=caps:Click:c=c1=20'
3804 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
3806 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
3810 Increase volume by 20dB using fast lookahead limiter from Steve Harris
3811 @code{SWH Plugins} collection:
3813 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
3817 Attenuate low frequencies using Multiband EQ from Steve Harris
3818 @code{SWH Plugins} collection:
3820 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
3824 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
3827 ladspa=caps:Narrower
3831 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
3833 ladspa=caps:White:.2
3837 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
3839 ladspa=caps:Fractal:c=c1=1
3843 Dynamic volume normalization using @code{VLevel} plugin:
3845 ladspa=vlevel-ladspa:vlevel_mono
3849 @subsection Commands
3851 This filter supports the following commands:
3854 Modify the @var{N}-th control value.
3856 If the specified value is not valid, it is ignored and prior one is kept.
3861 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
3862 Support for both single pass (livestreams, files) and double pass (files) modes.
3863 This algorithm can target IL, LRA, and maximum true peak. To accurately detect true peaks,
3864 the audio stream will be upsampled to 192 kHz unless the normalization mode is linear.
3865 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
3867 The filter accepts the following options:
3871 Set integrated loudness target.
3872 Range is -70.0 - -5.0. Default value is -24.0.
3875 Set loudness range target.
3876 Range is 1.0 - 20.0. Default value is 7.0.
3879 Set maximum true peak.
3880 Range is -9.0 - +0.0. Default value is -2.0.
3882 @item measured_I, measured_i
3883 Measured IL of input file.
3884 Range is -99.0 - +0.0.
3886 @item measured_LRA, measured_lra
3887 Measured LRA of input file.
3888 Range is 0.0 - 99.0.
3890 @item measured_TP, measured_tp
3891 Measured true peak of input file.
3892 Range is -99.0 - +99.0.
3894 @item measured_thresh
3895 Measured threshold of input file.
3896 Range is -99.0 - +0.0.
3899 Set offset gain. Gain is applied before the true-peak limiter.
3900 Range is -99.0 - +99.0. Default is +0.0.
3903 Normalize linearly if possible.
3904 measured_I, measured_LRA, measured_TP, and measured_thresh must also
3905 to be specified in order to use this mode.
3906 Options are true or false. Default is true.
3909 Treat mono input files as "dual-mono". If a mono file is intended for playback
3910 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
3911 If set to @code{true}, this option will compensate for this effect.
3912 Multi-channel input files are not affected by this option.
3913 Options are true or false. Default is false.
3916 Set print format for stats. Options are summary, json, or none.
3917 Default value is none.
3922 Apply a low-pass filter with 3dB point frequency.
3923 The filter can be either single-pole or double-pole (the default).
3924 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3926 The filter accepts the following options:
3930 Set frequency in Hz. Default is 500.
3933 Set number of poles. Default is 2.
3936 Set method to specify band-width of filter.
3951 Specify the band-width of a filter in width_type units.
3952 Applies only to double-pole filter.
3953 The default is 0.707q and gives a Butterworth response.
3956 Specify which channels to filter, by default all available are filtered.
3959 @subsection Examples
3962 Lowpass only LFE channel, it LFE is not present it does nothing:
3968 @subsection Commands
3970 This filter supports the following commands:
3973 Change lowpass frequency.
3974 Syntax for the command is : "@var{frequency}"
3977 Change lowpass width_type.
3978 Syntax for the command is : "@var{width_type}"
3981 Change lowpass width.
3982 Syntax for the command is : "@var{width}"
3987 Load a LV2 (LADSPA Version 2) plugin.
3989 To enable compilation of this filter you need to configure FFmpeg with
3990 @code{--enable-lv2}.
3994 Specifies the plugin URI. You may need to escape ':'.
3997 Set the '|' separated list of controls which are zero or more floating point
3998 values that determine the behavior of the loaded plugin (for example delay,
4000 If @option{controls} is set to @code{help}, all available controls and
4001 their valid ranges are printed.
4003 @item sample_rate, s
4004 Specify the sample rate, default to 44100. Only used if plugin have
4008 Set the number of samples per channel per each output frame, default
4009 is 1024. Only used if plugin have zero inputs.
4012 Set the minimum duration of the sourced audio. See
4013 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4014 for the accepted syntax.
4015 Note that the resulting duration may be greater than the specified duration,
4016 as the generated audio is always cut at the end of a complete frame.
4017 If not specified, or the expressed duration is negative, the audio is
4018 supposed to be generated forever.
4019 Only used if plugin have zero inputs.
4022 @subsection Examples
4026 Apply bass enhancer plugin from Calf:
4028 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4032 Apply vinyl plugin from Calf:
4034 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4038 Apply bit crusher plugin from ArtyFX:
4040 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4045 Multiband Compress or expand the audio's dynamic range.
4047 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4048 This is akin to the crossover of a loudspeaker, and results in flat frequency
4049 response when absent compander action.
4051 It accepts the following parameters:
4055 This option syntax is:
4056 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4057 For explanation of each item refer to compand filter documentation.
4063 Mix channels with specific gain levels. The filter accepts the output
4064 channel layout followed by a set of channels definitions.
4066 This filter is also designed to efficiently remap the channels of an audio
4069 The filter accepts parameters of the form:
4070 "@var{l}|@var{outdef}|@var{outdef}|..."
4074 output channel layout or number of channels
4077 output channel specification, of the form:
4078 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4081 output channel to define, either a channel name (FL, FR, etc.) or a channel
4082 number (c0, c1, etc.)
4085 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4088 input channel to use, see out_name for details; it is not possible to mix
4089 named and numbered input channels
4092 If the `=' in a channel specification is replaced by `<', then the gains for
4093 that specification will be renormalized so that the total is 1, thus
4094 avoiding clipping noise.
4096 @subsection Mixing examples
4098 For example, if you want to down-mix from stereo to mono, but with a bigger
4099 factor for the left channel:
4101 pan=1c|c0=0.9*c0+0.1*c1
4104 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4105 7-channels surround:
4107 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4110 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4111 that should be preferred (see "-ac" option) unless you have very specific
4114 @subsection Remapping examples
4116 The channel remapping will be effective if, and only if:
4119 @item gain coefficients are zeroes or ones,
4120 @item only one input per channel output,
4123 If all these conditions are satisfied, the filter will notify the user ("Pure
4124 channel mapping detected"), and use an optimized and lossless method to do the
4127 For example, if you have a 5.1 source and want a stereo audio stream by
4128 dropping the extra channels:
4130 pan="stereo| c0=FL | c1=FR"
4133 Given the same source, you can also switch front left and front right channels
4134 and keep the input channel layout:
4136 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4139 If the input is a stereo audio stream, you can mute the front left channel (and
4140 still keep the stereo channel layout) with:
4145 Still with a stereo audio stream input, you can copy the right channel in both
4146 front left and right:
4148 pan="stereo| c0=FR | c1=FR"
4153 ReplayGain scanner filter. This filter takes an audio stream as an input and
4154 outputs it unchanged.
4155 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4159 Convert the audio sample format, sample rate and channel layout. It is
4160 not meant to be used directly.
4163 Apply time-stretching and pitch-shifting with librubberband.
4165 To enable compilation of this filter, you need to configure FFmpeg with
4166 @code{--enable-librubberband}.
4168 The filter accepts the following options:
4172 Set tempo scale factor.
4175 Set pitch scale factor.
4178 Set transients detector.
4179 Possible values are:
4188 Possible values are:
4197 Possible values are:
4204 Set processing window size.
4205 Possible values are:
4214 Possible values are:
4221 Enable formant preservation when shift pitching.
4222 Possible values are:
4230 Possible values are:
4239 Possible values are:
4246 @section sidechaincompress
4248 This filter acts like normal compressor but has the ability to compress
4249 detected signal using second input signal.
4250 It needs two input streams and returns one output stream.
4251 First input stream will be processed depending on second stream signal.
4252 The filtered signal then can be filtered with other filters in later stages of
4253 processing. See @ref{pan} and @ref{amerge} filter.
4255 The filter accepts the following options:
4259 Set input gain. Default is 1. Range is between 0.015625 and 64.
4262 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4263 Default is @code{downward}.
4266 If a signal of second stream raises above this level it will affect the gain
4267 reduction of first stream.
4268 By default is 0.125. Range is between 0.00097563 and 1.
4271 Set a ratio about which the signal is reduced. 1:2 means that if the level
4272 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4273 Default is 2. Range is between 1 and 20.
4276 Amount of milliseconds the signal has to rise above the threshold before gain
4277 reduction starts. Default is 20. Range is between 0.01 and 2000.
4280 Amount of milliseconds the signal has to fall below the threshold before
4281 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4284 Set the amount by how much signal will be amplified after processing.
4285 Default is 1. Range is from 1 to 64.
4288 Curve the sharp knee around the threshold to enter gain reduction more softly.
4289 Default is 2.82843. Range is between 1 and 8.
4292 Choose if the @code{average} level between all channels of side-chain stream
4293 or the louder(@code{maximum}) channel of side-chain stream affects the
4294 reduction. Default is @code{average}.
4297 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4298 of @code{rms}. Default is @code{rms} which is mainly smoother.
4301 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4304 How much to use compressed signal in output. Default is 1.
4305 Range is between 0 and 1.
4308 @subsection Examples
4312 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4313 depending on the signal of 2nd input and later compressed signal to be
4314 merged with 2nd input:
4316 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4320 @section sidechaingate
4322 A sidechain gate acts like a normal (wideband) gate but has the ability to
4323 filter the detected signal before sending it to the gain reduction stage.
4324 Normally a gate uses the full range signal to detect a level above the
4326 For example: If you cut all lower frequencies from your sidechain signal
4327 the gate will decrease the volume of your track only if not enough highs
4328 appear. With this technique you are able to reduce the resonation of a
4329 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4331 It needs two input streams and returns one output stream.
4332 First input stream will be processed depending on second stream signal.
4334 The filter accepts the following options:
4338 Set input level before filtering.
4339 Default is 1. Allowed range is from 0.015625 to 64.
4342 Set the mode of operation. Can be @code{upward} or @code{downward}.
4343 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
4344 will be amplified, expanding dynamic range in upward direction.
4345 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
4348 Set the level of gain reduction when the signal is below the threshold.
4349 Default is 0.06125. Allowed range is from 0 to 1.
4350 Setting this to 0 disables reduction and then filter behaves like expander.
4353 If a signal rises above this level the gain reduction is released.
4354 Default is 0.125. Allowed range is from 0 to 1.
4357 Set a ratio about which the signal is reduced.
4358 Default is 2. Allowed range is from 1 to 9000.
4361 Amount of milliseconds the signal has to rise above the threshold before gain
4363 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4366 Amount of milliseconds the signal has to fall below the threshold before the
4367 reduction is increased again. Default is 250 milliseconds.
4368 Allowed range is from 0.01 to 9000.
4371 Set amount of amplification of signal after processing.
4372 Default is 1. Allowed range is from 1 to 64.
4375 Curve the sharp knee around the threshold to enter gain reduction more softly.
4376 Default is 2.828427125. Allowed range is from 1 to 8.
4379 Choose if exact signal should be taken for detection or an RMS like one.
4380 Default is rms. Can be peak or rms.
4383 Choose if the average level between all channels or the louder channel affects
4385 Default is average. Can be average or maximum.
4388 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4391 @section silencedetect
4393 Detect silence in an audio stream.
4395 This filter logs a message when it detects that the input audio volume is less
4396 or equal to a noise tolerance value for a duration greater or equal to the
4397 minimum detected noise duration.
4399 The printed times and duration are expressed in seconds.
4401 The filter accepts the following options:
4405 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4406 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4409 Set silence duration until notification (default is 2 seconds).
4412 Process each channel separately, instead of combined. By default is disabled.
4415 @subsection Examples
4419 Detect 5 seconds of silence with -50dB noise tolerance:
4421 silencedetect=n=-50dB:d=5
4425 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4426 tolerance in @file{silence.mp3}:
4428 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4432 @section silenceremove
4434 Remove silence from the beginning, middle or end of the audio.
4436 The filter accepts the following options:
4440 This value is used to indicate if audio should be trimmed at beginning of
4441 the audio. A value of zero indicates no silence should be trimmed from the
4442 beginning. When specifying a non-zero value, it trims audio up until it
4443 finds non-silence. Normally, when trimming silence from beginning of audio
4444 the @var{start_periods} will be @code{1} but it can be increased to higher
4445 values to trim all audio up to specific count of non-silence periods.
4446 Default value is @code{0}.
4448 @item start_duration
4449 Specify the amount of time that non-silence must be detected before it stops
4450 trimming audio. By increasing the duration, bursts of noises can be treated
4451 as silence and trimmed off. Default value is @code{0}.
4453 @item start_threshold
4454 This indicates what sample value should be treated as silence. For digital
4455 audio, a value of @code{0} may be fine but for audio recorded from analog,
4456 you may wish to increase the value to account for background noise.
4457 Can be specified in dB (in case "dB" is appended to the specified value)
4458 or amplitude ratio. Default value is @code{0}.
4461 Specify max duration of silence at beginning that will be kept after
4462 trimming. Default is 0, which is equal to trimming all samples detected
4466 Specify mode of detection of silence end in start of multi-channel audio.
4467 Can be @var{any} or @var{all}. Default is @var{any}.
4468 With @var{any}, any sample that is detected as non-silence will cause
4469 stopped trimming of silence.
4470 With @var{all}, only if all channels are detected as non-silence will cause
4471 stopped trimming of silence.
4474 Set the count for trimming silence from the end of audio.
4475 To remove silence from the middle of a file, specify a @var{stop_periods}
4476 that is negative. This value is then treated as a positive value and is
4477 used to indicate the effect should restart processing as specified by
4478 @var{start_periods}, making it suitable for removing periods of silence
4479 in the middle of the audio.
4480 Default value is @code{0}.
4483 Specify a duration of silence that must exist before audio is not copied any
4484 more. By specifying a higher duration, silence that is wanted can be left in
4486 Default value is @code{0}.
4488 @item stop_threshold
4489 This is the same as @option{start_threshold} but for trimming silence from
4491 Can be specified in dB (in case "dB" is appended to the specified value)
4492 or amplitude ratio. Default value is @code{0}.
4495 Specify max duration of silence at end that will be kept after
4496 trimming. Default is 0, which is equal to trimming all samples detected
4500 Specify mode of detection of silence start in end of multi-channel audio.
4501 Can be @var{any} or @var{all}. Default is @var{any}.
4502 With @var{any}, any sample that is detected as non-silence will cause
4503 stopped trimming of silence.
4504 With @var{all}, only if all channels are detected as non-silence will cause
4505 stopped trimming of silence.
4508 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4509 and works better with digital silence which is exactly 0.
4510 Default value is @code{rms}.
4513 Set duration in number of seconds used to calculate size of window in number
4514 of samples for detecting silence.
4515 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4518 @subsection Examples
4522 The following example shows how this filter can be used to start a recording
4523 that does not contain the delay at the start which usually occurs between
4524 pressing the record button and the start of the performance:
4526 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
4530 Trim all silence encountered from beginning to end where there is more than 1
4531 second of silence in audio:
4533 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
4539 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4540 loudspeakers around the user for binaural listening via headphones (audio
4541 formats up to 9 channels supported).
4542 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4543 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4544 Austrian Academy of Sciences.
4546 To enable compilation of this filter you need to configure FFmpeg with
4547 @code{--enable-libmysofa}.
4549 The filter accepts the following options:
4553 Set the SOFA file used for rendering.
4556 Set gain applied to audio. Value is in dB. Default is 0.
4559 Set rotation of virtual loudspeakers in deg. Default is 0.
4562 Set elevation of virtual speakers in deg. Default is 0.
4565 Set distance in meters between loudspeakers and the listener with near-field
4566 HRTFs. Default is 1.
4569 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4570 processing audio in time domain which is slow.
4571 @var{freq} is processing audio in frequency domain which is fast.
4572 Default is @var{freq}.
4575 Set custom positions of virtual loudspeakers. Syntax for this option is:
4576 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4577 Each virtual loudspeaker is described with short channel name following with
4578 azimuth and elevation in degrees.
4579 Each virtual loudspeaker description is separated by '|'.
4580 For example to override front left and front right channel positions use:
4581 'speakers=FL 45 15|FR 345 15'.
4582 Descriptions with unrecognised channel names are ignored.
4585 Set custom gain for LFE channels. Value is in dB. Default is 0.
4588 Set custom frame size in number of samples. Default is 1024.
4589 Allowed range is from 1024 to 96000. Only used if option @samp{type}
4590 is set to @var{freq}.
4593 Should all IRs be normalized upon importing SOFA file.
4594 By default is enabled.
4597 Should nearest IRs be interpolated with neighbor IRs if exact position
4598 does not match. By default is disabled.
4601 Minphase all IRs upon loading of SOFA file. By default is disabled.
4604 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
4607 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
4610 @subsection Examples
4614 Using ClubFritz6 sofa file:
4616 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
4620 Using ClubFritz12 sofa file and bigger radius with small rotation:
4622 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
4626 Similar as above but with custom speaker positions for front left, front right, back left and back right
4627 and also with custom gain:
4629 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
4633 @section stereotools
4635 This filter has some handy utilities to manage stereo signals, for converting
4636 M/S stereo recordings to L/R signal while having control over the parameters
4637 or spreading the stereo image of master track.
4639 The filter accepts the following options:
4643 Set input level before filtering for both channels. Defaults is 1.
4644 Allowed range is from 0.015625 to 64.
4647 Set output level after filtering for both channels. Defaults is 1.
4648 Allowed range is from 0.015625 to 64.
4651 Set input balance between both channels. Default is 0.
4652 Allowed range is from -1 to 1.
4655 Set output balance between both channels. Default is 0.
4656 Allowed range is from -1 to 1.
4659 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
4660 clipping. Disabled by default.
4663 Mute the left channel. Disabled by default.
4666 Mute the right channel. Disabled by default.
4669 Change the phase of the left channel. Disabled by default.
4672 Change the phase of the right channel. Disabled by default.
4675 Set stereo mode. Available values are:
4679 Left/Right to Left/Right, this is default.
4682 Left/Right to Mid/Side.
4685 Mid/Side to Left/Right.
4688 Left/Right to Left/Left.
4691 Left/Right to Right/Right.
4694 Left/Right to Left + Right.
4697 Left/Right to Right/Left.
4700 Mid/Side to Left/Left.
4703 Mid/Side to Right/Right.
4707 Set level of side signal. Default is 1.
4708 Allowed range is from 0.015625 to 64.
4711 Set balance of side signal. Default is 0.
4712 Allowed range is from -1 to 1.
4715 Set level of the middle signal. Default is 1.
4716 Allowed range is from 0.015625 to 64.
4719 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
4722 Set stereo base between mono and inversed channels. Default is 0.
4723 Allowed range is from -1 to 1.
4726 Set delay in milliseconds how much to delay left from right channel and
4727 vice versa. Default is 0. Allowed range is from -20 to 20.
4730 Set S/C level. Default is 1. Allowed range is from 1 to 100.
4733 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
4735 @item bmode_in, bmode_out
4736 Set balance mode for balance_in/balance_out option.
4738 Can be one of the following:
4742 Classic balance mode. Attenuate one channel at time.
4743 Gain is raised up to 1.
4746 Similar as classic mode above but gain is raised up to 2.
4749 Equal power distribution, from -6dB to +6dB range.
4753 @subsection Examples
4757 Apply karaoke like effect:
4759 stereotools=mlev=0.015625
4763 Convert M/S signal to L/R:
4765 "stereotools=mode=ms>lr"
4769 @section stereowiden
4771 This filter enhance the stereo effect by suppressing signal common to both
4772 channels and by delaying the signal of left into right and vice versa,
4773 thereby widening the stereo effect.
4775 The filter accepts the following options:
4779 Time in milliseconds of the delay of left signal into right and vice versa.
4780 Default is 20 milliseconds.
4783 Amount of gain in delayed signal into right and vice versa. Gives a delay
4784 effect of left signal in right output and vice versa which gives widening
4785 effect. Default is 0.3.
4788 Cross feed of left into right with inverted phase. This helps in suppressing
4789 the mono. If the value is 1 it will cancel all the signal common to both
4790 channels. Default is 0.3.
4793 Set level of input signal of original channel. Default is 0.8.
4796 @section superequalizer
4797 Apply 18 band equalizer.
4799 The filter accepts the following options:
4806 Set 131Hz band gain.
4808 Set 185Hz band gain.
4810 Set 262Hz band gain.
4812 Set 370Hz band gain.
4814 Set 523Hz band gain.
4816 Set 740Hz band gain.
4818 Set 1047Hz band gain.
4820 Set 1480Hz band gain.
4822 Set 2093Hz band gain.
4824 Set 2960Hz band gain.
4826 Set 4186Hz band gain.
4828 Set 5920Hz band gain.
4830 Set 8372Hz band gain.
4832 Set 11840Hz band gain.
4834 Set 16744Hz band gain.
4836 Set 20000Hz band gain.
4840 Apply audio surround upmix filter.
4842 This filter allows to produce multichannel output from audio stream.
4844 The filter accepts the following options:
4848 Set output channel layout. By default, this is @var{5.1}.
4850 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4851 for the required syntax.
4854 Set input channel layout. By default, this is @var{stereo}.
4856 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4857 for the required syntax.
4860 Set input volume level. By default, this is @var{1}.
4863 Set output volume level. By default, this is @var{1}.
4866 Enable LFE channel output if output channel layout has it. By default, this is enabled.
4869 Set LFE low cut off frequency. By default, this is @var{128} Hz.
4872 Set LFE high cut off frequency. By default, this is @var{256} Hz.
4875 Set front center input volume. By default, this is @var{1}.
4878 Set front center output volume. By default, this is @var{1}.
4881 Set LFE input volume. By default, this is @var{1}.
4884 Set LFE output volume. By default, this is @var{1}.
4887 Set window function.
4889 It accepts the following values:
4912 Default is @code{hann}.
4915 Set window overlap. If set to 1, the recommended overlap for selected
4916 window function will be picked. Default is @code{0.5}.
4919 @section treble, highshelf
4921 Boost or cut treble (upper) frequencies of the audio using a two-pole
4922 shelving filter with a response similar to that of a standard
4923 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
4925 The filter accepts the following options:
4929 Give the gain at whichever is the lower of ~22 kHz and the
4930 Nyquist frequency. Its useful range is about -20 (for a large cut)
4931 to +20 (for a large boost). Beware of clipping when using a positive gain.
4934 Set the filter's central frequency and so can be used
4935 to extend or reduce the frequency range to be boosted or cut.
4936 The default value is @code{3000} Hz.
4939 Set method to specify band-width of filter.
4954 Determine how steep is the filter's shelf transition.
4957 Specify which channels to filter, by default all available are filtered.
4960 @subsection Commands
4962 This filter supports the following commands:
4965 Change treble frequency.
4966 Syntax for the command is : "@var{frequency}"
4969 Change treble width_type.
4970 Syntax for the command is : "@var{width_type}"
4973 Change treble width.
4974 Syntax for the command is : "@var{width}"
4978 Syntax for the command is : "@var{gain}"
4983 Sinusoidal amplitude modulation.
4985 The filter accepts the following options:
4989 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
4990 (20 Hz or lower) will result in a tremolo effect.
4991 This filter may also be used as a ring modulator by specifying
4992 a modulation frequency higher than 20 Hz.
4993 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
4996 Depth of modulation as a percentage. Range is 0.0 - 1.0.
4997 Default value is 0.5.
5002 Sinusoidal phase modulation.
5004 The filter accepts the following options:
5008 Modulation frequency in Hertz.
5009 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5012 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5013 Default value is 0.5.
5018 Adjust the input audio volume.
5020 It accepts the following parameters:
5024 Set audio volume expression.
5026 Output values are clipped to the maximum value.
5028 The output audio volume is given by the relation:
5030 @var{output_volume} = @var{volume} * @var{input_volume}
5033 The default value for @var{volume} is "1.0".
5036 This parameter represents the mathematical precision.
5038 It determines which input sample formats will be allowed, which affects the
5039 precision of the volume scaling.
5043 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5045 32-bit floating-point; this limits input sample format to FLT. (default)
5047 64-bit floating-point; this limits input sample format to DBL.
5051 Choose the behaviour on encountering ReplayGain side data in input frames.
5055 Remove ReplayGain side data, ignoring its contents (the default).
5058 Ignore ReplayGain side data, but leave it in the frame.
5061 Prefer the track gain, if present.
5064 Prefer the album gain, if present.
5067 @item replaygain_preamp
5068 Pre-amplification gain in dB to apply to the selected replaygain gain.
5070 Default value for @var{replaygain_preamp} is 0.0.
5073 Set when the volume expression is evaluated.
5075 It accepts the following values:
5078 only evaluate expression once during the filter initialization, or
5079 when the @samp{volume} command is sent
5082 evaluate expression for each incoming frame
5085 Default value is @samp{once}.
5088 The volume expression can contain the following parameters.
5092 frame number (starting at zero)
5095 @item nb_consumed_samples
5096 number of samples consumed by the filter
5098 number of samples in the current frame
5100 original frame position in the file
5106 PTS at start of stream
5108 time at start of stream
5114 last set volume value
5117 Note that when @option{eval} is set to @samp{once} only the
5118 @var{sample_rate} and @var{tb} variables are available, all other
5119 variables will evaluate to NAN.
5121 @subsection Commands
5123 This filter supports the following commands:
5126 Modify the volume expression.
5127 The command accepts the same syntax of the corresponding option.
5129 If the specified expression is not valid, it is kept at its current
5131 @item replaygain_noclip
5132 Prevent clipping by limiting the gain applied.
5134 Default value for @var{replaygain_noclip} is 1.
5138 @subsection Examples
5142 Halve the input audio volume:
5146 volume=volume=-6.0206dB
5149 In all the above example the named key for @option{volume} can be
5150 omitted, for example like in:
5156 Increase input audio power by 6 decibels using fixed-point precision:
5158 volume=volume=6dB:precision=fixed
5162 Fade volume after time 10 with an annihilation period of 5 seconds:
5164 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5168 @section volumedetect
5170 Detect the volume of the input video.
5172 The filter has no parameters. The input is not modified. Statistics about
5173 the volume will be printed in the log when the input stream end is reached.
5175 In particular it will show the mean volume (root mean square), maximum
5176 volume (on a per-sample basis), and the beginning of a histogram of the
5177 registered volume values (from the maximum value to a cumulated 1/1000 of
5180 All volumes are in decibels relative to the maximum PCM value.
5182 @subsection Examples
5184 Here is an excerpt of the output:
5186 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5187 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5188 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5189 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5190 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5191 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5192 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5193 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5194 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5200 The mean square energy is approximately -27 dB, or 10^-2.7.
5202 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5204 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5207 In other words, raising the volume by +4 dB does not cause any clipping,
5208 raising it by +5 dB causes clipping for 6 samples, etc.
5210 @c man end AUDIO FILTERS
5212 @chapter Audio Sources
5213 @c man begin AUDIO SOURCES
5215 Below is a description of the currently available audio sources.
5219 Buffer audio frames, and make them available to the filter chain.
5221 This source is mainly intended for a programmatic use, in particular
5222 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5224 It accepts the following parameters:
5228 The timebase which will be used for timestamps of submitted frames. It must be
5229 either a floating-point number or in @var{numerator}/@var{denominator} form.
5232 The sample rate of the incoming audio buffers.
5235 The sample format of the incoming audio buffers.
5236 Either a sample format name or its corresponding integer representation from
5237 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5239 @item channel_layout
5240 The channel layout of the incoming audio buffers.
5241 Either a channel layout name from channel_layout_map in
5242 @file{libavutil/channel_layout.c} or its corresponding integer representation
5243 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5246 The number of channels of the incoming audio buffers.
5247 If both @var{channels} and @var{channel_layout} are specified, then they
5252 @subsection Examples
5255 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5258 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5259 Since the sample format with name "s16p" corresponds to the number
5260 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5263 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5268 Generate an audio signal specified by an expression.
5270 This source accepts in input one or more expressions (one for each
5271 channel), which are evaluated and used to generate a corresponding
5274 This source accepts the following options:
5278 Set the '|'-separated expressions list for each separate channel. In case the
5279 @option{channel_layout} option is not specified, the selected channel layout
5280 depends on the number of provided expressions. Otherwise the last
5281 specified expression is applied to the remaining output channels.
5283 @item channel_layout, c
5284 Set the channel layout. The number of channels in the specified layout
5285 must be equal to the number of specified expressions.
5288 Set the minimum duration of the sourced audio. See
5289 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5290 for the accepted syntax.
5291 Note that the resulting duration may be greater than the specified
5292 duration, as the generated audio is always cut at the end of a
5295 If not specified, or the expressed duration is negative, the audio is
5296 supposed to be generated forever.
5299 Set the number of samples per channel per each output frame,
5302 @item sample_rate, s
5303 Specify the sample rate, default to 44100.
5306 Each expression in @var{exprs} can contain the following constants:
5310 number of the evaluated sample, starting from 0
5313 time of the evaluated sample expressed in seconds, starting from 0
5320 @subsection Examples
5330 Generate a sin signal with frequency of 440 Hz, set sample rate to
5333 aevalsrc="sin(440*2*PI*t):s=8000"
5337 Generate a two channels signal, specify the channel layout (Front
5338 Center + Back Center) explicitly:
5340 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5344 Generate white noise:
5346 aevalsrc="-2+random(0)"
5350 Generate an amplitude modulated signal:
5352 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5356 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5358 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5365 The null audio source, return unprocessed audio frames. It is mainly useful
5366 as a template and to be employed in analysis / debugging tools, or as
5367 the source for filters which ignore the input data (for example the sox
5370 This source accepts the following options:
5374 @item channel_layout, cl
5376 Specifies the channel layout, and can be either an integer or a string
5377 representing a channel layout. The default value of @var{channel_layout}
5380 Check the channel_layout_map definition in
5381 @file{libavutil/channel_layout.c} for the mapping between strings and
5382 channel layout values.
5384 @item sample_rate, r
5385 Specifies the sample rate, and defaults to 44100.
5388 Set the number of samples per requested frames.
5392 @subsection Examples
5396 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5398 anullsrc=r=48000:cl=4
5402 Do the same operation with a more obvious syntax:
5404 anullsrc=r=48000:cl=mono
5408 All the parameters need to be explicitly defined.
5412 Synthesize a voice utterance using the libflite library.
5414 To enable compilation of this filter you need to configure FFmpeg with
5415 @code{--enable-libflite}.
5417 Note that versions of the flite library prior to 2.0 are not thread-safe.
5419 The filter accepts the following options:
5424 If set to 1, list the names of the available voices and exit
5425 immediately. Default value is 0.
5428 Set the maximum number of samples per frame. Default value is 512.
5431 Set the filename containing the text to speak.
5434 Set the text to speak.
5437 Set the voice to use for the speech synthesis. Default value is
5438 @code{kal}. See also the @var{list_voices} option.
5441 @subsection Examples
5445 Read from file @file{speech.txt}, and synthesize the text using the
5446 standard flite voice:
5448 flite=textfile=speech.txt
5452 Read the specified text selecting the @code{slt} voice:
5454 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5458 Input text to ffmpeg:
5460 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5464 Make @file{ffplay} speak the specified text, using @code{flite} and
5465 the @code{lavfi} device:
5467 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
5471 For more information about libflite, check:
5472 @url{http://www.festvox.org/flite/}
5476 Generate a noise audio signal.
5478 The filter accepts the following options:
5481 @item sample_rate, r
5482 Specify the sample rate. Default value is 48000 Hz.
5485 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
5489 Specify the duration of the generated audio stream. Not specifying this option
5490 results in noise with an infinite length.
5492 @item color, colour, c
5493 Specify the color of noise. Available noise colors are white, pink, brown,
5494 blue and violet. Default color is white.
5497 Specify a value used to seed the PRNG.
5500 Set the number of samples per each output frame, default is 1024.
5503 @subsection Examples
5508 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
5510 anoisesrc=d=60:c=pink:r=44100:a=0.5
5516 Generate odd-tap Hilbert transform FIR coefficients.
5518 The resulting stream can be used with @ref{afir} filter for phase-shifting
5519 the signal by 90 degrees.
5521 This is used in many matrix coding schemes and for analytic signal generation.
5522 The process is often written as a multiplication by i (or j), the imaginary unit.
5524 The filter accepts the following options:
5528 @item sample_rate, s
5529 Set sample rate, default is 44100.
5532 Set length of FIR filter, default is 22051.
5535 Set number of samples per each frame.
5538 Set window function to be used when generating FIR coefficients.
5543 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
5545 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
5547 The filter accepts the following options:
5550 @item sample_rate, r
5551 Set sample rate, default is 44100.
5554 Set number of samples per each frame. Default is 1024.
5557 Set high-pass frequency. Default is 0.
5560 Set low-pass frequency. Default is 0.
5561 If high-pass frequency is lower than low-pass frequency and low-pass frequency
5562 is higher than 0 then filter will create band-pass filter coefficients,
5563 otherwise band-reject filter coefficients.
5566 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
5569 Set Kaiser window beta.
5572 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
5575 Enable rounding, by default is disabled.
5578 Set number of taps for high-pass filter.
5581 Set number of taps for low-pass filter.
5586 Generate an audio signal made of a sine wave with amplitude 1/8.
5588 The audio signal is bit-exact.
5590 The filter accepts the following options:
5595 Set the carrier frequency. Default is 440 Hz.
5597 @item beep_factor, b
5598 Enable a periodic beep every second with frequency @var{beep_factor} times
5599 the carrier frequency. Default is 0, meaning the beep is disabled.
5601 @item sample_rate, r
5602 Specify the sample rate, default is 44100.
5605 Specify the duration of the generated audio stream.
5607 @item samples_per_frame
5608 Set the number of samples per output frame.
5610 The expression can contain the following constants:
5614 The (sequential) number of the output audio frame, starting from 0.
5617 The PTS (Presentation TimeStamp) of the output audio frame,
5618 expressed in @var{TB} units.
5621 The PTS of the output audio frame, expressed in seconds.
5624 The timebase of the output audio frames.
5627 Default is @code{1024}.
5630 @subsection Examples
5635 Generate a simple 440 Hz sine wave:
5641 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
5645 sine=frequency=220:beep_factor=4:duration=5
5649 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
5652 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
5656 @c man end AUDIO SOURCES
5658 @chapter Audio Sinks
5659 @c man begin AUDIO SINKS
5661 Below is a description of the currently available audio sinks.
5663 @section abuffersink
5665 Buffer audio frames, and make them available to the end of filter chain.
5667 This sink is mainly intended for programmatic use, in particular
5668 through the interface defined in @file{libavfilter/buffersink.h}
5669 or the options system.
5671 It accepts a pointer to an AVABufferSinkContext structure, which
5672 defines the incoming buffers' formats, to be passed as the opaque
5673 parameter to @code{avfilter_init_filter} for initialization.
5676 Null audio sink; do absolutely nothing with the input audio. It is
5677 mainly useful as a template and for use in analysis / debugging
5680 @c man end AUDIO SINKS
5682 @chapter Video Filters
5683 @c man begin VIDEO FILTERS
5685 When you configure your FFmpeg build, you can disable any of the
5686 existing filters using @code{--disable-filters}.
5687 The configure output will show the video filters included in your
5690 Below is a description of the currently available video filters.
5692 @section alphaextract
5694 Extract the alpha component from the input as a grayscale video. This
5695 is especially useful with the @var{alphamerge} filter.
5699 Add or replace the alpha component of the primary input with the
5700 grayscale value of a second input. This is intended for use with
5701 @var{alphaextract} to allow the transmission or storage of frame
5702 sequences that have alpha in a format that doesn't support an alpha
5705 For example, to reconstruct full frames from a normal YUV-encoded video
5706 and a separate video created with @var{alphaextract}, you might use:
5708 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
5711 Since this filter is designed for reconstruction, it operates on frame
5712 sequences without considering timestamps, and terminates when either
5713 input reaches end of stream. This will cause problems if your encoding
5714 pipeline drops frames. If you're trying to apply an image as an
5715 overlay to a video stream, consider the @var{overlay} filter instead.
5719 Amplify differences between current pixel and pixels of adjacent frames in
5720 same pixel location.
5722 This filter accepts the following options:
5726 Set frame radius. Default is 2. Allowed range is from 1 to 63.
5727 For example radius of 3 will instruct filter to calculate average of 7 frames.
5730 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
5733 Set threshold for difference amplification. Any difference greater or equal to
5734 this value will not alter source pixel. Default is 10.
5735 Allowed range is from 0 to 65535.
5738 Set tolerance for difference amplification. Any difference lower to
5739 this value will not alter source pixel. Default is 0.
5740 Allowed range is from 0 to 65535.
5743 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
5744 This option controls maximum possible value that will decrease source pixel value.
5747 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
5748 This option controls maximum possible value that will increase source pixel value.
5751 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
5756 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
5757 and libavformat to work. On the other hand, it is limited to ASS (Advanced
5758 Substation Alpha) subtitles files.
5760 This filter accepts the following option in addition to the common options from
5761 the @ref{subtitles} filter:
5765 Set the shaping engine
5767 Available values are:
5770 The default libass shaping engine, which is the best available.
5772 Fast, font-agnostic shaper that can do only substitutions
5774 Slower shaper using OpenType for substitutions and positioning
5777 The default is @code{auto}.
5781 Apply an Adaptive Temporal Averaging Denoiser to the video input.
5783 The filter accepts the following options:
5787 Set threshold A for 1st plane. Default is 0.02.
5788 Valid range is 0 to 0.3.
5791 Set threshold B for 1st plane. Default is 0.04.
5792 Valid range is 0 to 5.
5795 Set threshold A for 2nd plane. Default is 0.02.
5796 Valid range is 0 to 0.3.
5799 Set threshold B for 2nd plane. Default is 0.04.
5800 Valid range is 0 to 5.
5803 Set threshold A for 3rd plane. Default is 0.02.
5804 Valid range is 0 to 0.3.
5807 Set threshold B for 3rd plane. Default is 0.04.
5808 Valid range is 0 to 5.
5810 Threshold A is designed to react on abrupt changes in the input signal and
5811 threshold B is designed to react on continuous changes in the input signal.
5814 Set number of frames filter will use for averaging. Default is 9. Must be odd
5815 number in range [5, 129].
5818 Set what planes of frame filter will use for averaging. Default is all.
5823 Apply average blur filter.
5825 The filter accepts the following options:
5829 Set horizontal radius size.
5832 Set which planes to filter. By default all planes are filtered.
5835 Set vertical radius size, if zero it will be same as @code{sizeX}.
5836 Default is @code{0}.
5841 Compute the bounding box for the non-black pixels in the input frame
5844 This filter computes the bounding box containing all the pixels with a
5845 luminance value greater than the minimum allowed value.
5846 The parameters describing the bounding box are printed on the filter
5849 The filter accepts the following option:
5853 Set the minimal luminance value. Default is @code{16}.
5856 @section bitplanenoise
5858 Show and measure bit plane noise.
5860 The filter accepts the following options:
5864 Set which plane to analyze. Default is @code{1}.
5867 Filter out noisy pixels from @code{bitplane} set above.
5868 Default is disabled.
5871 @section blackdetect
5873 Detect video intervals that are (almost) completely black. Can be
5874 useful to detect chapter transitions, commercials, or invalid
5875 recordings. Output lines contains the time for the start, end and
5876 duration of the detected black interval expressed in seconds.
5878 In order to display the output lines, you need to set the loglevel at
5879 least to the AV_LOG_INFO value.
5881 The filter accepts the following options:
5884 @item black_min_duration, d
5885 Set the minimum detected black duration expressed in seconds. It must
5886 be a non-negative floating point number.
5888 Default value is 2.0.
5890 @item picture_black_ratio_th, pic_th
5891 Set the threshold for considering a picture "black".
5892 Express the minimum value for the ratio:
5894 @var{nb_black_pixels} / @var{nb_pixels}
5897 for which a picture is considered black.
5898 Default value is 0.98.
5900 @item pixel_black_th, pix_th
5901 Set the threshold for considering a pixel "black".
5903 The threshold expresses the maximum pixel luminance value for which a
5904 pixel is considered "black". The provided value is scaled according to
5905 the following equation:
5907 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
5910 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
5911 the input video format, the range is [0-255] for YUV full-range
5912 formats and [16-235] for YUV non full-range formats.
5914 Default value is 0.10.
5917 The following example sets the maximum pixel threshold to the minimum
5918 value, and detects only black intervals of 2 or more seconds:
5920 blackdetect=d=2:pix_th=0.00
5925 Detect frames that are (almost) completely black. Can be useful to
5926 detect chapter transitions or commercials. Output lines consist of
5927 the frame number of the detected frame, the percentage of blackness,
5928 the position in the file if known or -1 and the timestamp in seconds.
5930 In order to display the output lines, you need to set the loglevel at
5931 least to the AV_LOG_INFO value.
5933 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
5934 The value represents the percentage of pixels in the picture that
5935 are below the threshold value.
5937 It accepts the following parameters:
5942 The percentage of the pixels that have to be below the threshold; it defaults to
5945 @item threshold, thresh
5946 The threshold below which a pixel value is considered black; it defaults to
5951 @section blend, tblend
5953 Blend two video frames into each other.
5955 The @code{blend} filter takes two input streams and outputs one
5956 stream, the first input is the "top" layer and second input is
5957 "bottom" layer. By default, the output terminates when the longest input terminates.
5959 The @code{tblend} (time blend) filter takes two consecutive frames
5960 from one single stream, and outputs the result obtained by blending
5961 the new frame on top of the old frame.
5963 A description of the accepted options follows.
5971 Set blend mode for specific pixel component or all pixel components in case
5972 of @var{all_mode}. Default value is @code{normal}.
5974 Available values for component modes are:
6016 Set blend opacity for specific pixel component or all pixel components in case
6017 of @var{all_opacity}. Only used in combination with pixel component blend modes.
6024 Set blend expression for specific pixel component or all pixel components in case
6025 of @var{all_expr}. Note that related mode options will be ignored if those are set.
6027 The expressions can use the following variables:
6031 The sequential number of the filtered frame, starting from @code{0}.
6035 the coordinates of the current sample
6039 the width and height of currently filtered plane
6043 Width and height scale for the plane being filtered. It is the
6044 ratio between the dimensions of the current plane to the luma plane,
6045 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
6046 the luma plane and @code{0.5,0.5} for the chroma planes.
6049 Time of the current frame, expressed in seconds.
6052 Value of pixel component at current location for first video frame (top layer).
6055 Value of pixel component at current location for second video frame (bottom layer).
6059 The @code{blend} filter also supports the @ref{framesync} options.
6061 @subsection Examples
6065 Apply transition from bottom layer to top layer in first 10 seconds:
6067 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
6071 Apply linear horizontal transition from top layer to bottom layer:
6073 blend=all_expr='A*(X/W)+B*(1-X/W)'
6077 Apply 1x1 checkerboard effect:
6079 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
6083 Apply uncover left effect:
6085 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6089 Apply uncover down effect:
6091 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6095 Apply uncover up-left effect:
6097 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6101 Split diagonally video and shows top and bottom layer on each side:
6103 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6107 Display differences between the current and the previous frame:
6109 tblend=all_mode=grainextract
6115 Denoise frames using Block-Matching 3D algorithm.
6117 The filter accepts the following options.
6121 Set denoising strength. Default value is 1.
6122 Allowed range is from 0 to 999.9.
6123 The denoising algorithm is very sensitive to sigma, so adjust it
6124 according to the source.
6127 Set local patch size. This sets dimensions in 2D.
6130 Set sliding step for processing blocks. Default value is 4.
6131 Allowed range is from 1 to 64.
6132 Smaller values allows processing more reference blocks and is slower.
6135 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6136 When set to 1, no block matching is done. Larger values allows more blocks
6138 Allowed range is from 1 to 256.
6141 Set radius for search block matching. Default is 9.
6142 Allowed range is from 1 to INT32_MAX.
6145 Set step between two search locations for block matching. Default is 1.
6146 Allowed range is from 1 to 64. Smaller is slower.
6149 Set threshold of mean square error for block matching. Valid range is 0 to
6153 Set thresholding parameter for hard thresholding in 3D transformed domain.
6154 Larger values results in stronger hard-thresholding filtering in frequency
6158 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6159 Default is @code{basic}.
6162 If enabled, filter will use 2nd stream for block matching.
6163 Default is disabled for @code{basic} value of @var{estim} option,
6164 and always enabled if value of @var{estim} is @code{final}.
6167 Set planes to filter. Default is all available except alpha.
6170 @subsection Examples
6174 Basic filtering with bm3d:
6176 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
6180 Same as above, but filtering only luma:
6182 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
6186 Same as above, but with both estimation modes:
6188 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
6192 Same as above, but prefilter with @ref{nlmeans} filter instead:
6194 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
6200 Apply a boxblur algorithm to the input video.
6202 It accepts the following parameters:
6206 @item luma_radius, lr
6207 @item luma_power, lp
6208 @item chroma_radius, cr
6209 @item chroma_power, cp
6210 @item alpha_radius, ar
6211 @item alpha_power, ap
6215 A description of the accepted options follows.
6218 @item luma_radius, lr
6219 @item chroma_radius, cr
6220 @item alpha_radius, ar
6221 Set an expression for the box radius in pixels used for blurring the
6222 corresponding input plane.
6224 The radius value must be a non-negative number, and must not be
6225 greater than the value of the expression @code{min(w,h)/2} for the
6226 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
6229 Default value for @option{luma_radius} is "2". If not specified,
6230 @option{chroma_radius} and @option{alpha_radius} default to the
6231 corresponding value set for @option{luma_radius}.
6233 The expressions can contain the following constants:
6237 The input width and height in pixels.
6241 The input chroma image width and height in pixels.
6245 The horizontal and vertical chroma subsample values. For example, for the
6246 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6249 @item luma_power, lp
6250 @item chroma_power, cp
6251 @item alpha_power, ap
6252 Specify how many times the boxblur filter is applied to the
6253 corresponding plane.
6255 Default value for @option{luma_power} is 2. If not specified,
6256 @option{chroma_power} and @option{alpha_power} default to the
6257 corresponding value set for @option{luma_power}.
6259 A value of 0 will disable the effect.
6262 @subsection Examples
6266 Apply a boxblur filter with the luma, chroma, and alpha radii
6269 boxblur=luma_radius=2:luma_power=1
6274 Set the luma radius to 2, and alpha and chroma radius to 0:
6276 boxblur=2:1:cr=0:ar=0
6280 Set the luma and chroma radii to a fraction of the video dimension:
6282 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6288 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6289 Deinterlacing Filter").
6291 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6292 interpolation algorithms.
6293 It accepts the following parameters:
6297 The interlacing mode to adopt. It accepts one of the following values:
6301 Output one frame for each frame.
6303 Output one frame for each field.
6306 The default value is @code{send_field}.
6309 The picture field parity assumed for the input interlaced video. It accepts one
6310 of the following values:
6314 Assume the top field is first.
6316 Assume the bottom field is first.
6318 Enable automatic detection of field parity.
6321 The default value is @code{auto}.
6322 If the interlacing is unknown or the decoder does not export this information,
6323 top field first will be assumed.
6326 Specify which frames to deinterlace. Accept one of the following
6331 Deinterlace all frames.
6333 Only deinterlace frames marked as interlaced.
6336 The default value is @code{all}.
6340 Remove all color information for all colors except for certain one.
6342 The filter accepts the following options:
6346 The color which will not be replaced with neutral chroma.
6349 Similarity percentage with the above color.
6350 0.01 matches only the exact key color, while 1.0 matches everything.
6353 Signals that the color passed is already in YUV instead of RGB.
6355 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6356 This can be used to pass exact YUV values as hexadecimal numbers.
6360 YUV colorspace color/chroma keying.
6362 The filter accepts the following options:
6366 The color which will be replaced with transparency.
6369 Similarity percentage with the key color.
6371 0.01 matches only the exact key color, while 1.0 matches everything.
6376 0.0 makes pixels either fully transparent, or not transparent at all.
6378 Higher values result in semi-transparent pixels, with a higher transparency
6379 the more similar the pixels color is to the key color.
6382 Signals that the color passed is already in YUV instead of RGB.
6384 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6385 This can be used to pass exact YUV values as hexadecimal numbers.
6388 @subsection Examples
6392 Make every green pixel in the input image transparent:
6394 ffmpeg -i input.png -vf chromakey=green out.png
6398 Overlay a greenscreen-video on top of a static black background.
6400 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
6404 @section chromashift
6405 Shift chroma pixels horizontally and/or vertically.
6407 The filter accepts the following options:
6410 Set amount to shift chroma-blue horizontally.
6412 Set amount to shift chroma-blue vertically.
6414 Set amount to shift chroma-red horizontally.
6416 Set amount to shift chroma-red vertically.
6418 Set edge mode, can be @var{smear}, default, or @var{warp}.
6423 Display CIE color diagram with pixels overlaid onto it.
6425 The filter accepts the following options:
6440 @item uhdtv, rec2020
6453 Set what gamuts to draw.
6455 See @code{system} option for available values.
6458 Set ciescope size, by default set to 512.
6461 Set intensity used to map input pixel values to CIE diagram.
6464 Set contrast used to draw tongue colors that are out of active color system gamut.
6467 Correct gamma displayed on scope, by default enabled.
6470 Show white point on CIE diagram, by default disabled.
6473 Set input gamma. Used only with XYZ input color space.
6478 Visualize information exported by some codecs.
6480 Some codecs can export information through frames using side-data or other
6481 means. For example, some MPEG based codecs export motion vectors through the
6482 @var{export_mvs} flag in the codec @option{flags2} option.
6484 The filter accepts the following option:
6488 Set motion vectors to visualize.
6490 Available flags for @var{mv} are:
6494 forward predicted MVs of P-frames
6496 forward predicted MVs of B-frames
6498 backward predicted MVs of B-frames
6502 Display quantization parameters using the chroma planes.
6505 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
6507 Available flags for @var{mv_type} are:
6511 forward predicted MVs
6513 backward predicted MVs
6516 @item frame_type, ft
6517 Set frame type to visualize motion vectors of.
6519 Available flags for @var{frame_type} are:
6523 intra-coded frames (I-frames)
6525 predicted frames (P-frames)
6527 bi-directionally predicted frames (B-frames)
6531 @subsection Examples
6535 Visualize forward predicted MVs of all frames using @command{ffplay}:
6537 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
6541 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
6543 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
6547 @section colorbalance
6548 Modify intensity of primary colors (red, green and blue) of input frames.
6550 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
6551 regions for the red-cyan, green-magenta or blue-yellow balance.
6553 A positive adjustment value shifts the balance towards the primary color, a negative
6554 value towards the complementary color.
6556 The filter accepts the following options:
6562 Adjust red, green and blue shadows (darkest pixels).
6567 Adjust red, green and blue midtones (medium pixels).
6572 Adjust red, green and blue highlights (brightest pixels).
6574 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
6577 @subsection Examples
6581 Add red color cast to shadows:
6588 RGB colorspace color keying.
6590 The filter accepts the following options:
6594 The color which will be replaced with transparency.
6597 Similarity percentage with the key color.
6599 0.01 matches only the exact key color, while 1.0 matches everything.
6604 0.0 makes pixels either fully transparent, or not transparent at all.
6606 Higher values result in semi-transparent pixels, with a higher transparency
6607 the more similar the pixels color is to the key color.
6610 @subsection Examples
6614 Make every green pixel in the input image transparent:
6616 ffmpeg -i input.png -vf colorkey=green out.png
6620 Overlay a greenscreen-video on top of a static background image.
6622 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
6626 @section colorlevels
6628 Adjust video input frames using levels.
6630 The filter accepts the following options:
6637 Adjust red, green, blue and alpha input black point.
6638 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
6644 Adjust red, green, blue and alpha input white point.
6645 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
6647 Input levels are used to lighten highlights (bright tones), darken shadows
6648 (dark tones), change the balance of bright and dark tones.
6654 Adjust red, green, blue and alpha output black point.
6655 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
6661 Adjust red, green, blue and alpha output white point.
6662 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
6664 Output levels allows manual selection of a constrained output level range.
6667 @subsection Examples
6671 Make video output darker:
6673 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
6679 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
6683 Make video output lighter:
6685 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
6689 Increase brightness:
6691 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
6695 @section colorchannelmixer
6697 Adjust video input frames by re-mixing color channels.
6699 This filter modifies a color channel by adding the values associated to
6700 the other channels of the same pixels. For example if the value to
6701 modify is red, the output value will be:
6703 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
6706 The filter accepts the following options:
6713 Adjust contribution of input red, green, blue and alpha channels for output red channel.
6714 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
6720 Adjust contribution of input red, green, blue and alpha channels for output green channel.
6721 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
6727 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
6728 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
6734 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
6735 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
6737 Allowed ranges for options are @code{[-2.0, 2.0]}.
6740 @subsection Examples
6744 Convert source to grayscale:
6746 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
6749 Simulate sepia tones:
6751 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
6755 @section colormatrix
6757 Convert color matrix.
6759 The filter accepts the following options:
6764 Specify the source and destination color matrix. Both values must be
6767 The accepted values are:
6795 For example to convert from BT.601 to SMPTE-240M, use the command:
6797 colormatrix=bt601:smpte240m
6802 Convert colorspace, transfer characteristics or color primaries.
6803 Input video needs to have an even size.
6805 The filter accepts the following options:
6810 Specify all color properties at once.
6812 The accepted values are:
6842 Specify output colorspace.
6844 The accepted values are:
6853 BT.470BG or BT.601-6 625
6856 SMPTE-170M or BT.601-6 525
6865 BT.2020 with non-constant luminance
6871 Specify output transfer characteristics.
6873 The accepted values are:
6885 Constant gamma of 2.2
6888 Constant gamma of 2.8
6891 SMPTE-170M, BT.601-6 625 or BT.601-6 525
6909 BT.2020 for 10-bits content
6912 BT.2020 for 12-bits content
6918 Specify output color primaries.
6920 The accepted values are:
6929 BT.470BG or BT.601-6 625
6932 SMPTE-170M or BT.601-6 525
6956 Specify output color range.
6958 The accepted values are:
6961 TV (restricted) range
6964 MPEG (restricted) range
6975 Specify output color format.
6977 The accepted values are:
6980 YUV 4:2:0 planar 8-bits
6983 YUV 4:2:0 planar 10-bits
6986 YUV 4:2:0 planar 12-bits
6989 YUV 4:2:2 planar 8-bits
6992 YUV 4:2:2 planar 10-bits
6995 YUV 4:2:2 planar 12-bits
6998 YUV 4:4:4 planar 8-bits
7001 YUV 4:4:4 planar 10-bits
7004 YUV 4:4:4 planar 12-bits
7009 Do a fast conversion, which skips gamma/primary correction. This will take
7010 significantly less CPU, but will be mathematically incorrect. To get output
7011 compatible with that produced by the colormatrix filter, use fast=1.
7014 Specify dithering mode.
7016 The accepted values are:
7022 Floyd-Steinberg dithering
7026 Whitepoint adaptation mode.
7028 The accepted values are:
7031 Bradford whitepoint adaptation
7034 von Kries whitepoint adaptation
7037 identity whitepoint adaptation (i.e. no whitepoint adaptation)
7041 Override all input properties at once. Same accepted values as @ref{all}.
7044 Override input colorspace. Same accepted values as @ref{space}.
7047 Override input color primaries. Same accepted values as @ref{primaries}.
7050 Override input transfer characteristics. Same accepted values as @ref{trc}.
7053 Override input color range. Same accepted values as @ref{range}.
7057 The filter converts the transfer characteristics, color space and color
7058 primaries to the specified user values. The output value, if not specified,
7059 is set to a default value based on the "all" property. If that property is
7060 also not specified, the filter will log an error. The output color range and
7061 format default to the same value as the input color range and format. The
7062 input transfer characteristics, color space, color primaries and color range
7063 should be set on the input data. If any of these are missing, the filter will
7064 log an error and no conversion will take place.
7066 For example to convert the input to SMPTE-240M, use the command:
7068 colorspace=smpte240m
7071 @section convolution
7073 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
7075 The filter accepts the following options:
7082 Set matrix for each plane.
7083 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
7084 and from 1 to 49 odd number of signed integers in @var{row} mode.
7090 Set multiplier for calculated value for each plane.
7091 If unset or 0, it will be sum of all matrix elements.
7097 Set bias for each plane. This value is added to the result of the multiplication.
7098 Useful for making the overall image brighter or darker. Default is 0.0.
7104 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
7105 Default is @var{square}.
7108 @subsection Examples
7114 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"
7120 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"
7126 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"
7132 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"
7136 Apply laplacian edge detector which includes diagonals:
7138 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"
7144 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"
7150 Apply 2D convolution of video stream in frequency domain using second stream
7153 The filter accepts the following options:
7157 Set which planes to process.
7160 Set which impulse video frames will be processed, can be @var{first}
7161 or @var{all}. Default is @var{all}.
7164 The @code{convolve} filter also supports the @ref{framesync} options.
7168 Copy the input video source unchanged to the output. This is mainly useful for
7173 Video filtering on GPU using Apple's CoreImage API on OSX.
7175 Hardware acceleration is based on an OpenGL context. Usually, this means it is
7176 processed by video hardware. However, software-based OpenGL implementations
7177 exist which means there is no guarantee for hardware processing. It depends on
7180 There are many filters and image generators provided by Apple that come with a
7181 large variety of options. The filter has to be referenced by its name along
7184 The coreimage filter accepts the following options:
7187 List all available filters and generators along with all their respective
7188 options as well as possible minimum and maximum values along with the default
7195 Specify all filters by their respective name and options.
7196 Use @var{list_filters} to determine all valid filter names and options.
7197 Numerical options are specified by a float value and are automatically clamped
7198 to their respective value range. Vector and color options have to be specified
7199 by a list of space separated float values. Character escaping has to be done.
7200 A special option name @code{default} is available to use default options for a
7203 It is required to specify either @code{default} or at least one of the filter options.
7204 All omitted options are used with their default values.
7205 The syntax of the filter string is as follows:
7207 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
7211 Specify a rectangle where the output of the filter chain is copied into the
7212 input image. It is given by a list of space separated float values:
7214 output_rect=x\ y\ width\ height
7216 If not given, the output rectangle equals the dimensions of the input image.
7217 The output rectangle is automatically cropped at the borders of the input
7218 image. Negative values are valid for each component.
7220 output_rect=25\ 25\ 100\ 100
7224 Several filters can be chained for successive processing without GPU-HOST
7225 transfers allowing for fast processing of complex filter chains.
7226 Currently, only filters with zero (generators) or exactly one (filters) input
7227 image and one output image are supported. Also, transition filters are not yet
7230 Some filters generate output images with additional padding depending on the
7231 respective filter kernel. The padding is automatically removed to ensure the
7232 filter output has the same size as the input image.
7234 For image generators, the size of the output image is determined by the
7235 previous output image of the filter chain or the input image of the whole
7236 filterchain, respectively. The generators do not use the pixel information of
7237 this image to generate their output. However, the generated output is
7238 blended onto this image, resulting in partial or complete coverage of the
7241 The @ref{coreimagesrc} video source can be used for generating input images
7242 which are directly fed into the filter chain. By using it, providing input
7243 images by another video source or an input video is not required.
7245 @subsection Examples
7250 List all filters available:
7252 coreimage=list_filters=true
7256 Use the CIBoxBlur filter with default options to blur an image:
7258 coreimage=filter=CIBoxBlur@@default
7262 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
7263 its center at 100x100 and a radius of 50 pixels:
7265 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
7269 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
7270 given as complete and escaped command-line for Apple's standard bash shell:
7272 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
7278 Crop the input video to given dimensions.
7280 It accepts the following parameters:
7284 The width of the output video. It defaults to @code{iw}.
7285 This expression is evaluated only once during the filter
7286 configuration, or when the @samp{w} or @samp{out_w} command is sent.
7289 The height of the output video. It defaults to @code{ih}.
7290 This expression is evaluated only once during the filter
7291 configuration, or when the @samp{h} or @samp{out_h} command is sent.
7294 The horizontal position, in the input video, of the left edge of the output
7295 video. It defaults to @code{(in_w-out_w)/2}.
7296 This expression is evaluated per-frame.
7299 The vertical position, in the input video, of the top edge of the output video.
7300 It defaults to @code{(in_h-out_h)/2}.
7301 This expression is evaluated per-frame.
7304 If set to 1 will force the output display aspect ratio
7305 to be the same of the input, by changing the output sample aspect
7306 ratio. It defaults to 0.
7309 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
7310 width/height/x/y as specified and will not be rounded to nearest smaller value.
7314 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
7315 expressions containing the following constants:
7320 The computed values for @var{x} and @var{y}. They are evaluated for
7325 The input width and height.
7329 These are the same as @var{in_w} and @var{in_h}.
7333 The output (cropped) width and height.
7337 These are the same as @var{out_w} and @var{out_h}.
7340 same as @var{iw} / @var{ih}
7343 input sample aspect ratio
7346 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
7350 horizontal and vertical chroma subsample values. For example for the
7351 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
7354 The number of the input frame, starting from 0.
7357 the position in the file of the input frame, NAN if unknown
7360 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
7364 The expression for @var{out_w} may depend on the value of @var{out_h},
7365 and the expression for @var{out_h} may depend on @var{out_w}, but they
7366 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
7367 evaluated after @var{out_w} and @var{out_h}.
7369 The @var{x} and @var{y} parameters specify the expressions for the
7370 position of the top-left corner of the output (non-cropped) area. They
7371 are evaluated for each frame. If the evaluated value is not valid, it
7372 is approximated to the nearest valid value.
7374 The expression for @var{x} may depend on @var{y}, and the expression
7375 for @var{y} may depend on @var{x}.
7377 @subsection Examples
7381 Crop area with size 100x100 at position (12,34).
7386 Using named options, the example above becomes:
7388 crop=w=100:h=100:x=12:y=34
7392 Crop the central input area with size 100x100:
7398 Crop the central input area with size 2/3 of the input video:
7400 crop=2/3*in_w:2/3*in_h
7404 Crop the input video central square:
7411 Delimit the rectangle with the top-left corner placed at position
7412 100:100 and the right-bottom corner corresponding to the right-bottom
7413 corner of the input image.
7415 crop=in_w-100:in_h-100:100:100
7419 Crop 10 pixels from the left and right borders, and 20 pixels from
7420 the top and bottom borders
7422 crop=in_w-2*10:in_h-2*20
7426 Keep only the bottom right quarter of the input image:
7428 crop=in_w/2:in_h/2:in_w/2:in_h/2
7432 Crop height for getting Greek harmony:
7434 crop=in_w:1/PHI*in_w
7438 Apply trembling effect:
7440 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)
7444 Apply erratic camera effect depending on timestamp:
7446 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)"
7450 Set x depending on the value of y:
7452 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
7456 @subsection Commands
7458 This filter supports the following commands:
7464 Set width/height of the output video and the horizontal/vertical position
7466 The command accepts the same syntax of the corresponding option.
7468 If the specified expression is not valid, it is kept at its current
7474 Auto-detect the crop size.
7476 It calculates the necessary cropping parameters and prints the
7477 recommended parameters via the logging system. The detected dimensions
7478 correspond to the non-black area of the input video.
7480 It accepts the following parameters:
7485 Set higher black value threshold, which can be optionally specified
7486 from nothing (0) to everything (255 for 8-bit based formats). An intensity
7487 value greater to the set value is considered non-black. It defaults to 24.
7488 You can also specify a value between 0.0 and 1.0 which will be scaled depending
7489 on the bitdepth of the pixel format.
7492 The value which the width/height should be divisible by. It defaults to
7493 16. The offset is automatically adjusted to center the video. Use 2 to
7494 get only even dimensions (needed for 4:2:2 video). 16 is best when
7495 encoding to most video codecs.
7497 @item reset_count, reset
7498 Set the counter that determines after how many frames cropdetect will
7499 reset the previously detected largest video area and start over to
7500 detect the current optimal crop area. Default value is 0.
7502 This can be useful when channel logos distort the video area. 0
7503 indicates 'never reset', and returns the largest area encountered during
7510 Delay video filtering until a given wallclock timestamp. The filter first
7511 passes on @option{preroll} amount of frames, then it buffers at most
7512 @option{buffer} amount of frames and waits for the cue. After reaching the cue
7513 it forwards the buffered frames and also any subsequent frames coming in its
7516 The filter can be used synchronize the output of multiple ffmpeg processes for
7517 realtime output devices like decklink. By putting the delay in the filtering
7518 chain and pre-buffering frames the process can pass on data to output almost
7519 immediately after the target wallclock timestamp is reached.
7521 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
7527 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
7530 The duration of content to pass on as preroll expressed in seconds. Default is 0.
7533 The maximum duration of content to buffer before waiting for the cue expressed
7534 in seconds. Default is 0.
7541 Apply color adjustments using curves.
7543 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
7544 component (red, green and blue) has its values defined by @var{N} key points
7545 tied from each other using a smooth curve. The x-axis represents the pixel
7546 values from the input frame, and the y-axis the new pixel values to be set for
7549 By default, a component curve is defined by the two points @var{(0;0)} and
7550 @var{(1;1)}. This creates a straight line where each original pixel value is
7551 "adjusted" to its own value, which means no change to the image.
7553 The filter allows you to redefine these two points and add some more. A new
7554 curve (using a natural cubic spline interpolation) will be define to pass
7555 smoothly through all these new coordinates. The new defined points needs to be
7556 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
7557 be in the @var{[0;1]} interval. If the computed curves happened to go outside
7558 the vector spaces, the values will be clipped accordingly.
7560 The filter accepts the following options:
7564 Select one of the available color presets. This option can be used in addition
7565 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
7566 options takes priority on the preset values.
7567 Available presets are:
7570 @item color_negative
7573 @item increase_contrast
7575 @item linear_contrast
7576 @item medium_contrast
7578 @item strong_contrast
7581 Default is @code{none}.
7583 Set the master key points. These points will define a second pass mapping. It
7584 is sometimes called a "luminance" or "value" mapping. It can be used with
7585 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
7586 post-processing LUT.
7588 Set the key points for the red component.
7590 Set the key points for the green component.
7592 Set the key points for the blue component.
7594 Set the key points for all components (not including master).
7595 Can be used in addition to the other key points component
7596 options. In this case, the unset component(s) will fallback on this
7597 @option{all} setting.
7599 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
7601 Save Gnuplot script of the curves in specified file.
7604 To avoid some filtergraph syntax conflicts, each key points list need to be
7605 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
7607 @subsection Examples
7611 Increase slightly the middle level of blue:
7613 curves=blue='0/0 0.5/0.58 1/1'
7619 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'
7621 Here we obtain the following coordinates for each components:
7624 @code{(0;0.11) (0.42;0.51) (1;0.95)}
7626 @code{(0;0) (0.50;0.48) (1;1)}
7628 @code{(0;0.22) (0.49;0.44) (1;0.80)}
7632 The previous example can also be achieved with the associated built-in preset:
7634 curves=preset=vintage
7644 Use a Photoshop preset and redefine the points of the green component:
7646 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
7650 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
7651 and @command{gnuplot}:
7653 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
7654 gnuplot -p /tmp/curves.plt
7660 Video data analysis filter.
7662 This filter shows hexadecimal pixel values of part of video.
7664 The filter accepts the following options:
7668 Set output video size.
7671 Set x offset from where to pick pixels.
7674 Set y offset from where to pick pixels.
7677 Set scope mode, can be one of the following:
7680 Draw hexadecimal pixel values with white color on black background.
7683 Draw hexadecimal pixel values with input video pixel color on black
7687 Draw hexadecimal pixel values on color background picked from input video,
7688 the text color is picked in such way so its always visible.
7692 Draw rows and columns numbers on left and top of video.
7695 Set background opacity.
7700 Denoise frames using 2D DCT (frequency domain filtering).
7702 This filter is not designed for real time.
7704 The filter accepts the following options:
7708 Set the noise sigma constant.
7710 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
7711 coefficient (absolute value) below this threshold with be dropped.
7713 If you need a more advanced filtering, see @option{expr}.
7715 Default is @code{0}.
7718 Set number overlapping pixels for each block. Since the filter can be slow, you
7719 may want to reduce this value, at the cost of a less effective filter and the
7720 risk of various artefacts.
7722 If the overlapping value doesn't permit processing the whole input width or
7723 height, a warning will be displayed and according borders won't be denoised.
7725 Default value is @var{blocksize}-1, which is the best possible setting.
7728 Set the coefficient factor expression.
7730 For each coefficient of a DCT block, this expression will be evaluated as a
7731 multiplier value for the coefficient.
7733 If this is option is set, the @option{sigma} option will be ignored.
7735 The absolute value of the coefficient can be accessed through the @var{c}
7739 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
7740 @var{blocksize}, which is the width and height of the processed blocks.
7742 The default value is @var{3} (8x8) and can be raised to @var{4} for a
7743 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
7744 on the speed processing. Also, a larger block size does not necessarily means a
7748 @subsection Examples
7750 Apply a denoise with a @option{sigma} of @code{4.5}:
7755 The same operation can be achieved using the expression system:
7757 dctdnoiz=e='gte(c, 4.5*3)'
7760 Violent denoise using a block size of @code{16x16}:
7767 Remove banding artifacts from input video.
7768 It works by replacing banded pixels with average value of referenced pixels.
7770 The filter accepts the following options:
7777 Set banding detection threshold for each plane. Default is 0.02.
7778 Valid range is 0.00003 to 0.5.
7779 If difference between current pixel and reference pixel is less than threshold,
7780 it will be considered as banded.
7783 Banding detection range in pixels. Default is 16. If positive, random number
7784 in range 0 to set value will be used. If negative, exact absolute value
7786 The range defines square of four pixels around current pixel.
7789 Set direction in radians from which four pixel will be compared. If positive,
7790 random direction from 0 to set direction will be picked. If negative, exact of
7791 absolute value will be picked. For example direction 0, -PI or -2*PI radians
7792 will pick only pixels on same row and -PI/2 will pick only pixels on same
7796 If enabled, current pixel is compared with average value of all four
7797 surrounding pixels. The default is enabled. If disabled current pixel is
7798 compared with all four surrounding pixels. The pixel is considered banded
7799 if only all four differences with surrounding pixels are less than threshold.
7802 If enabled, current pixel is changed if and only if all pixel components are banded,
7803 e.g. banding detection threshold is triggered for all color components.
7804 The default is disabled.
7809 Remove blocking artifacts from input video.
7811 The filter accepts the following options:
7815 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
7816 This controls what kind of deblocking is applied.
7819 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
7825 Set blocking detection thresholds. Allowed range is 0 to 1.
7826 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
7827 Using higher threshold gives more deblocking strength.
7828 Setting @var{alpha} controls threshold detection at exact edge of block.
7829 Remaining options controls threshold detection near the edge. Each one for
7830 below/above or left/right. Setting any of those to @var{0} disables
7834 Set planes to filter. Default is to filter all available planes.
7837 @subsection Examples
7841 Deblock using weak filter and block size of 4 pixels.
7843 deblock=filter=weak:block=4
7847 Deblock using strong filter, block size of 4 pixels and custom thresholds for
7848 deblocking more edges.
7850 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
7854 Similar as above, but filter only first plane.
7856 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
7860 Similar as above, but filter only second and third plane.
7862 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
7869 Drop duplicated frames at regular intervals.
7871 The filter accepts the following options:
7875 Set the number of frames from which one will be dropped. Setting this to
7876 @var{N} means one frame in every batch of @var{N} frames will be dropped.
7877 Default is @code{5}.
7880 Set the threshold for duplicate detection. If the difference metric for a frame
7881 is less than or equal to this value, then it is declared as duplicate. Default
7885 Set scene change threshold. Default is @code{15}.
7889 Set the size of the x and y-axis blocks used during metric calculations.
7890 Larger blocks give better noise suppression, but also give worse detection of
7891 small movements. Must be a power of two. Default is @code{32}.
7894 Mark main input as a pre-processed input and activate clean source input
7895 stream. This allows the input to be pre-processed with various filters to help
7896 the metrics calculation while keeping the frame selection lossless. When set to
7897 @code{1}, the first stream is for the pre-processed input, and the second
7898 stream is the clean source from where the kept frames are chosen. Default is
7902 Set whether or not chroma is considered in the metric calculations. Default is
7908 Apply 2D deconvolution of video stream in frequency domain using second stream
7911 The filter accepts the following options:
7915 Set which planes to process.
7918 Set which impulse video frames will be processed, can be @var{first}
7919 or @var{all}. Default is @var{all}.
7922 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
7923 and height are not same and not power of 2 or if stream prior to convolving
7927 The @code{deconvolve} filter also supports the @ref{framesync} options.
7931 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
7933 It accepts the following options:
7937 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
7938 @var{rainbows} for cross-color reduction.
7941 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
7944 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
7947 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
7950 Set temporal chroma threshold. Lower values increases reduction of cross-color.
7955 Apply deflate effect to the video.
7957 This filter replaces the pixel by the local(3x3) average by taking into account
7958 only values lower than the pixel.
7960 It accepts the following options:
7967 Limit the maximum change for each plane, default is 65535.
7968 If 0, plane will remain unchanged.
7973 Remove temporal frame luminance variations.
7975 It accepts the following options:
7979 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
7982 Set averaging mode to smooth temporal luminance variations.
7984 Available values are:
8009 Do not actually modify frame. Useful when one only wants metadata.
8014 Remove judder produced by partially interlaced telecined content.
8016 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
8017 source was partially telecined content then the output of @code{pullup,dejudder}
8018 will have a variable frame rate. May change the recorded frame rate of the
8019 container. Aside from that change, this filter will not affect constant frame
8022 The option available in this filter is:
8026 Specify the length of the window over which the judder repeats.
8028 Accepts any integer greater than 1. Useful values are:
8032 If the original was telecined from 24 to 30 fps (Film to NTSC).
8035 If the original was telecined from 25 to 30 fps (PAL to NTSC).
8038 If a mixture of the two.
8041 The default is @samp{4}.
8046 Suppress a TV station logo by a simple interpolation of the surrounding
8047 pixels. Just set a rectangle covering the logo and watch it disappear
8048 (and sometimes something even uglier appear - your mileage may vary).
8050 It accepts the following parameters:
8055 Specify the top left corner coordinates of the logo. They must be
8060 Specify the width and height of the logo to clear. They must be
8064 Specify the thickness of the fuzzy edge of the rectangle (added to
8065 @var{w} and @var{h}). The default value is 1. This option is
8066 deprecated, setting higher values should no longer be necessary and
8070 When set to 1, a green rectangle is drawn on the screen to simplify
8071 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
8072 The default value is 0.
8074 The rectangle is drawn on the outermost pixels which will be (partly)
8075 replaced with interpolated values. The values of the next pixels
8076 immediately outside this rectangle in each direction will be used to
8077 compute the interpolated pixel values inside the rectangle.
8081 @subsection Examples
8085 Set a rectangle covering the area with top left corner coordinates 0,0
8086 and size 100x77, and a band of size 10:
8088 delogo=x=0:y=0:w=100:h=77:band=10
8095 Attempt to fix small changes in horizontal and/or vertical shift. This
8096 filter helps remove camera shake from hand-holding a camera, bumping a
8097 tripod, moving on a vehicle, etc.
8099 The filter accepts the following options:
8107 Specify a rectangular area where to limit the search for motion
8109 If desired the search for motion vectors can be limited to a
8110 rectangular area of the frame defined by its top left corner, width
8111 and height. These parameters have the same meaning as the drawbox
8112 filter which can be used to visualise the position of the bounding
8115 This is useful when simultaneous movement of subjects within the frame
8116 might be confused for camera motion by the motion vector search.
8118 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
8119 then the full frame is used. This allows later options to be set
8120 without specifying the bounding box for the motion vector search.
8122 Default - search the whole frame.
8126 Specify the maximum extent of movement in x and y directions in the
8127 range 0-64 pixels. Default 16.
8130 Specify how to generate pixels to fill blanks at the edge of the
8131 frame. Available values are:
8134 Fill zeroes at blank locations
8136 Original image at blank locations
8138 Extruded edge value at blank locations
8140 Mirrored edge at blank locations
8142 Default value is @samp{mirror}.
8145 Specify the blocksize to use for motion search. Range 4-128 pixels,
8149 Specify the contrast threshold for blocks. Only blocks with more than
8150 the specified contrast (difference between darkest and lightest
8151 pixels) will be considered. Range 1-255, default 125.
8154 Specify the search strategy. Available values are:
8157 Set exhaustive search
8159 Set less exhaustive search.
8161 Default value is @samp{exhaustive}.
8164 If set then a detailed log of the motion search is written to the
8171 Remove unwanted contamination of foreground colors, caused by reflected color of
8172 greenscreen or bluescreen.
8174 This filter accepts the following options:
8178 Set what type of despill to use.
8181 Set how spillmap will be generated.
8184 Set how much to get rid of still remaining spill.
8187 Controls amount of red in spill area.
8190 Controls amount of green in spill area.
8191 Should be -1 for greenscreen.
8194 Controls amount of blue in spill area.
8195 Should be -1 for bluescreen.
8198 Controls brightness of spill area, preserving colors.
8201 Modify alpha from generated spillmap.
8206 Apply an exact inverse of the telecine operation. It requires a predefined
8207 pattern specified using the pattern option which must be the same as that passed
8208 to the telecine filter.
8210 This filter accepts the following options:
8219 The default value is @code{top}.
8223 A string of numbers representing the pulldown pattern you wish to apply.
8224 The default value is @code{23}.
8227 A number representing position of the first frame with respect to the telecine
8228 pattern. This is to be used if the stream is cut. The default value is @code{0}.
8233 Apply dilation effect to the video.
8235 This filter replaces the pixel by the local(3x3) maximum.
8237 It accepts the following options:
8244 Limit the maximum change for each plane, default is 65535.
8245 If 0, plane will remain unchanged.
8248 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
8251 Flags to local 3x3 coordinates maps like this:
8260 Displace pixels as indicated by second and third input stream.
8262 It takes three input streams and outputs one stream, the first input is the
8263 source, and second and third input are displacement maps.
8265 The second input specifies how much to displace pixels along the
8266 x-axis, while the third input specifies how much to displace pixels
8268 If one of displacement map streams terminates, last frame from that
8269 displacement map will be used.
8271 Note that once generated, displacements maps can be reused over and over again.
8273 A description of the accepted options follows.
8277 Set displace behavior for pixels that are out of range.
8279 Available values are:
8282 Missing pixels are replaced by black pixels.
8285 Adjacent pixels will spread out to replace missing pixels.
8288 Out of range pixels are wrapped so they point to pixels of other side.
8291 Out of range pixels will be replaced with mirrored pixels.
8293 Default is @samp{smear}.
8297 @subsection Examples
8301 Add ripple effect to rgb input of video size hd720:
8303 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
8307 Add wave effect to rgb input of video size hd720:
8309 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
8315 Draw a colored box on the input image.
8317 It accepts the following parameters:
8322 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
8326 The expressions which specify the width and height of the box; if 0 they are interpreted as
8327 the input width and height. It defaults to 0.
8330 Specify the color of the box to write. For the general syntax of this option,
8331 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8332 value @code{invert} is used, the box edge color is the same as the
8333 video with inverted luma.
8336 The expression which sets the thickness of the box edge.
8337 A value of @code{fill} will create a filled box. Default value is @code{3}.
8339 See below for the list of accepted constants.
8342 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
8343 will overwrite the video's color and alpha pixels.
8344 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
8347 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8348 following constants:
8352 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8356 horizontal and vertical chroma subsample values. For example for the
8357 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8361 The input width and height.
8364 The input sample aspect ratio.
8368 The x and y offset coordinates where the box is drawn.
8372 The width and height of the drawn box.
8375 The thickness of the drawn box.
8377 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8378 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8382 @subsection Examples
8386 Draw a black box around the edge of the input image:
8392 Draw a box with color red and an opacity of 50%:
8394 drawbox=10:20:200:60:red@@0.5
8397 The previous example can be specified as:
8399 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
8403 Fill the box with pink color:
8405 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
8409 Draw a 2-pixel red 2.40:1 mask:
8411 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
8417 Draw a grid on the input image.
8419 It accepts the following parameters:
8424 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
8428 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
8429 input width and height, respectively, minus @code{thickness}, so image gets
8430 framed. Default to 0.
8433 Specify the color of the grid. For the general syntax of this option,
8434 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8435 value @code{invert} is used, the grid color is the same as the
8436 video with inverted luma.
8439 The expression which sets the thickness of the grid line. Default value is @code{1}.
8441 See below for the list of accepted constants.
8444 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
8445 will overwrite the video's color and alpha pixels.
8446 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
8449 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8450 following constants:
8454 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8458 horizontal and vertical chroma subsample values. For example for the
8459 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8463 The input grid cell width and height.
8466 The input sample aspect ratio.
8470 The x and y coordinates of some point of grid intersection (meant to configure offset).
8474 The width and height of the drawn cell.
8477 The thickness of the drawn cell.
8479 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8480 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8484 @subsection Examples
8488 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
8490 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
8494 Draw a white 3x3 grid with an opacity of 50%:
8496 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
8503 Draw a text string or text from a specified file on top of a video, using the
8504 libfreetype library.
8506 To enable compilation of this filter, you need to configure FFmpeg with
8507 @code{--enable-libfreetype}.
8508 To enable default font fallback and the @var{font} option you need to
8509 configure FFmpeg with @code{--enable-libfontconfig}.
8510 To enable the @var{text_shaping} option, you need to configure FFmpeg with
8511 @code{--enable-libfribidi}.
8515 It accepts the following parameters:
8520 Used to draw a box around text using the background color.
8521 The value must be either 1 (enable) or 0 (disable).
8522 The default value of @var{box} is 0.
8525 Set the width of the border to be drawn around the box using @var{boxcolor}.
8526 The default value of @var{boxborderw} is 0.
8529 The color to be used for drawing box around text. For the syntax of this
8530 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8532 The default value of @var{boxcolor} is "white".
8535 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
8536 The default value of @var{line_spacing} is 0.
8539 Set the width of the border to be drawn around the text using @var{bordercolor}.
8540 The default value of @var{borderw} is 0.
8543 Set the color to be used for drawing border around text. For the syntax of this
8544 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8546 The default value of @var{bordercolor} is "black".
8549 Select how the @var{text} is expanded. Can be either @code{none},
8550 @code{strftime} (deprecated) or
8551 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
8555 Set a start time for the count. Value is in microseconds. Only applied
8556 in the deprecated strftime expansion mode. To emulate in normal expansion
8557 mode use the @code{pts} function, supplying the start time (in seconds)
8558 as the second argument.
8561 If true, check and fix text coords to avoid clipping.
8564 The color to be used for drawing fonts. For the syntax of this option, check
8565 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8567 The default value of @var{fontcolor} is "black".
8569 @item fontcolor_expr
8570 String which is expanded the same way as @var{text} to obtain dynamic
8571 @var{fontcolor} value. By default this option has empty value and is not
8572 processed. When this option is set, it overrides @var{fontcolor} option.
8575 The font family to be used for drawing text. By default Sans.
8578 The font file to be used for drawing text. The path must be included.
8579 This parameter is mandatory if the fontconfig support is disabled.
8582 Draw the text applying alpha blending. The value can
8583 be a number between 0.0 and 1.0.
8584 The expression accepts the same variables @var{x, y} as well.
8585 The default value is 1.
8586 Please see @var{fontcolor_expr}.
8589 The font size to be used for drawing text.
8590 The default value of @var{fontsize} is 16.
8593 If set to 1, attempt to shape the text (for example, reverse the order of
8594 right-to-left text and join Arabic characters) before drawing it.
8595 Otherwise, just draw the text exactly as given.
8596 By default 1 (if supported).
8599 The flags to be used for loading the fonts.
8601 The flags map the corresponding flags supported by libfreetype, and are
8602 a combination of the following values:
8609 @item vertical_layout
8610 @item force_autohint
8613 @item ignore_global_advance_width
8615 @item ignore_transform
8621 Default value is "default".
8623 For more information consult the documentation for the FT_LOAD_*
8627 The color to be used for drawing a shadow behind the drawn text. For the
8628 syntax of this option, check the @ref{color syntax,,"Color" section in the
8629 ffmpeg-utils manual,ffmpeg-utils}.
8631 The default value of @var{shadowcolor} is "black".
8635 The x and y offsets for the text shadow position with respect to the
8636 position of the text. They can be either positive or negative
8637 values. The default value for both is "0".
8640 The starting frame number for the n/frame_num variable. The default value
8644 The size in number of spaces to use for rendering the tab.
8648 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
8649 format. It can be used with or without text parameter. @var{timecode_rate}
8650 option must be specified.
8652 @item timecode_rate, rate, r
8653 Set the timecode frame rate (timecode only). Value will be rounded to nearest
8654 integer. Minimum value is "1".
8655 Drop-frame timecode is supported for frame rates 30 & 60.
8658 If set to 1, the output of the timecode option will wrap around at 24 hours.
8659 Default is 0 (disabled).
8662 The text string to be drawn. The text must be a sequence of UTF-8
8664 This parameter is mandatory if no file is specified with the parameter
8668 A text file containing text to be drawn. The text must be a sequence
8669 of UTF-8 encoded characters.
8671 This parameter is mandatory if no text string is specified with the
8672 parameter @var{text}.
8674 If both @var{text} and @var{textfile} are specified, an error is thrown.
8677 If set to 1, the @var{textfile} will be reloaded before each frame.
8678 Be sure to update it atomically, or it may be read partially, or even fail.
8682 The expressions which specify the offsets where text will be drawn
8683 within the video frame. They are relative to the top/left border of the
8686 The default value of @var{x} and @var{y} is "0".
8688 See below for the list of accepted constants and functions.
8691 The parameters for @var{x} and @var{y} are expressions containing the
8692 following constants and functions:
8696 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
8700 horizontal and vertical chroma subsample values. For example for the
8701 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8704 the height of each text line
8712 @item max_glyph_a, ascent
8713 the maximum distance from the baseline to the highest/upper grid
8714 coordinate used to place a glyph outline point, for all the rendered
8716 It is a positive value, due to the grid's orientation with the Y axis
8719 @item max_glyph_d, descent
8720 the maximum distance from the baseline to the lowest grid coordinate
8721 used to place a glyph outline point, for all the rendered glyphs.
8722 This is a negative value, due to the grid's orientation, with the Y axis
8726 maximum glyph height, that is the maximum height for all the glyphs
8727 contained in the rendered text, it is equivalent to @var{ascent} -
8731 maximum glyph width, that is the maximum width for all the glyphs
8732 contained in the rendered text
8735 the number of input frame, starting from 0
8737 @item rand(min, max)
8738 return a random number included between @var{min} and @var{max}
8741 The input sample aspect ratio.
8744 timestamp expressed in seconds, NAN if the input timestamp is unknown
8747 the height of the rendered text
8750 the width of the rendered text
8754 the x and y offset coordinates where the text is drawn.
8756 These parameters allow the @var{x} and @var{y} expressions to refer
8757 each other, so you can for example specify @code{y=x/dar}.
8760 @anchor{drawtext_expansion}
8761 @subsection Text expansion
8763 If @option{expansion} is set to @code{strftime},
8764 the filter recognizes strftime() sequences in the provided text and
8765 expands them accordingly. Check the documentation of strftime(). This
8766 feature is deprecated.
8768 If @option{expansion} is set to @code{none}, the text is printed verbatim.
8770 If @option{expansion} is set to @code{normal} (which is the default),
8771 the following expansion mechanism is used.
8773 The backslash character @samp{\}, followed by any character, always expands to
8774 the second character.
8776 Sequences of the form @code{%@{...@}} are expanded. The text between the
8777 braces is a function name, possibly followed by arguments separated by ':'.
8778 If the arguments contain special characters or delimiters (':' or '@}'),
8779 they should be escaped.
8781 Note that they probably must also be escaped as the value for the
8782 @option{text} option in the filter argument string and as the filter
8783 argument in the filtergraph description, and possibly also for the shell,
8784 that makes up to four levels of escaping; using a text file avoids these
8787 The following functions are available:
8792 The expression evaluation result.
8794 It must take one argument specifying the expression to be evaluated,
8795 which accepts the same constants and functions as the @var{x} and
8796 @var{y} values. Note that not all constants should be used, for
8797 example the text size is not known when evaluating the expression, so
8798 the constants @var{text_w} and @var{text_h} will have an undefined
8801 @item expr_int_format, eif
8802 Evaluate the expression's value and output as formatted integer.
8804 The first argument is the expression to be evaluated, just as for the @var{expr} function.
8805 The second argument specifies the output format. Allowed values are @samp{x},
8806 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
8807 @code{printf} function.
8808 The third parameter is optional and sets the number of positions taken by the output.
8809 It can be used to add padding with zeros from the left.
8812 The time at which the filter is running, expressed in UTC.
8813 It can accept an argument: a strftime() format string.
8816 The time at which the filter is running, expressed in the local time zone.
8817 It can accept an argument: a strftime() format string.
8820 Frame metadata. Takes one or two arguments.
8822 The first argument is mandatory and specifies the metadata key.
8824 The second argument is optional and specifies a default value, used when the
8825 metadata key is not found or empty.
8828 The frame number, starting from 0.
8831 A 1 character description of the current picture type.
8834 The timestamp of the current frame.
8835 It can take up to three arguments.
8837 The first argument is the format of the timestamp; it defaults to @code{flt}
8838 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
8839 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
8840 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
8841 @code{localtime} stands for the timestamp of the frame formatted as
8842 local time zone time.
8844 The second argument is an offset added to the timestamp.
8846 If the format is set to @code{hms}, a third argument @code{24HH} may be
8847 supplied to present the hour part of the formatted timestamp in 24h format
8850 If the format is set to @code{localtime} or @code{gmtime},
8851 a third argument may be supplied: a strftime() format string.
8852 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
8855 @subsection Examples
8859 Draw "Test Text" with font FreeSerif, using the default values for the
8860 optional parameters.
8863 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
8867 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
8868 and y=50 (counting from the top-left corner of the screen), text is
8869 yellow with a red box around it. Both the text and the box have an
8873 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
8874 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
8877 Note that the double quotes are not necessary if spaces are not used
8878 within the parameter list.
8881 Show the text at the center of the video frame:
8883 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
8887 Show the text at a random position, switching to a new position every 30 seconds:
8889 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)"
8893 Show a text line sliding from right to left in the last row of the video
8894 frame. The file @file{LONG_LINE} is assumed to contain a single line
8897 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
8901 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
8903 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
8907 Draw a single green letter "g", at the center of the input video.
8908 The glyph baseline is placed at half screen height.
8910 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
8914 Show text for 1 second every 3 seconds:
8916 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
8920 Use fontconfig to set the font. Note that the colons need to be escaped.
8922 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
8926 Print the date of a real-time encoding (see strftime(3)):
8928 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
8932 Show text fading in and out (appearing/disappearing):
8935 DS=1.0 # display start
8936 DE=10.0 # display end
8937 FID=1.5 # fade in duration
8938 FOD=5 # fade out duration
8939 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 @}"
8943 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
8944 and the @option{fontsize} value are included in the @option{y} offset.
8946 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
8947 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
8952 For more information about libfreetype, check:
8953 @url{http://www.freetype.org/}.
8955 For more information about fontconfig, check:
8956 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
8958 For more information about libfribidi, check:
8959 @url{http://fribidi.org/}.
8963 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
8965 The filter accepts the following options:
8970 Set low and high threshold values used by the Canny thresholding
8973 The high threshold selects the "strong" edge pixels, which are then
8974 connected through 8-connectivity with the "weak" edge pixels selected
8975 by the low threshold.
8977 @var{low} and @var{high} threshold values must be chosen in the range
8978 [0,1], and @var{low} should be lesser or equal to @var{high}.
8980 Default value for @var{low} is @code{20/255}, and default value for @var{high}
8984 Define the drawing mode.
8988 Draw white/gray wires on black background.
8991 Mix the colors to create a paint/cartoon effect.
8994 Apply Canny edge detector on all selected planes.
8996 Default value is @var{wires}.
8999 Select planes for filtering. By default all available planes are filtered.
9002 @subsection Examples
9006 Standard edge detection with custom values for the hysteresis thresholding:
9008 edgedetect=low=0.1:high=0.4
9012 Painting effect without thresholding:
9014 edgedetect=mode=colormix:high=0
9019 Set brightness, contrast, saturation and approximate gamma adjustment.
9021 The filter accepts the following options:
9025 Set the contrast expression. The value must be a float value in range
9026 @code{-2.0} to @code{2.0}. The default value is "1".
9029 Set the brightness expression. The value must be a float value in
9030 range @code{-1.0} to @code{1.0}. The default value is "0".
9033 Set the saturation expression. The value must be a float in
9034 range @code{0.0} to @code{3.0}. The default value is "1".
9037 Set the gamma expression. The value must be a float in range
9038 @code{0.1} to @code{10.0}. The default value is "1".
9041 Set the gamma expression for red. The value must be a float in
9042 range @code{0.1} to @code{10.0}. The default value is "1".
9045 Set the gamma expression for green. The value must be a float in range
9046 @code{0.1} to @code{10.0}. The default value is "1".
9049 Set the gamma expression for blue. The value must be a float in range
9050 @code{0.1} to @code{10.0}. The default value is "1".
9053 Set the gamma weight expression. It can be used to reduce the effect
9054 of a high gamma value on bright image areas, e.g. keep them from
9055 getting overamplified and just plain white. The value must be a float
9056 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
9057 gamma correction all the way down while @code{1.0} leaves it at its
9058 full strength. Default is "1".
9061 Set when the expressions for brightness, contrast, saturation and
9062 gamma expressions are evaluated.
9064 It accepts the following values:
9067 only evaluate expressions once during the filter initialization or
9068 when a command is processed
9071 evaluate expressions for each incoming frame
9074 Default value is @samp{init}.
9077 The expressions accept the following parameters:
9080 frame count of the input frame starting from 0
9083 byte position of the corresponding packet in the input file, NAN if
9087 frame rate of the input video, NAN if the input frame rate is unknown
9090 timestamp expressed in seconds, NAN if the input timestamp is unknown
9093 @subsection Commands
9094 The filter supports the following commands:
9098 Set the contrast expression.
9101 Set the brightness expression.
9104 Set the saturation expression.
9107 Set the gamma expression.
9110 Set the gamma_r expression.
9113 Set gamma_g expression.
9116 Set gamma_b expression.
9119 Set gamma_weight expression.
9121 The command accepts the same syntax of the corresponding option.
9123 If the specified expression is not valid, it is kept at its current
9130 Apply erosion effect to the video.
9132 This filter replaces the pixel by the local(3x3) minimum.
9134 It accepts the following options:
9141 Limit the maximum change for each plane, default is 65535.
9142 If 0, plane will remain unchanged.
9145 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9148 Flags to local 3x3 coordinates maps like this:
9155 @section extractplanes
9157 Extract color channel components from input video stream into
9158 separate grayscale video streams.
9160 The filter accepts the following option:
9164 Set plane(s) to extract.
9166 Available values for planes are:
9177 Choosing planes not available in the input will result in an error.
9178 That means you cannot select @code{r}, @code{g}, @code{b} planes
9179 with @code{y}, @code{u}, @code{v} planes at same time.
9182 @subsection Examples
9186 Extract luma, u and v color channel component from input video frame
9187 into 3 grayscale outputs:
9189 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
9195 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
9197 For each input image, the filter will compute the optimal mapping from
9198 the input to the output given the codebook length, that is the number
9199 of distinct output colors.
9201 This filter accepts the following options.
9204 @item codebook_length, l
9205 Set codebook length. The value must be a positive integer, and
9206 represents the number of distinct output colors. Default value is 256.
9209 Set the maximum number of iterations to apply for computing the optimal
9210 mapping. The higher the value the better the result and the higher the
9211 computation time. Default value is 1.
9214 Set a random seed, must be an integer included between 0 and
9215 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
9216 will try to use a good random seed on a best effort basis.
9219 Set pal8 output pixel format. This option does not work with codebook
9220 length greater than 256.
9225 Measure graylevel entropy in histogram of color channels of video frames.
9227 It accepts the following parameters:
9231 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
9233 @var{diff} mode measures entropy of histogram delta values, absolute differences
9234 between neighbour histogram values.
9239 Apply a fade-in/out effect to the input video.
9241 It accepts the following parameters:
9245 The effect type can be either "in" for a fade-in, or "out" for a fade-out
9247 Default is @code{in}.
9249 @item start_frame, s
9250 Specify the number of the frame to start applying the fade
9251 effect at. Default is 0.
9254 The number of frames that the fade effect lasts. At the end of the
9255 fade-in effect, the output video will have the same intensity as the input video.
9256 At the end of the fade-out transition, the output video will be filled with the
9257 selected @option{color}.
9261 If set to 1, fade only alpha channel, if one exists on the input.
9264 @item start_time, st
9265 Specify the timestamp (in seconds) of the frame to start to apply the fade
9266 effect. If both start_frame and start_time are specified, the fade will start at
9267 whichever comes last. Default is 0.
9270 The number of seconds for which the fade effect has to last. At the end of the
9271 fade-in effect the output video will have the same intensity as the input video,
9272 at the end of the fade-out transition the output video will be filled with the
9273 selected @option{color}.
9274 If both duration and nb_frames are specified, duration is used. Default is 0
9275 (nb_frames is used by default).
9278 Specify the color of the fade. Default is "black".
9281 @subsection Examples
9285 Fade in the first 30 frames of video:
9290 The command above is equivalent to:
9296 Fade out the last 45 frames of a 200-frame video:
9299 fade=type=out:start_frame=155:nb_frames=45
9303 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
9305 fade=in:0:25, fade=out:975:25
9309 Make the first 5 frames yellow, then fade in from frame 5-24:
9311 fade=in:5:20:color=yellow
9315 Fade in alpha over first 25 frames of video:
9317 fade=in:0:25:alpha=1
9321 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
9323 fade=t=in:st=5.5:d=0.5
9329 Apply arbitrary expressions to samples in frequency domain
9333 Adjust the dc value (gain) of the luma plane of the image. The filter
9334 accepts an integer value in range @code{0} to @code{1000}. The default
9335 value is set to @code{0}.
9338 Adjust the dc value (gain) of the 1st chroma plane of the image. The
9339 filter accepts an integer value in range @code{0} to @code{1000}. The
9340 default value is set to @code{0}.
9343 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
9344 filter accepts an integer value in range @code{0} to @code{1000}. The
9345 default value is set to @code{0}.
9348 Set the frequency domain weight expression for the luma plane.
9351 Set the frequency domain weight expression for the 1st chroma plane.
9354 Set the frequency domain weight expression for the 2nd chroma plane.
9357 Set when the expressions are evaluated.
9359 It accepts the following values:
9362 Only evaluate expressions once during the filter initialization.
9365 Evaluate expressions for each incoming frame.
9368 Default value is @samp{init}.
9370 The filter accepts the following variables:
9373 The coordinates of the current sample.
9377 The width and height of the image.
9380 The number of input frame, starting from 0.
9383 @subsection Examples
9389 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
9395 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
9401 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
9407 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
9413 Denoise frames using 3D FFT (frequency domain filtering).
9415 The filter accepts the following options:
9419 Set the noise sigma constant. This sets denoising strength.
9420 Default value is 1. Allowed range is from 0 to 30.
9421 Using very high sigma with low overlap may give blocking artifacts.
9424 Set amount of denoising. By default all detected noise is reduced.
9425 Default value is 1. Allowed range is from 0 to 1.
9428 Set size of block, Default is 4, can be 3, 4, 5 or 6.
9429 Actual size of block in pixels is 2 to power of @var{block}, so by default
9430 block size in pixels is 2^4 which is 16.
9433 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
9436 Set number of previous frames to use for denoising. By default is set to 0.
9439 Set number of next frames to to use for denoising. By default is set to 0.
9442 Set planes which will be filtered, by default are all available filtered
9448 Extract a single field from an interlaced image using stride
9449 arithmetic to avoid wasting CPU time. The output frames are marked as
9452 The filter accepts the following options:
9456 Specify whether to extract the top (if the value is @code{0} or
9457 @code{top}) or the bottom field (if the value is @code{1} or
9463 Create new frames by copying the top and bottom fields from surrounding frames
9464 supplied as numbers by the hint file.
9468 Set file containing hints: absolute/relative frame numbers.
9470 There must be one line for each frame in a clip. Each line must contain two
9471 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
9472 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
9473 is current frame number for @code{absolute} mode or out of [-1, 1] range
9474 for @code{relative} mode. First number tells from which frame to pick up top
9475 field and second number tells from which frame to pick up bottom field.
9477 If optionally followed by @code{+} output frame will be marked as interlaced,
9478 else if followed by @code{-} output frame will be marked as progressive, else
9479 it will be marked same as input frame.
9480 If line starts with @code{#} or @code{;} that line is skipped.
9483 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
9486 Example of first several lines of @code{hint} file for @code{relative} mode:
9489 1,0 - # second frame, use third's frame top field and second's frame bottom field
9490 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
9507 Field matching filter for inverse telecine. It is meant to reconstruct the
9508 progressive frames from a telecined stream. The filter does not drop duplicated
9509 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
9510 followed by a decimation filter such as @ref{decimate} in the filtergraph.
9512 The separation of the field matching and the decimation is notably motivated by
9513 the possibility of inserting a de-interlacing filter fallback between the two.
9514 If the source has mixed telecined and real interlaced content,
9515 @code{fieldmatch} will not be able to match fields for the interlaced parts.
9516 But these remaining combed frames will be marked as interlaced, and thus can be
9517 de-interlaced by a later filter such as @ref{yadif} before decimation.
9519 In addition to the various configuration options, @code{fieldmatch} can take an
9520 optional second stream, activated through the @option{ppsrc} option. If
9521 enabled, the frames reconstruction will be based on the fields and frames from
9522 this second stream. This allows the first input to be pre-processed in order to
9523 help the various algorithms of the filter, while keeping the output lossless
9524 (assuming the fields are matched properly). Typically, a field-aware denoiser,
9525 or brightness/contrast adjustments can help.
9527 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
9528 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
9529 which @code{fieldmatch} is based on. While the semantic and usage are very
9530 close, some behaviour and options names can differ.
9532 The @ref{decimate} filter currently only works for constant frame rate input.
9533 If your input has mixed telecined (30fps) and progressive content with a lower
9534 framerate like 24fps use the following filterchain to produce the necessary cfr
9535 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
9537 The filter accepts the following options:
9541 Specify the assumed field order of the input stream. Available values are:
9545 Auto detect parity (use FFmpeg's internal parity value).
9547 Assume bottom field first.
9549 Assume top field first.
9552 Note that it is sometimes recommended not to trust the parity announced by the
9555 Default value is @var{auto}.
9558 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
9559 sense that it won't risk creating jerkiness due to duplicate frames when
9560 possible, but if there are bad edits or blended fields it will end up
9561 outputting combed frames when a good match might actually exist. On the other
9562 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
9563 but will almost always find a good frame if there is one. The other values are
9564 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
9565 jerkiness and creating duplicate frames versus finding good matches in sections
9566 with bad edits, orphaned fields, blended fields, etc.
9568 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
9570 Available values are:
9574 2-way matching (p/c)
9576 2-way matching, and trying 3rd match if still combed (p/c + n)
9578 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
9580 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
9581 still combed (p/c + n + u/b)
9583 3-way matching (p/c/n)
9585 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
9586 detected as combed (p/c/n + u/b)
9589 The parenthesis at the end indicate the matches that would be used for that
9590 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
9593 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
9596 Default value is @var{pc_n}.
9599 Mark the main input stream as a pre-processed input, and enable the secondary
9600 input stream as the clean source to pick the fields from. See the filter
9601 introduction for more details. It is similar to the @option{clip2} feature from
9604 Default value is @code{0} (disabled).
9607 Set the field to match from. It is recommended to set this to the same value as
9608 @option{order} unless you experience matching failures with that setting. In
9609 certain circumstances changing the field that is used to match from can have a
9610 large impact on matching performance. Available values are:
9614 Automatic (same value as @option{order}).
9616 Match from the bottom field.
9618 Match from the top field.
9621 Default value is @var{auto}.
9624 Set whether or not chroma is included during the match comparisons. In most
9625 cases it is recommended to leave this enabled. You should set this to @code{0}
9626 only if your clip has bad chroma problems such as heavy rainbowing or other
9627 artifacts. Setting this to @code{0} could also be used to speed things up at
9628 the cost of some accuracy.
9630 Default value is @code{1}.
9634 These define an exclusion band which excludes the lines between @option{y0} and
9635 @option{y1} from being included in the field matching decision. An exclusion
9636 band can be used to ignore subtitles, a logo, or other things that may
9637 interfere with the matching. @option{y0} sets the starting scan line and
9638 @option{y1} sets the ending line; all lines in between @option{y0} and
9639 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
9640 @option{y0} and @option{y1} to the same value will disable the feature.
9641 @option{y0} and @option{y1} defaults to @code{0}.
9644 Set the scene change detection threshold as a percentage of maximum change on
9645 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
9646 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
9647 @option{scthresh} is @code{[0.0, 100.0]}.
9649 Default value is @code{12.0}.
9652 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
9653 account the combed scores of matches when deciding what match to use as the
9654 final match. Available values are:
9658 No final matching based on combed scores.
9660 Combed scores are only used when a scene change is detected.
9662 Use combed scores all the time.
9665 Default is @var{sc}.
9668 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
9669 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
9670 Available values are:
9674 No forced calculation.
9676 Force p/c/n calculations.
9678 Force p/c/n/u/b calculations.
9681 Default value is @var{none}.
9684 This is the area combing threshold used for combed frame detection. This
9685 essentially controls how "strong" or "visible" combing must be to be detected.
9686 Larger values mean combing must be more visible and smaller values mean combing
9687 can be less visible or strong and still be detected. Valid settings are from
9688 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
9689 be detected as combed). This is basically a pixel difference value. A good
9690 range is @code{[8, 12]}.
9692 Default value is @code{9}.
9695 Sets whether or not chroma is considered in the combed frame decision. Only
9696 disable this if your source has chroma problems (rainbowing, etc.) that are
9697 causing problems for the combed frame detection with chroma enabled. Actually,
9698 using @option{chroma}=@var{0} is usually more reliable, except for the case
9699 where there is chroma only combing in the source.
9701 Default value is @code{0}.
9705 Respectively set the x-axis and y-axis size of the window used during combed
9706 frame detection. This has to do with the size of the area in which
9707 @option{combpel} pixels are required to be detected as combed for a frame to be
9708 declared combed. See the @option{combpel} parameter description for more info.
9709 Possible values are any number that is a power of 2 starting at 4 and going up
9712 Default value is @code{16}.
9715 The number of combed pixels inside any of the @option{blocky} by
9716 @option{blockx} size blocks on the frame for the frame to be detected as
9717 combed. While @option{cthresh} controls how "visible" the combing must be, this
9718 setting controls "how much" combing there must be in any localized area (a
9719 window defined by the @option{blockx} and @option{blocky} settings) on the
9720 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
9721 which point no frames will ever be detected as combed). This setting is known
9722 as @option{MI} in TFM/VFM vocabulary.
9724 Default value is @code{80}.
9727 @anchor{p/c/n/u/b meaning}
9728 @subsection p/c/n/u/b meaning
9730 @subsubsection p/c/n
9732 We assume the following telecined stream:
9735 Top fields: 1 2 2 3 4
9736 Bottom fields: 1 2 3 4 4
9739 The numbers correspond to the progressive frame the fields relate to. Here, the
9740 first two frames are progressive, the 3rd and 4th are combed, and so on.
9742 When @code{fieldmatch} is configured to run a matching from bottom
9743 (@option{field}=@var{bottom}) this is how this input stream get transformed:
9748 B 1 2 3 4 4 <-- matching reference
9757 As a result of the field matching, we can see that some frames get duplicated.
9758 To perform a complete inverse telecine, you need to rely on a decimation filter
9759 after this operation. See for instance the @ref{decimate} filter.
9761 The same operation now matching from top fields (@option{field}=@var{top})
9766 T 1 2 2 3 4 <-- matching reference
9776 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
9777 basically, they refer to the frame and field of the opposite parity:
9780 @item @var{p} matches the field of the opposite parity in the previous frame
9781 @item @var{c} matches the field of the opposite parity in the current frame
9782 @item @var{n} matches the field of the opposite parity in the next frame
9787 The @var{u} and @var{b} matching are a bit special in the sense that they match
9788 from the opposite parity flag. In the following examples, we assume that we are
9789 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
9790 'x' is placed above and below each matched fields.
9792 With bottom matching (@option{field}=@var{bottom}):
9797 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
9798 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
9806 With top matching (@option{field}=@var{top}):
9811 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
9812 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
9820 @subsection Examples
9822 Simple IVTC of a top field first telecined stream:
9824 fieldmatch=order=tff:combmatch=none, decimate
9827 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
9829 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
9834 Transform the field order of the input video.
9836 It accepts the following parameters:
9841 The output field order. Valid values are @var{tff} for top field first or @var{bff}
9842 for bottom field first.
9845 The default value is @samp{tff}.
9847 The transformation is done by shifting the picture content up or down
9848 by one line, and filling the remaining line with appropriate picture content.
9849 This method is consistent with most broadcast field order converters.
9851 If the input video is not flagged as being interlaced, or it is already
9852 flagged as being of the required output field order, then this filter does
9853 not alter the incoming video.
9855 It is very useful when converting to or from PAL DV material,
9856 which is bottom field first.
9860 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
9863 @section fifo, afifo
9865 Buffer input images and send them when they are requested.
9867 It is mainly useful when auto-inserted by the libavfilter
9870 It does not take parameters.
9872 @section fillborders
9874 Fill borders of the input video, without changing video stream dimensions.
9875 Sometimes video can have garbage at the four edges and you may not want to
9876 crop video input to keep size multiple of some number.
9878 This filter accepts the following options:
9882 Number of pixels to fill from left border.
9885 Number of pixels to fill from right border.
9888 Number of pixels to fill from top border.
9891 Number of pixels to fill from bottom border.
9896 It accepts the following values:
9899 fill pixels using outermost pixels
9902 fill pixels using mirroring
9905 fill pixels with constant value
9908 Default is @var{smear}.
9911 Set color for pixels in fixed mode. Default is @var{black}.
9916 Find a rectangular object
9918 It accepts the following options:
9922 Filepath of the object image, needs to be in gray8.
9925 Detection threshold, default is 0.5.
9928 Number of mipmaps, default is 3.
9930 @item xmin, ymin, xmax, ymax
9931 Specifies the rectangle in which to search.
9934 @subsection Examples
9938 Generate a representative palette of a given video using @command{ffmpeg}:
9940 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
9946 Cover a rectangular object
9948 It accepts the following options:
9952 Filepath of the optional cover image, needs to be in yuv420.
9957 It accepts the following values:
9960 cover it by the supplied image
9962 cover it by interpolating the surrounding pixels
9965 Default value is @var{blur}.
9968 @subsection Examples
9972 Generate a representative palette of a given video using @command{ffmpeg}:
9974 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
9980 Flood area with values of same pixel components with another values.
9982 It accepts the following options:
9985 Set pixel x coordinate.
9988 Set pixel y coordinate.
9991 Set source #0 component value.
9994 Set source #1 component value.
9997 Set source #2 component value.
10000 Set source #3 component value.
10003 Set destination #0 component value.
10006 Set destination #1 component value.
10009 Set destination #2 component value.
10012 Set destination #3 component value.
10018 Convert the input video to one of the specified pixel formats.
10019 Libavfilter will try to pick one that is suitable as input to
10022 It accepts the following parameters:
10026 A '|'-separated list of pixel format names, such as
10027 "pix_fmts=yuv420p|monow|rgb24".
10031 @subsection Examples
10035 Convert the input video to the @var{yuv420p} format
10037 format=pix_fmts=yuv420p
10040 Convert the input video to any of the formats in the list
10042 format=pix_fmts=yuv420p|yuv444p|yuv410p
10049 Convert the video to specified constant frame rate by duplicating or dropping
10050 frames as necessary.
10052 It accepts the following parameters:
10056 The desired output frame rate. The default is @code{25}.
10059 Assume the first PTS should be the given value, in seconds. This allows for
10060 padding/trimming at the start of stream. By default, no assumption is made
10061 about the first frame's expected PTS, so no padding or trimming is done.
10062 For example, this could be set to 0 to pad the beginning with duplicates of
10063 the first frame if a video stream starts after the audio stream or to trim any
10064 frames with a negative PTS.
10067 Timestamp (PTS) rounding method.
10069 Possible values are:
10076 round towards -infinity
10078 round towards +infinity
10082 The default is @code{near}.
10085 Action performed when reading the last frame.
10087 Possible values are:
10090 Use same timestamp rounding method as used for other frames.
10092 Pass through last frame if input duration has not been reached yet.
10094 The default is @code{round}.
10098 Alternatively, the options can be specified as a flat string:
10099 @var{fps}[:@var{start_time}[:@var{round}]].
10101 See also the @ref{setpts} filter.
10103 @subsection Examples
10107 A typical usage in order to set the fps to 25:
10113 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
10115 fps=fps=film:round=near
10121 Pack two different video streams into a stereoscopic video, setting proper
10122 metadata on supported codecs. The two views should have the same size and
10123 framerate and processing will stop when the shorter video ends. Please note
10124 that you may conveniently adjust view properties with the @ref{scale} and
10127 It accepts the following parameters:
10131 The desired packing format. Supported values are:
10136 The views are next to each other (default).
10139 The views are on top of each other.
10142 The views are packed by line.
10145 The views are packed by column.
10148 The views are temporally interleaved.
10157 # Convert left and right views into a frame-sequential video
10158 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
10160 # Convert views into a side-by-side video with the same output resolution as the input
10161 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
10166 Change the frame rate by interpolating new video output frames from the source
10169 This filter is not designed to function correctly with interlaced media. If
10170 you wish to change the frame rate of interlaced media then you are required
10171 to deinterlace before this filter and re-interlace after this filter.
10173 A description of the accepted options follows.
10177 Specify the output frames per second. This option can also be specified
10178 as a value alone. The default is @code{50}.
10181 Specify the start of a range where the output frame will be created as a
10182 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10183 the default is @code{15}.
10186 Specify the end of a range where the output frame will be created as a
10187 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10188 the default is @code{240}.
10191 Specify the level at which a scene change is detected as a value between
10192 0 and 100 to indicate a new scene; a low value reflects a low
10193 probability for the current frame to introduce a new scene, while a higher
10194 value means the current frame is more likely to be one.
10195 The default is @code{8.2}.
10198 Specify flags influencing the filter process.
10200 Available value for @var{flags} is:
10203 @item scene_change_detect, scd
10204 Enable scene change detection using the value of the option @var{scene}.
10205 This flag is enabled by default.
10211 Select one frame every N-th frame.
10213 This filter accepts the following option:
10216 Select frame after every @code{step} frames.
10217 Allowed values are positive integers higher than 0. Default value is @code{1}.
10220 @section freezedetect
10222 Detect frozen video.
10224 This filter logs a message and sets frame metadata when it detects that the
10225 input video has no significant change in content during a specified duration.
10226 Video freeze detection calculates the mean average absolute difference of all
10227 the components of video frames and compares it to a noise floor.
10229 The printed times and duration are expressed in seconds. The
10230 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
10231 whose timestamp equals or exceeds the detection duration and it contains the
10232 timestamp of the first frame of the freeze. The
10233 @code{lavfi.freezedetect.freeze_duration} and
10234 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
10237 The filter accepts the following options:
10241 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
10242 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
10246 Set freeze duration until notification (default is 2 seconds).
10252 Apply a frei0r effect to the input video.
10254 To enable the compilation of this filter, you need to install the frei0r
10255 header and configure FFmpeg with @code{--enable-frei0r}.
10257 It accepts the following parameters:
10262 The name of the frei0r effect to load. If the environment variable
10263 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
10264 directories specified by the colon-separated list in @env{FREI0R_PATH}.
10265 Otherwise, the standard frei0r paths are searched, in this order:
10266 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
10267 @file{/usr/lib/frei0r-1/}.
10269 @item filter_params
10270 A '|'-separated list of parameters to pass to the frei0r effect.
10274 A frei0r effect parameter can be a boolean (its value is either
10275 "y" or "n"), a double, a color (specified as
10276 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
10277 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
10278 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
10279 a position (specified as @var{X}/@var{Y}, where
10280 @var{X} and @var{Y} are floating point numbers) and/or a string.
10282 The number and types of parameters depend on the loaded effect. If an
10283 effect parameter is not specified, the default value is set.
10285 @subsection Examples
10289 Apply the distort0r effect, setting the first two double parameters:
10291 frei0r=filter_name=distort0r:filter_params=0.5|0.01
10295 Apply the colordistance effect, taking a color as the first parameter:
10297 frei0r=colordistance:0.2/0.3/0.4
10298 frei0r=colordistance:violet
10299 frei0r=colordistance:0x112233
10303 Apply the perspective effect, specifying the top left and top right image
10306 frei0r=perspective:0.2/0.2|0.8/0.2
10310 For more information, see
10311 @url{http://frei0r.dyne.org}
10315 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
10317 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
10318 processing filter, one of them is performed once per block, not per pixel.
10319 This allows for much higher speed.
10321 The filter accepts the following options:
10325 Set quality. This option defines the number of levels for averaging. It accepts
10326 an integer in the range 4-5. Default value is @code{4}.
10329 Force a constant quantization parameter. It accepts an integer in range 0-63.
10330 If not set, the filter will use the QP from the video stream (if available).
10333 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
10334 more details but also more artifacts, while higher values make the image smoother
10335 but also blurrier. Default value is @code{0} − PSNR optimal.
10337 @item use_bframe_qp
10338 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
10339 option may cause flicker since the B-Frames have often larger QP. Default is
10340 @code{0} (not enabled).
10346 Apply Gaussian blur filter.
10348 The filter accepts the following options:
10352 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
10355 Set number of steps for Gaussian approximation. Default is @code{1}.
10358 Set which planes to filter. By default all planes are filtered.
10361 Set vertical sigma, if negative it will be same as @code{sigma}.
10362 Default is @code{-1}.
10367 Apply generic equation to each pixel.
10369 The filter accepts the following options:
10372 @item lum_expr, lum
10373 Set the luminance expression.
10375 Set the chrominance blue expression.
10377 Set the chrominance red expression.
10378 @item alpha_expr, a
10379 Set the alpha expression.
10381 Set the red expression.
10382 @item green_expr, g
10383 Set the green expression.
10385 Set the blue expression.
10388 The colorspace is selected according to the specified options. If one
10389 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
10390 options is specified, the filter will automatically select a YCbCr
10391 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
10392 @option{blue_expr} options is specified, it will select an RGB
10395 If one of the chrominance expression is not defined, it falls back on the other
10396 one. If no alpha expression is specified it will evaluate to opaque value.
10397 If none of chrominance expressions are specified, they will evaluate
10398 to the luminance expression.
10400 The expressions can use the following variables and functions:
10404 The sequential number of the filtered frame, starting from @code{0}.
10408 The coordinates of the current sample.
10412 The width and height of the image.
10416 Width and height scale depending on the currently filtered plane. It is the
10417 ratio between the corresponding luma plane number of pixels and the current
10418 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
10419 @code{0.5,0.5} for chroma planes.
10422 Time of the current frame, expressed in seconds.
10425 Return the value of the pixel at location (@var{x},@var{y}) of the current
10429 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
10433 Return the value of the pixel at location (@var{x},@var{y}) of the
10434 blue-difference chroma plane. Return 0 if there is no such plane.
10437 Return the value of the pixel at location (@var{x},@var{y}) of the
10438 red-difference chroma plane. Return 0 if there is no such plane.
10443 Return the value of the pixel at location (@var{x},@var{y}) of the
10444 red/green/blue component. Return 0 if there is no such component.
10447 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
10448 plane. Return 0 if there is no such plane.
10451 For functions, if @var{x} and @var{y} are outside the area, the value will be
10452 automatically clipped to the closer edge.
10454 @subsection Examples
10458 Flip the image horizontally:
10464 Generate a bidimensional sine wave, with angle @code{PI/3} and a
10465 wavelength of 100 pixels:
10467 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
10471 Generate a fancy enigmatic moving light:
10473 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
10477 Generate a quick emboss effect:
10479 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
10483 Modify RGB components depending on pixel position:
10485 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
10489 Create a radial gradient that is the same size as the input (also see
10490 the @ref{vignette} filter):
10492 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
10498 Fix the banding artifacts that are sometimes introduced into nearly flat
10499 regions by truncation to 8-bit color depth.
10500 Interpolate the gradients that should go where the bands are, and
10503 It is designed for playback only. Do not use it prior to
10504 lossy compression, because compression tends to lose the dither and
10505 bring back the bands.
10507 It accepts the following parameters:
10512 The maximum amount by which the filter will change any one pixel. This is also
10513 the threshold for detecting nearly flat regions. Acceptable values range from
10514 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
10518 The neighborhood to fit the gradient to. A larger radius makes for smoother
10519 gradients, but also prevents the filter from modifying the pixels near detailed
10520 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
10521 values will be clipped to the valid range.
10525 Alternatively, the options can be specified as a flat string:
10526 @var{strength}[:@var{radius}]
10528 @subsection Examples
10532 Apply the filter with a @code{3.5} strength and radius of @code{8}:
10538 Specify radius, omitting the strength (which will fall-back to the default
10546 @section graphmonitor, agraphmonitor
10547 Show various filtergraph stats.
10549 With this filter one can debug complete filtergraph.
10550 Especially issues with links filling with queued frames.
10552 The filter accepts the following options:
10556 Set video output size. Default is @var{hd720}.
10559 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
10562 Set output mode, can be @var{fulll} or @var{compact}.
10563 In @var{compact} mode only filters with some queued frames have displayed stats.
10566 Set flags which enable which stats are shown in video.
10568 Available values for flags are:
10571 Display number of queued frames in each link.
10573 @item frame_count_in
10574 Display number of frames taken from filter.
10576 @item frame_count_out
10577 Display number of frames given out from filter.
10580 Display current filtered frame pts.
10583 Display current filtered frame time.
10586 Display time base for filter link.
10589 Display used format for filter link.
10592 Display video size or number of audio channels in case of audio used by filter link.
10595 Display video frame rate or sample rate in case of audio used by filter link.
10599 Set upper limit for video rate of output stream, Default value is @var{25}.
10600 This guarantee that output video frame rate will not be higher than this value.
10604 A color constancy variation filter which estimates scene illumination via grey edge algorithm
10605 and corrects the scene colors accordingly.
10607 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
10609 The filter accepts the following options:
10613 The order of differentiation to be applied on the scene. Must be chosen in the range
10614 [0,2] and default value is 1.
10617 The Minkowski parameter to be used for calculating the Minkowski distance. Must
10618 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
10619 max value instead of calculating Minkowski distance.
10622 The standard deviation of Gaussian blur to be applied on the scene. Must be
10623 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
10624 can't be equal to 0 if @var{difford} is greater than 0.
10627 @subsection Examples
10633 greyedge=difford=1:minknorm=5:sigma=2
10639 greyedge=difford=1:minknorm=0:sigma=2
10647 Apply a Hald CLUT to a video stream.
10649 First input is the video stream to process, and second one is the Hald CLUT.
10650 The Hald CLUT input can be a simple picture or a complete video stream.
10652 The filter accepts the following options:
10656 Force termination when the shortest input terminates. Default is @code{0}.
10658 Continue applying the last CLUT after the end of the stream. A value of
10659 @code{0} disable the filter after the last frame of the CLUT is reached.
10660 Default is @code{1}.
10663 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
10664 filters share the same internals).
10666 More information about the Hald CLUT can be found on Eskil Steenberg's website
10667 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
10669 @subsection Workflow examples
10671 @subsubsection Hald CLUT video stream
10673 Generate an identity Hald CLUT stream altered with various effects:
10675 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
10678 Note: make sure you use a lossless codec.
10680 Then use it with @code{haldclut} to apply it on some random stream:
10682 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
10685 The Hald CLUT will be applied to the 10 first seconds (duration of
10686 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
10687 to the remaining frames of the @code{mandelbrot} stream.
10689 @subsubsection Hald CLUT with preview
10691 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
10692 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
10693 biggest possible square starting at the top left of the picture. The remaining
10694 padding pixels (bottom or right) will be ignored. This area can be used to add
10695 a preview of the Hald CLUT.
10697 Typically, the following generated Hald CLUT will be supported by the
10698 @code{haldclut} filter:
10701 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
10702 pad=iw+320 [padded_clut];
10703 smptebars=s=320x256, split [a][b];
10704 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
10705 [main][b] overlay=W-320" -frames:v 1 clut.png
10708 It contains the original and a preview of the effect of the CLUT: SMPTE color
10709 bars are displayed on the right-top, and below the same color bars processed by
10712 Then, the effect of this Hald CLUT can be visualized with:
10714 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
10719 Flip the input video horizontally.
10721 For example, to horizontally flip the input video with @command{ffmpeg}:
10723 ffmpeg -i in.avi -vf "hflip" out.avi
10727 This filter applies a global color histogram equalization on a
10730 It can be used to correct video that has a compressed range of pixel
10731 intensities. The filter redistributes the pixel intensities to
10732 equalize their distribution across the intensity range. It may be
10733 viewed as an "automatically adjusting contrast filter". This filter is
10734 useful only for correcting degraded or poorly captured source
10737 The filter accepts the following options:
10741 Determine the amount of equalization to be applied. As the strength
10742 is reduced, the distribution of pixel intensities more-and-more
10743 approaches that of the input frame. The value must be a float number
10744 in the range [0,1] and defaults to 0.200.
10747 Set the maximum intensity that can generated and scale the output
10748 values appropriately. The strength should be set as desired and then
10749 the intensity can be limited if needed to avoid washing-out. The value
10750 must be a float number in the range [0,1] and defaults to 0.210.
10753 Set the antibanding level. If enabled the filter will randomly vary
10754 the luminance of output pixels by a small amount to avoid banding of
10755 the histogram. Possible values are @code{none}, @code{weak} or
10756 @code{strong}. It defaults to @code{none}.
10761 Compute and draw a color distribution histogram for the input video.
10763 The computed histogram is a representation of the color component
10764 distribution in an image.
10766 Standard histogram displays the color components distribution in an image.
10767 Displays color graph for each color component. Shows distribution of
10768 the Y, U, V, A or R, G, B components, depending on input format, in the
10769 current frame. Below each graph a color component scale meter is shown.
10771 The filter accepts the following options:
10775 Set height of level. Default value is @code{200}.
10776 Allowed range is [50, 2048].
10779 Set height of color scale. Default value is @code{12}.
10780 Allowed range is [0, 40].
10784 It accepts the following values:
10787 Per color component graphs are placed below each other.
10790 Per color component graphs are placed side by side.
10793 Presents information identical to that in the @code{parade}, except
10794 that the graphs representing color components are superimposed directly
10797 Default is @code{stack}.
10800 Set mode. Can be either @code{linear}, or @code{logarithmic}.
10801 Default is @code{linear}.
10804 Set what color components to display.
10805 Default is @code{7}.
10808 Set foreground opacity. Default is @code{0.7}.
10811 Set background opacity. Default is @code{0.5}.
10814 @subsection Examples
10819 Calculate and draw histogram:
10821 ffplay -i input -vf histogram
10829 This is a high precision/quality 3d denoise filter. It aims to reduce
10830 image noise, producing smooth images and making still images really
10831 still. It should enhance compressibility.
10833 It accepts the following optional parameters:
10837 A non-negative floating point number which specifies spatial luma strength.
10838 It defaults to 4.0.
10840 @item chroma_spatial
10841 A non-negative floating point number which specifies spatial chroma strength.
10842 It defaults to 3.0*@var{luma_spatial}/4.0.
10845 A floating point number which specifies luma temporal strength. It defaults to
10846 6.0*@var{luma_spatial}/4.0.
10849 A floating point number which specifies chroma temporal strength. It defaults to
10850 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
10853 @anchor{hwdownload}
10854 @section hwdownload
10856 Download hardware frames to system memory.
10858 The input must be in hardware frames, and the output a non-hardware format.
10859 Not all formats will be supported on the output - it may be necessary to insert
10860 an additional @option{format} filter immediately following in the graph to get
10861 the output in a supported format.
10865 Map hardware frames to system memory or to another device.
10867 This filter has several different modes of operation; which one is used depends
10868 on the input and output formats:
10871 Hardware frame input, normal frame output
10873 Map the input frames to system memory and pass them to the output. If the
10874 original hardware frame is later required (for example, after overlaying
10875 something else on part of it), the @option{hwmap} filter can be used again
10876 in the next mode to retrieve it.
10878 Normal frame input, hardware frame output
10880 If the input is actually a software-mapped hardware frame, then unmap it -
10881 that is, return the original hardware frame.
10883 Otherwise, a device must be provided. Create new hardware surfaces on that
10884 device for the output, then map them back to the software format at the input
10885 and give those frames to the preceding filter. This will then act like the
10886 @option{hwupload} filter, but may be able to avoid an additional copy when
10887 the input is already in a compatible format.
10889 Hardware frame input and output
10891 A device must be supplied for the output, either directly or with the
10892 @option{derive_device} option. The input and output devices must be of
10893 different types and compatible - the exact meaning of this is
10894 system-dependent, but typically it means that they must refer to the same
10895 underlying hardware context (for example, refer to the same graphics card).
10897 If the input frames were originally created on the output device, then unmap
10898 to retrieve the original frames.
10900 Otherwise, map the frames to the output device - create new hardware frames
10901 on the output corresponding to the frames on the input.
10904 The following additional parameters are accepted:
10908 Set the frame mapping mode. Some combination of:
10911 The mapped frame should be readable.
10913 The mapped frame should be writeable.
10915 The mapping will always overwrite the entire frame.
10917 This may improve performance in some cases, as the original contents of the
10918 frame need not be loaded.
10920 The mapping must not involve any copying.
10922 Indirect mappings to copies of frames are created in some cases where either
10923 direct mapping is not possible or it would have unexpected properties.
10924 Setting this flag ensures that the mapping is direct and will fail if that is
10927 Defaults to @var{read+write} if not specified.
10929 @item derive_device @var{type}
10930 Rather than using the device supplied at initialisation, instead derive a new
10931 device of type @var{type} from the device the input frames exist on.
10934 In a hardware to hardware mapping, map in reverse - create frames in the sink
10935 and map them back to the source. This may be necessary in some cases where
10936 a mapping in one direction is required but only the opposite direction is
10937 supported by the devices being used.
10939 This option is dangerous - it may break the preceding filter in undefined
10940 ways if there are any additional constraints on that filter's output.
10941 Do not use it without fully understanding the implications of its use.
10947 Upload system memory frames to hardware surfaces.
10949 The device to upload to must be supplied when the filter is initialised. If
10950 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
10953 @anchor{hwupload_cuda}
10954 @section hwupload_cuda
10956 Upload system memory frames to a CUDA device.
10958 It accepts the following optional parameters:
10962 The number of the CUDA device to use
10967 Apply a high-quality magnification filter designed for pixel art. This filter
10968 was originally created by Maxim Stepin.
10970 It accepts the following option:
10974 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
10975 @code{hq3x} and @code{4} for @code{hq4x}.
10976 Default is @code{3}.
10980 Stack input videos horizontally.
10982 All streams must be of same pixel format and of same height.
10984 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
10985 to create same output.
10987 The filter accept the following option:
10991 Set number of input streams. Default is 2.
10994 If set to 1, force the output to terminate when the shortest input
10995 terminates. Default value is 0.
11000 Modify the hue and/or the saturation of the input.
11002 It accepts the following parameters:
11006 Specify the hue angle as a number of degrees. It accepts an expression,
11007 and defaults to "0".
11010 Specify the saturation in the [-10,10] range. It accepts an expression and
11014 Specify the hue angle as a number of radians. It accepts an
11015 expression, and defaults to "0".
11018 Specify the brightness in the [-10,10] range. It accepts an expression and
11022 @option{h} and @option{H} are mutually exclusive, and can't be
11023 specified at the same time.
11025 The @option{b}, @option{h}, @option{H} and @option{s} option values are
11026 expressions containing the following constants:
11030 frame count of the input frame starting from 0
11033 presentation timestamp of the input frame expressed in time base units
11036 frame rate of the input video, NAN if the input frame rate is unknown
11039 timestamp expressed in seconds, NAN if the input timestamp is unknown
11042 time base of the input video
11045 @subsection Examples
11049 Set the hue to 90 degrees and the saturation to 1.0:
11055 Same command but expressing the hue in radians:
11061 Rotate hue and make the saturation swing between 0
11062 and 2 over a period of 1 second:
11064 hue="H=2*PI*t: s=sin(2*PI*t)+1"
11068 Apply a 3 seconds saturation fade-in effect starting at 0:
11070 hue="s=min(t/3\,1)"
11073 The general fade-in expression can be written as:
11075 hue="s=min(0\, max((t-START)/DURATION\, 1))"
11079 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
11081 hue="s=max(0\, min(1\, (8-t)/3))"
11084 The general fade-out expression can be written as:
11086 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
11091 @subsection Commands
11093 This filter supports the following commands:
11099 Modify the hue and/or the saturation and/or brightness of the input video.
11100 The command accepts the same syntax of the corresponding option.
11102 If the specified expression is not valid, it is kept at its current
11106 @section hysteresis
11108 Grow first stream into second stream by connecting components.
11109 This makes it possible to build more robust edge masks.
11111 This filter accepts the following options:
11115 Set which planes will be processed as bitmap, unprocessed planes will be
11116 copied from first stream.
11117 By default value 0xf, all planes will be processed.
11120 Set threshold which is used in filtering. If pixel component value is higher than
11121 this value filter algorithm for connecting components is activated.
11122 By default value is 0.
11127 Detect video interlacing type.
11129 This filter tries to detect if the input frames are interlaced, progressive,
11130 top or bottom field first. It will also try to detect fields that are
11131 repeated between adjacent frames (a sign of telecine).
11133 Single frame detection considers only immediately adjacent frames when classifying each frame.
11134 Multiple frame detection incorporates the classification history of previous frames.
11136 The filter will log these metadata values:
11139 @item single.current_frame
11140 Detected type of current frame using single-frame detection. One of:
11141 ``tff'' (top field first), ``bff'' (bottom field first),
11142 ``progressive'', or ``undetermined''
11145 Cumulative number of frames detected as top field first using single-frame detection.
11148 Cumulative number of frames detected as top field first using multiple-frame detection.
11151 Cumulative number of frames detected as bottom field first using single-frame detection.
11153 @item multiple.current_frame
11154 Detected type of current frame using multiple-frame detection. One of:
11155 ``tff'' (top field first), ``bff'' (bottom field first),
11156 ``progressive'', or ``undetermined''
11159 Cumulative number of frames detected as bottom field first using multiple-frame detection.
11161 @item single.progressive
11162 Cumulative number of frames detected as progressive using single-frame detection.
11164 @item multiple.progressive
11165 Cumulative number of frames detected as progressive using multiple-frame detection.
11167 @item single.undetermined
11168 Cumulative number of frames that could not be classified using single-frame detection.
11170 @item multiple.undetermined
11171 Cumulative number of frames that could not be classified using multiple-frame detection.
11173 @item repeated.current_frame
11174 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
11176 @item repeated.neither
11177 Cumulative number of frames with no repeated field.
11180 Cumulative number of frames with the top field repeated from the previous frame's top field.
11182 @item repeated.bottom
11183 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
11186 The filter accepts the following options:
11190 Set interlacing threshold.
11192 Set progressive threshold.
11194 Threshold for repeated field detection.
11196 Number of frames after which a given frame's contribution to the
11197 statistics is halved (i.e., it contributes only 0.5 to its
11198 classification). The default of 0 means that all frames seen are given
11199 full weight of 1.0 forever.
11200 @item analyze_interlaced_flag
11201 When this is not 0 then idet will use the specified number of frames to determine
11202 if the interlaced flag is accurate, it will not count undetermined frames.
11203 If the flag is found to be accurate it will be used without any further
11204 computations, if it is found to be inaccurate it will be cleared without any
11205 further computations. This allows inserting the idet filter as a low computational
11206 method to clean up the interlaced flag
11211 Deinterleave or interleave fields.
11213 This filter allows one to process interlaced images fields without
11214 deinterlacing them. Deinterleaving splits the input frame into 2
11215 fields (so called half pictures). Odd lines are moved to the top
11216 half of the output image, even lines to the bottom half.
11217 You can process (filter) them independently and then re-interleave them.
11219 The filter accepts the following options:
11223 @item chroma_mode, c
11224 @item alpha_mode, a
11225 Available values for @var{luma_mode}, @var{chroma_mode} and
11226 @var{alpha_mode} are:
11232 @item deinterleave, d
11233 Deinterleave fields, placing one above the other.
11235 @item interleave, i
11236 Interleave fields. Reverse the effect of deinterleaving.
11238 Default value is @code{none}.
11240 @item luma_swap, ls
11241 @item chroma_swap, cs
11242 @item alpha_swap, as
11243 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
11248 Apply inflate effect to the video.
11250 This filter replaces the pixel by the local(3x3) average by taking into account
11251 only values higher than the pixel.
11253 It accepts the following options:
11260 Limit the maximum change for each plane, default is 65535.
11261 If 0, plane will remain unchanged.
11266 Simple interlacing filter from progressive contents. This interleaves upper (or
11267 lower) lines from odd frames with lower (or upper) lines from even frames,
11268 halving the frame rate and preserving image height.
11271 Original Original New Frame
11272 Frame 'j' Frame 'j+1' (tff)
11273 ========== =========== ==================
11274 Line 0 --------------------> Frame 'j' Line 0
11275 Line 1 Line 1 ----> Frame 'j+1' Line 1
11276 Line 2 ---------------------> Frame 'j' Line 2
11277 Line 3 Line 3 ----> Frame 'j+1' Line 3
11279 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
11282 It accepts the following optional parameters:
11286 This determines whether the interlaced frame is taken from the even
11287 (tff - default) or odd (bff) lines of the progressive frame.
11290 Vertical lowpass filter to avoid twitter interlacing and
11291 reduce moire patterns.
11295 Disable vertical lowpass filter
11298 Enable linear filter (default)
11301 Enable complex filter. This will slightly less reduce twitter and moire
11302 but better retain detail and subjective sharpness impression.
11309 Deinterlace input video by applying Donald Graft's adaptive kernel
11310 deinterling. Work on interlaced parts of a video to produce
11311 progressive frames.
11313 The description of the accepted parameters follows.
11317 Set the threshold which affects the filter's tolerance when
11318 determining if a pixel line must be processed. It must be an integer
11319 in the range [0,255] and defaults to 10. A value of 0 will result in
11320 applying the process on every pixels.
11323 Paint pixels exceeding the threshold value to white if set to 1.
11327 Set the fields order. Swap fields if set to 1, leave fields alone if
11331 Enable additional sharpening if set to 1. Default is 0.
11334 Enable twoway sharpening if set to 1. Default is 0.
11337 @subsection Examples
11341 Apply default values:
11343 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
11347 Enable additional sharpening:
11353 Paint processed pixels in white:
11361 Slowly update darker pixels.
11363 This filter makes short flashes of light appear longer.
11364 This filter accepts the following options:
11368 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
11371 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
11374 @section lenscorrection
11376 Correct radial lens distortion
11378 This filter can be used to correct for radial distortion as can result from the use
11379 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
11380 one can use tools available for example as part of opencv or simply trial-and-error.
11381 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
11382 and extract the k1 and k2 coefficients from the resulting matrix.
11384 Note that effectively the same filter is available in the open-source tools Krita and
11385 Digikam from the KDE project.
11387 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
11388 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
11389 brightness distribution, so you may want to use both filters together in certain
11390 cases, though you will have to take care of ordering, i.e. whether vignetting should
11391 be applied before or after lens correction.
11393 @subsection Options
11395 The filter accepts the following options:
11399 Relative x-coordinate of the focal point of the image, and thereby the center of the
11400 distortion. This value has a range [0,1] and is expressed as fractions of the image
11401 width. Default is 0.5.
11403 Relative y-coordinate of the focal point of the image, and thereby the center of the
11404 distortion. This value has a range [0,1] and is expressed as fractions of the image
11405 height. Default is 0.5.
11407 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
11408 no correction. Default is 0.
11410 Coefficient of the double quadratic correction term. This value has a range [-1,1].
11411 0 means no correction. Default is 0.
11414 The formula that generates the correction is:
11416 @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)
11418 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
11419 distances from the focal point in the source and target images, respectively.
11423 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
11425 The @code{lensfun} filter requires the camera make, camera model, and lens model
11426 to apply the lens correction. The filter will load the lensfun database and
11427 query it to find the corresponding camera and lens entries in the database. As
11428 long as these entries can be found with the given options, the filter can
11429 perform corrections on frames. Note that incomplete strings will result in the
11430 filter choosing the best match with the given options, and the filter will
11431 output the chosen camera and lens models (logged with level "info"). You must
11432 provide the make, camera model, and lens model as they are required.
11434 The filter accepts the following options:
11438 The make of the camera (for example, "Canon"). This option is required.
11441 The model of the camera (for example, "Canon EOS 100D"). This option is
11445 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
11446 option is required.
11449 The type of correction to apply. The following values are valid options:
11453 Enables fixing lens vignetting.
11456 Enables fixing lens geometry. This is the default.
11459 Enables fixing chromatic aberrations.
11462 Enables fixing lens vignetting and lens geometry.
11465 Enables fixing lens vignetting and chromatic aberrations.
11468 Enables fixing both lens geometry and chromatic aberrations.
11471 Enables all possible corrections.
11475 The focal length of the image/video (zoom; expected constant for video). For
11476 example, a 18--55mm lens has focal length range of [18--55], so a value in that
11477 range should be chosen when using that lens. Default 18.
11480 The aperture of the image/video (expected constant for video). Note that
11481 aperture is only used for vignetting correction. Default 3.5.
11483 @item focus_distance
11484 The focus distance of the image/video (expected constant for video). Note that
11485 focus distance is only used for vignetting and only slightly affects the
11486 vignetting correction process. If unknown, leave it at the default value (which
11490 The scale factor which is applied after transformation. After correction the
11491 video is no longer necessarily rectangular. This parameter controls how much of
11492 the resulting image is visible. The value 0 means that a value will be chosen
11493 automatically such that there is little or no unmapped area in the output
11494 image. 1.0 means that no additional scaling is done. Lower values may result
11495 in more of the corrected image being visible, while higher values may avoid
11496 unmapped areas in the output.
11498 @item target_geometry
11499 The target geometry of the output image/video. The following values are valid
11503 @item rectilinear (default)
11506 @item equirectangular
11507 @item fisheye_orthographic
11508 @item fisheye_stereographic
11509 @item fisheye_equisolid
11510 @item fisheye_thoby
11513 Apply the reverse of image correction (instead of correcting distortion, apply
11516 @item interpolation
11517 The type of interpolation used when correcting distortion. The following values
11522 @item linear (default)
11527 @subsection Examples
11531 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
11532 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
11536 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
11540 Apply the same as before, but only for the first 5 seconds of video.
11543 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
11550 Obtain the VMAF (Video Multi-Method Assessment Fusion)
11551 score between two input videos.
11553 The obtained VMAF score is printed through the logging system.
11555 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
11556 After installing the library it can be enabled using:
11557 @code{./configure --enable-libvmaf --enable-version3}.
11558 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
11560 The filter has following options:
11564 Set the model path which is to be used for SVM.
11565 Default value: @code{"vmaf_v0.6.1.pkl"}
11568 Set the file path to be used to store logs.
11571 Set the format of the log file (xml or json).
11573 @item enable_transform
11574 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
11575 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
11576 Default value: @code{false}
11579 Invokes the phone model which will generate VMAF scores higher than in the
11580 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
11583 Enables computing psnr along with vmaf.
11586 Enables computing ssim along with vmaf.
11589 Enables computing ms_ssim along with vmaf.
11592 Set the pool method (mean, min or harmonic mean) to be used for computing vmaf.
11595 Set number of threads to be used when computing vmaf.
11598 Set interval for frame subsampling used when computing vmaf.
11600 @item enable_conf_interval
11601 Enables confidence interval.
11604 This filter also supports the @ref{framesync} options.
11606 On the below examples the input file @file{main.mpg} being processed is
11607 compared with the reference file @file{ref.mpg}.
11610 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
11613 Example with options:
11615 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
11620 Limits the pixel components values to the specified range [min, max].
11622 The filter accepts the following options:
11626 Lower bound. Defaults to the lowest allowed value for the input.
11629 Upper bound. Defaults to the highest allowed value for the input.
11632 Specify which planes will be processed. Defaults to all available.
11639 The filter accepts the following options:
11643 Set the number of loops. Setting this value to -1 will result in infinite loops.
11647 Set maximal size in number of frames. Default is 0.
11650 Set first frame of loop. Default is 0.
11653 @subsection Examples
11657 Loop single first frame infinitely:
11659 loop=loop=-1:size=1:start=0
11663 Loop single first frame 10 times:
11665 loop=loop=10:size=1:start=0
11669 Loop 10 first frames 5 times:
11671 loop=loop=5:size=10:start=0
11677 Apply a 1D LUT to an input video.
11679 The filter accepts the following options:
11683 Set the 1D LUT file name.
11685 Currently supported formats:
11694 Select interpolation mode.
11696 Available values are:
11700 Use values from the nearest defined point.
11702 Interpolate values using the linear interpolation.
11704 Interpolate values using the cosine interpolation.
11706 Interpolate values using the cubic interpolation.
11708 Interpolate values using the spline interpolation.
11715 Apply a 3D LUT to an input video.
11717 The filter accepts the following options:
11721 Set the 3D LUT file name.
11723 Currently supported formats:
11737 Select interpolation mode.
11739 Available values are:
11743 Use values from the nearest defined point.
11745 Interpolate values using the 8 points defining a cube.
11747 Interpolate values using a tetrahedron.
11751 This filter also supports the @ref{framesync} options.
11755 Turn certain luma values into transparency.
11757 The filter accepts the following options:
11761 Set the luma which will be used as base for transparency.
11762 Default value is @code{0}.
11765 Set the range of luma values to be keyed out.
11766 Default value is @code{0}.
11769 Set the range of softness. Default value is @code{0}.
11770 Use this to control gradual transition from zero to full transparency.
11773 @section lut, lutrgb, lutyuv
11775 Compute a look-up table for binding each pixel component input value
11776 to an output value, and apply it to the input video.
11778 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
11779 to an RGB input video.
11781 These filters accept the following parameters:
11784 set first pixel component expression
11786 set second pixel component expression
11788 set third pixel component expression
11790 set fourth pixel component expression, corresponds to the alpha component
11793 set red component expression
11795 set green component expression
11797 set blue component expression
11799 alpha component expression
11802 set Y/luminance component expression
11804 set U/Cb component expression
11806 set V/Cr component expression
11809 Each of them specifies the expression to use for computing the lookup table for
11810 the corresponding pixel component values.
11812 The exact component associated to each of the @var{c*} options depends on the
11815 The @var{lut} filter requires either YUV or RGB pixel formats in input,
11816 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
11818 The expressions can contain the following constants and functions:
11823 The input width and height.
11826 The input value for the pixel component.
11829 The input value, clipped to the @var{minval}-@var{maxval} range.
11832 The maximum value for the pixel component.
11835 The minimum value for the pixel component.
11838 The negated value for the pixel component value, clipped to the
11839 @var{minval}-@var{maxval} range; it corresponds to the expression
11840 "maxval-clipval+minval".
11843 The computed value in @var{val}, clipped to the
11844 @var{minval}-@var{maxval} range.
11846 @item gammaval(gamma)
11847 The computed gamma correction value of the pixel component value,
11848 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
11850 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
11854 All expressions default to "val".
11856 @subsection Examples
11860 Negate input video:
11862 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
11863 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
11866 The above is the same as:
11868 lutrgb="r=negval:g=negval:b=negval"
11869 lutyuv="y=negval:u=negval:v=negval"
11879 Remove chroma components, turning the video into a graytone image:
11881 lutyuv="u=128:v=128"
11885 Apply a luma burning effect:
11891 Remove green and blue components:
11897 Set a constant alpha channel value on input:
11899 format=rgba,lutrgb=a="maxval-minval/2"
11903 Correct luminance gamma by a factor of 0.5:
11905 lutyuv=y=gammaval(0.5)
11909 Discard least significant bits of luma:
11911 lutyuv=y='bitand(val, 128+64+32)'
11915 Technicolor like effect:
11917 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
11921 @section lut2, tlut2
11923 The @code{lut2} filter takes two input streams and outputs one
11926 The @code{tlut2} (time lut2) filter takes two consecutive frames
11927 from one single stream.
11929 This filter accepts the following parameters:
11932 set first pixel component expression
11934 set second pixel component expression
11936 set third pixel component expression
11938 set fourth pixel component expression, corresponds to the alpha component
11941 set output bit depth, only available for @code{lut2} filter. By default is 0,
11942 which means bit depth is automatically picked from first input format.
11945 Each of them specifies the expression to use for computing the lookup table for
11946 the corresponding pixel component values.
11948 The exact component associated to each of the @var{c*} options depends on the
11951 The expressions can contain the following constants:
11956 The input width and height.
11959 The first input value for the pixel component.
11962 The second input value for the pixel component.
11965 The first input video bit depth.
11968 The second input video bit depth.
11971 All expressions default to "x".
11973 @subsection Examples
11977 Highlight differences between two RGB video streams:
11979 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)'
11983 Highlight differences between two YUV video streams:
11985 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)'
11989 Show max difference between two video streams:
11991 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)))'
11995 @section maskedclamp
11997 Clamp the first input stream with the second input and third input stream.
11999 Returns the value of first stream to be between second input
12000 stream - @code{undershoot} and third input stream + @code{overshoot}.
12002 This filter accepts the following options:
12005 Default value is @code{0}.
12008 Default value is @code{0}.
12011 Set which planes will be processed as bitmap, unprocessed planes will be
12012 copied from first stream.
12013 By default value 0xf, all planes will be processed.
12016 @section maskedmerge
12018 Merge the first input stream with the second input stream using per pixel
12019 weights in the third input stream.
12021 A value of 0 in the third stream pixel component means that pixel component
12022 from first stream is returned unchanged, while maximum value (eg. 255 for
12023 8-bit videos) means that pixel component from second stream is returned
12024 unchanged. Intermediate values define the amount of merging between both
12025 input stream's pixel components.
12027 This filter accepts the following options:
12030 Set which planes will be processed as bitmap, unprocessed planes will be
12031 copied from first stream.
12032 By default value 0xf, all planes will be processed.
12036 Create mask from input video.
12038 For example it is useful to create motion masks after @code{tblend} filter.
12040 This filter accepts the following options:
12044 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
12047 Set high threshold. Any pixel component higher than this value will be set to max value
12048 allowed for current pixel format.
12051 Set planes to filter, by default all available planes are filtered.
12054 Fill all frame pixels with this value.
12057 Set max average pixel value for frame. If sum of all pixel components is higher that this
12058 average, output frame will be completely filled with value set by @var{fill} option.
12059 Typically useful for scene changes when used in combination with @code{tblend} filter.
12064 Apply motion-compensation deinterlacing.
12066 It needs one field per frame as input and must thus be used together
12067 with yadif=1/3 or equivalent.
12069 This filter accepts the following options:
12072 Set the deinterlacing mode.
12074 It accepts one of the following values:
12079 use iterative motion estimation
12081 like @samp{slow}, but use multiple reference frames.
12083 Default value is @samp{fast}.
12086 Set the picture field parity assumed for the input video. It must be
12087 one of the following values:
12091 assume top field first
12093 assume bottom field first
12096 Default value is @samp{bff}.
12099 Set per-block quantization parameter (QP) used by the internal
12102 Higher values should result in a smoother motion vector field but less
12103 optimal individual vectors. Default value is 1.
12106 @section mergeplanes
12108 Merge color channel components from several video streams.
12110 The filter accepts up to 4 input streams, and merge selected input
12111 planes to the output video.
12113 This filter accepts the following options:
12116 Set input to output plane mapping. Default is @code{0}.
12118 The mappings is specified as a bitmap. It should be specified as a
12119 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
12120 mapping for the first plane of the output stream. 'A' sets the number of
12121 the input stream to use (from 0 to 3), and 'a' the plane number of the
12122 corresponding input to use (from 0 to 3). The rest of the mappings is
12123 similar, 'Bb' describes the mapping for the output stream second
12124 plane, 'Cc' describes the mapping for the output stream third plane and
12125 'Dd' describes the mapping for the output stream fourth plane.
12128 Set output pixel format. Default is @code{yuva444p}.
12131 @subsection Examples
12135 Merge three gray video streams of same width and height into single video stream:
12137 [a0][a1][a2]mergeplanes=0x001020:yuv444p
12141 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
12143 [a0][a1]mergeplanes=0x00010210:yuva444p
12147 Swap Y and A plane in yuva444p stream:
12149 format=yuva444p,mergeplanes=0x03010200:yuva444p
12153 Swap U and V plane in yuv420p stream:
12155 format=yuv420p,mergeplanes=0x000201:yuv420p
12159 Cast a rgb24 clip to yuv444p:
12161 format=rgb24,mergeplanes=0x000102:yuv444p
12167 Estimate and export motion vectors using block matching algorithms.
12168 Motion vectors are stored in frame side data to be used by other filters.
12170 This filter accepts the following options:
12173 Specify the motion estimation method. Accepts one of the following values:
12177 Exhaustive search algorithm.
12179 Three step search algorithm.
12181 Two dimensional logarithmic search algorithm.
12183 New three step search algorithm.
12185 Four step search algorithm.
12187 Diamond search algorithm.
12189 Hexagon-based search algorithm.
12191 Enhanced predictive zonal search algorithm.
12193 Uneven multi-hexagon search algorithm.
12195 Default value is @samp{esa}.
12198 Macroblock size. Default @code{16}.
12201 Search parameter. Default @code{7}.
12204 @section midequalizer
12206 Apply Midway Image Equalization effect using two video streams.
12208 Midway Image Equalization adjusts a pair of images to have the same
12209 histogram, while maintaining their dynamics as much as possible. It's
12210 useful for e.g. matching exposures from a pair of stereo cameras.
12212 This filter has two inputs and one output, which must be of same pixel format, but
12213 may be of different sizes. The output of filter is first input adjusted with
12214 midway histogram of both inputs.
12216 This filter accepts the following option:
12220 Set which planes to process. Default is @code{15}, which is all available planes.
12223 @section minterpolate
12225 Convert the video to specified frame rate using motion interpolation.
12227 This filter accepts the following options:
12230 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}.
12233 Motion interpolation mode. Following values are accepted:
12236 Duplicate previous or next frame for interpolating new ones.
12238 Blend source frames. Interpolated frame is mean of previous and next frames.
12240 Motion compensated interpolation. Following options are effective when this mode is selected:
12244 Motion compensation mode. Following values are accepted:
12247 Overlapped block motion compensation.
12249 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
12251 Default mode is @samp{obmc}.
12254 Motion estimation mode. Following values are accepted:
12257 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
12259 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
12261 Default mode is @samp{bilat}.
12264 The algorithm to be used for motion estimation. Following values are accepted:
12267 Exhaustive search algorithm.
12269 Three step search algorithm.
12271 Two dimensional logarithmic search algorithm.
12273 New three step search algorithm.
12275 Four step search algorithm.
12277 Diamond search algorithm.
12279 Hexagon-based search algorithm.
12281 Enhanced predictive zonal search algorithm.
12283 Uneven multi-hexagon search algorithm.
12285 Default algorithm is @samp{epzs}.
12288 Macroblock size. Default @code{16}.
12291 Motion estimation search parameter. Default @code{32}.
12294 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).
12299 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:
12302 Disable scene change detection.
12304 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
12306 Default method is @samp{fdiff}.
12308 @item scd_threshold
12309 Scene change detection threshold. Default is @code{5.0}.
12314 Mix several video input streams into one video stream.
12316 A description of the accepted options follows.
12320 The number of inputs. If unspecified, it defaults to 2.
12323 Specify weight of each input video stream as sequence.
12324 Each weight is separated by space. If number of weights
12325 is smaller than number of @var{frames} last specified
12326 weight will be used for all remaining unset weights.
12329 Specify scale, if it is set it will be multiplied with sum
12330 of each weight multiplied with pixel values to give final destination
12331 pixel value. By default @var{scale} is auto scaled to sum of weights.
12334 Specify how end of stream is determined.
12337 The duration of the longest input. (default)
12340 The duration of the shortest input.
12343 The duration of the first input.
12347 @section mpdecimate
12349 Drop frames that do not differ greatly from the previous frame in
12350 order to reduce frame rate.
12352 The main use of this filter is for very-low-bitrate encoding
12353 (e.g. streaming over dialup modem), but it could in theory be used for
12354 fixing movies that were inverse-telecined incorrectly.
12356 A description of the accepted options follows.
12360 Set the maximum number of consecutive frames which can be dropped (if
12361 positive), or the minimum interval between dropped frames (if
12362 negative). If the value is 0, the frame is dropped disregarding the
12363 number of previous sequentially dropped frames.
12365 Default value is 0.
12370 Set the dropping threshold values.
12372 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
12373 represent actual pixel value differences, so a threshold of 64
12374 corresponds to 1 unit of difference for each pixel, or the same spread
12375 out differently over the block.
12377 A frame is a candidate for dropping if no 8x8 blocks differ by more
12378 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
12379 meaning the whole image) differ by more than a threshold of @option{lo}.
12381 Default value for @option{hi} is 64*12, default value for @option{lo} is
12382 64*5, and default value for @option{frac} is 0.33.
12388 Negate (invert) the input video.
12390 It accepts the following option:
12395 With value 1, it negates the alpha component, if present. Default value is 0.
12401 Denoise frames using Non-Local Means algorithm.
12403 Each pixel is adjusted by looking for other pixels with similar contexts. This
12404 context similarity is defined by comparing their surrounding patches of size
12405 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
12408 Note that the research area defines centers for patches, which means some
12409 patches will be made of pixels outside that research area.
12411 The filter accepts the following options.
12415 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
12418 Set patch size. Default is 7. Must be odd number in range [0, 99].
12421 Same as @option{p} but for chroma planes.
12423 The default value is @var{0} and means automatic.
12426 Set research size. Default is 15. Must be odd number in range [0, 99].
12429 Same as @option{r} but for chroma planes.
12431 The default value is @var{0} and means automatic.
12436 Deinterlace video using neural network edge directed interpolation.
12438 This filter accepts the following options:
12442 Mandatory option, without binary file filter can not work.
12443 Currently file can be found here:
12444 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
12447 Set which frames to deinterlace, by default it is @code{all}.
12448 Can be @code{all} or @code{interlaced}.
12451 Set mode of operation.
12453 Can be one of the following:
12457 Use frame flags, both fields.
12459 Use frame flags, single field.
12461 Use top field only.
12463 Use bottom field only.
12465 Use both fields, top first.
12467 Use both fields, bottom first.
12471 Set which planes to process, by default filter process all frames.
12474 Set size of local neighborhood around each pixel, used by the predictor neural
12477 Can be one of the following:
12490 Set the number of neurons in predictor neural network.
12491 Can be one of the following:
12502 Controls the number of different neural network predictions that are blended
12503 together to compute the final output value. Can be @code{fast}, default or
12507 Set which set of weights to use in the predictor.
12508 Can be one of the following:
12512 weights trained to minimize absolute error
12514 weights trained to minimize squared error
12518 Controls whether or not the prescreener neural network is used to decide
12519 which pixels should be processed by the predictor neural network and which
12520 can be handled by simple cubic interpolation.
12521 The prescreener is trained to know whether cubic interpolation will be
12522 sufficient for a pixel or whether it should be predicted by the predictor nn.
12523 The computational complexity of the prescreener nn is much less than that of
12524 the predictor nn. Since most pixels can be handled by cubic interpolation,
12525 using the prescreener generally results in much faster processing.
12526 The prescreener is pretty accurate, so the difference between using it and not
12527 using it is almost always unnoticeable.
12529 Can be one of the following:
12537 Default is @code{new}.
12540 Set various debugging flags.
12545 Force libavfilter not to use any of the specified pixel formats for the
12546 input to the next filter.
12548 It accepts the following parameters:
12552 A '|'-separated list of pixel format names, such as
12553 pix_fmts=yuv420p|monow|rgb24".
12557 @subsection Examples
12561 Force libavfilter to use a format different from @var{yuv420p} for the
12562 input to the vflip filter:
12564 noformat=pix_fmts=yuv420p,vflip
12568 Convert the input video to any of the formats not contained in the list:
12570 noformat=yuv420p|yuv444p|yuv410p
12576 Add noise on video input frame.
12578 The filter accepts the following options:
12586 Set noise seed for specific pixel component or all pixel components in case
12587 of @var{all_seed}. Default value is @code{123457}.
12589 @item all_strength, alls
12590 @item c0_strength, c0s
12591 @item c1_strength, c1s
12592 @item c2_strength, c2s
12593 @item c3_strength, c3s
12594 Set noise strength for specific pixel component or all pixel components in case
12595 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
12597 @item all_flags, allf
12598 @item c0_flags, c0f
12599 @item c1_flags, c1f
12600 @item c2_flags, c2f
12601 @item c3_flags, c3f
12602 Set pixel component flags or set flags for all components if @var{all_flags}.
12603 Available values for component flags are:
12606 averaged temporal noise (smoother)
12608 mix random noise with a (semi)regular pattern
12610 temporal noise (noise pattern changes between frames)
12612 uniform noise (gaussian otherwise)
12616 @subsection Examples
12618 Add temporal and uniform noise to input video:
12620 noise=alls=20:allf=t+u
12625 Normalize RGB video (aka histogram stretching, contrast stretching).
12626 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
12628 For each channel of each frame, the filter computes the input range and maps
12629 it linearly to the user-specified output range. The output range defaults
12630 to the full dynamic range from pure black to pure white.
12632 Temporal smoothing can be used on the input range to reduce flickering (rapid
12633 changes in brightness) caused when small dark or bright objects enter or leave
12634 the scene. This is similar to the auto-exposure (automatic gain control) on a
12635 video camera, and, like a video camera, it may cause a period of over- or
12636 under-exposure of the video.
12638 The R,G,B channels can be normalized independently, which may cause some
12639 color shifting, or linked together as a single channel, which prevents
12640 color shifting. Linked normalization preserves hue. Independent normalization
12641 does not, so it can be used to remove some color casts. Independent and linked
12642 normalization can be combined in any ratio.
12644 The normalize filter accepts the following options:
12649 Colors which define the output range. The minimum input value is mapped to
12650 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
12651 The defaults are black and white respectively. Specifying white for
12652 @var{blackpt} and black for @var{whitept} will give color-inverted,
12653 normalized video. Shades of grey can be used to reduce the dynamic range
12654 (contrast). Specifying saturated colors here can create some interesting
12658 The number of previous frames to use for temporal smoothing. The input range
12659 of each channel is smoothed using a rolling average over the current frame
12660 and the @var{smoothing} previous frames. The default is 0 (no temporal
12664 Controls the ratio of independent (color shifting) channel normalization to
12665 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
12666 independent. Defaults to 1.0 (fully independent).
12669 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
12670 expensive no-op. Defaults to 1.0 (full strength).
12674 @subsection Examples
12676 Stretch video contrast to use the full dynamic range, with no temporal
12677 smoothing; may flicker depending on the source content:
12679 normalize=blackpt=black:whitept=white:smoothing=0
12682 As above, but with 50 frames of temporal smoothing; flicker should be
12683 reduced, depending on the source content:
12685 normalize=blackpt=black:whitept=white:smoothing=50
12688 As above, but with hue-preserving linked channel normalization:
12690 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
12693 As above, but with half strength:
12695 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
12698 Map the darkest input color to red, the brightest input color to cyan:
12700 normalize=blackpt=red:whitept=cyan
12705 Pass the video source unchanged to the output.
12708 Optical Character Recognition
12710 This filter uses Tesseract for optical character recognition. To enable
12711 compilation of this filter, you need to configure FFmpeg with
12712 @code{--enable-libtesseract}.
12714 It accepts the following options:
12718 Set datapath to tesseract data. Default is to use whatever was
12719 set at installation.
12722 Set language, default is "eng".
12725 Set character whitelist.
12728 Set character blacklist.
12731 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
12735 Apply a video transform using libopencv.
12737 To enable this filter, install the libopencv library and headers and
12738 configure FFmpeg with @code{--enable-libopencv}.
12740 It accepts the following parameters:
12745 The name of the libopencv filter to apply.
12747 @item filter_params
12748 The parameters to pass to the libopencv filter. If not specified, the default
12749 values are assumed.
12753 Refer to the official libopencv documentation for more precise
12755 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
12757 Several libopencv filters are supported; see the following subsections.
12762 Dilate an image by using a specific structuring element.
12763 It corresponds to the libopencv function @code{cvDilate}.
12765 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
12767 @var{struct_el} represents a structuring element, and has the syntax:
12768 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
12770 @var{cols} and @var{rows} represent the number of columns and rows of
12771 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
12772 point, and @var{shape} the shape for the structuring element. @var{shape}
12773 must be "rect", "cross", "ellipse", or "custom".
12775 If the value for @var{shape} is "custom", it must be followed by a
12776 string of the form "=@var{filename}". The file with name
12777 @var{filename} is assumed to represent a binary image, with each
12778 printable character corresponding to a bright pixel. When a custom
12779 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
12780 or columns and rows of the read file are assumed instead.
12782 The default value for @var{struct_el} is "3x3+0x0/rect".
12784 @var{nb_iterations} specifies the number of times the transform is
12785 applied to the image, and defaults to 1.
12789 # Use the default values
12792 # Dilate using a structuring element with a 5x5 cross, iterating two times
12793 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
12795 # Read the shape from the file diamond.shape, iterating two times.
12796 # The file diamond.shape may contain a pattern of characters like this
12802 # The specified columns and rows are ignored
12803 # but the anchor point coordinates are not
12804 ocv=dilate:0x0+2x2/custom=diamond.shape|2
12809 Erode an image by using a specific structuring element.
12810 It corresponds to the libopencv function @code{cvErode}.
12812 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
12813 with the same syntax and semantics as the @ref{dilate} filter.
12817 Smooth the input video.
12819 The filter takes the following parameters:
12820 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
12822 @var{type} is the type of smooth filter to apply, and must be one of
12823 the following values: "blur", "blur_no_scale", "median", "gaussian",
12824 or "bilateral". The default value is "gaussian".
12826 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
12827 depend on the smooth type. @var{param1} and
12828 @var{param2} accept integer positive values or 0. @var{param3} and
12829 @var{param4} accept floating point values.
12831 The default value for @var{param1} is 3. The default value for the
12832 other parameters is 0.
12834 These parameters correspond to the parameters assigned to the
12835 libopencv function @code{cvSmooth}.
12837 @section oscilloscope
12839 2D Video Oscilloscope.
12841 Useful to measure spatial impulse, step responses, chroma delays, etc.
12843 It accepts the following parameters:
12847 Set scope center x position.
12850 Set scope center y position.
12853 Set scope size, relative to frame diagonal.
12856 Set scope tilt/rotation.
12862 Set trace center x position.
12865 Set trace center y position.
12868 Set trace width, relative to width of frame.
12871 Set trace height, relative to height of frame.
12874 Set which components to trace. By default it traces first three components.
12877 Draw trace grid. By default is enabled.
12880 Draw some statistics. By default is enabled.
12883 Draw scope. By default is enabled.
12886 @subsection Examples
12890 Inspect full first row of video frame.
12892 oscilloscope=x=0.5:y=0:s=1
12896 Inspect full last row of video frame.
12898 oscilloscope=x=0.5:y=1:s=1
12902 Inspect full 5th line of video frame of height 1080.
12904 oscilloscope=x=0.5:y=5/1080:s=1
12908 Inspect full last column of video frame.
12910 oscilloscope=x=1:y=0.5:s=1:t=1
12918 Overlay one video on top of another.
12920 It takes two inputs and has one output. The first input is the "main"
12921 video on which the second input is overlaid.
12923 It accepts the following parameters:
12925 A description of the accepted options follows.
12930 Set the expression for the x and y coordinates of the overlaid video
12931 on the main video. Default value is "0" for both expressions. In case
12932 the expression is invalid, it is set to a huge value (meaning that the
12933 overlay will not be displayed within the output visible area).
12936 See @ref{framesync}.
12939 Set when the expressions for @option{x}, and @option{y} are evaluated.
12941 It accepts the following values:
12944 only evaluate expressions once during the filter initialization or
12945 when a command is processed
12948 evaluate expressions for each incoming frame
12951 Default value is @samp{frame}.
12954 See @ref{framesync}.
12957 Set the format for the output video.
12959 It accepts the following values:
12962 force YUV420 output
12965 force YUV422 output
12968 force YUV444 output
12971 force packed RGB output
12974 force planar RGB output
12977 automatically pick format
12980 Default value is @samp{yuv420}.
12983 See @ref{framesync}.
12986 Set format of alpha of the overlaid video, it can be @var{straight} or
12987 @var{premultiplied}. Default is @var{straight}.
12990 The @option{x}, and @option{y} expressions can contain the following
12996 The main input width and height.
13000 The overlay input width and height.
13004 The computed values for @var{x} and @var{y}. They are evaluated for
13009 horizontal and vertical chroma subsample values of the output
13010 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
13014 the number of input frame, starting from 0
13017 the position in the file of the input frame, NAN if unknown
13020 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
13024 This filter also supports the @ref{framesync} options.
13026 Note that the @var{n}, @var{pos}, @var{t} variables are available only
13027 when evaluation is done @emph{per frame}, and will evaluate to NAN
13028 when @option{eval} is set to @samp{init}.
13030 Be aware that frames are taken from each input video in timestamp
13031 order, hence, if their initial timestamps differ, it is a good idea
13032 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
13033 have them begin in the same zero timestamp, as the example for
13034 the @var{movie} filter does.
13036 You can chain together more overlays but you should test the
13037 efficiency of such approach.
13039 @subsection Commands
13041 This filter supports the following commands:
13045 Modify the x and y of the overlay input.
13046 The command accepts the same syntax of the corresponding option.
13048 If the specified expression is not valid, it is kept at its current
13052 @subsection Examples
13056 Draw the overlay at 10 pixels from the bottom right corner of the main
13059 overlay=main_w-overlay_w-10:main_h-overlay_h-10
13062 Using named options the example above becomes:
13064 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
13068 Insert a transparent PNG logo in the bottom left corner of the input,
13069 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
13071 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
13075 Insert 2 different transparent PNG logos (second logo on bottom
13076 right corner) using the @command{ffmpeg} tool:
13078 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
13082 Add a transparent color layer on top of the main video; @code{WxH}
13083 must specify the size of the main input to the overlay filter:
13085 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
13089 Play an original video and a filtered version (here with the deshake
13090 filter) side by side using the @command{ffplay} tool:
13092 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
13095 The above command is the same as:
13097 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
13101 Make a sliding overlay appearing from the left to the right top part of the
13102 screen starting since time 2:
13104 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
13108 Compose output by putting two input videos side to side:
13110 ffmpeg -i left.avi -i right.avi -filter_complex "
13111 nullsrc=size=200x100 [background];
13112 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
13113 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
13114 [background][left] overlay=shortest=1 [background+left];
13115 [background+left][right] overlay=shortest=1:x=100 [left+right]
13120 Mask 10-20 seconds of a video by applying the delogo filter to a section
13122 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
13123 -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]'
13128 Chain several overlays in cascade:
13130 nullsrc=s=200x200 [bg];
13131 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
13132 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
13133 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
13134 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
13135 [in3] null, [mid2] overlay=100:100 [out0]
13142 Apply Overcomplete Wavelet denoiser.
13144 The filter accepts the following options:
13150 Larger depth values will denoise lower frequency components more, but
13151 slow down filtering.
13153 Must be an int in the range 8-16, default is @code{8}.
13155 @item luma_strength, ls
13158 Must be a double value in the range 0-1000, default is @code{1.0}.
13160 @item chroma_strength, cs
13161 Set chroma strength.
13163 Must be a double value in the range 0-1000, default is @code{1.0}.
13169 Add paddings to the input image, and place the original input at the
13170 provided @var{x}, @var{y} coordinates.
13172 It accepts the following parameters:
13177 Specify an expression for the size of the output image with the
13178 paddings added. If the value for @var{width} or @var{height} is 0, the
13179 corresponding input size is used for the output.
13181 The @var{width} expression can reference the value set by the
13182 @var{height} expression, and vice versa.
13184 The default value of @var{width} and @var{height} is 0.
13188 Specify the offsets to place the input image at within the padded area,
13189 with respect to the top/left border of the output image.
13191 The @var{x} expression can reference the value set by the @var{y}
13192 expression, and vice versa.
13194 The default value of @var{x} and @var{y} is 0.
13196 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
13197 so the input image is centered on the padded area.
13200 Specify the color of the padded area. For the syntax of this option,
13201 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
13202 manual,ffmpeg-utils}.
13204 The default value of @var{color} is "black".
13207 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
13209 It accepts the following values:
13213 Only evaluate expressions once during the filter initialization or when
13214 a command is processed.
13217 Evaluate expressions for each incoming frame.
13221 Default value is @samp{init}.
13224 Pad to aspect instead to a resolution.
13228 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
13229 options are expressions containing the following constants:
13234 The input video width and height.
13238 These are the same as @var{in_w} and @var{in_h}.
13242 The output width and height (the size of the padded area), as
13243 specified by the @var{width} and @var{height} expressions.
13247 These are the same as @var{out_w} and @var{out_h}.
13251 The x and y offsets as specified by the @var{x} and @var{y}
13252 expressions, or NAN if not yet specified.
13255 same as @var{iw} / @var{ih}
13258 input sample aspect ratio
13261 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
13265 The horizontal and vertical chroma subsample values. For example for the
13266 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
13269 @subsection Examples
13273 Add paddings with the color "violet" to the input video. The output video
13274 size is 640x480, and the top-left corner of the input video is placed at
13277 pad=640:480:0:40:violet
13280 The example above is equivalent to the following command:
13282 pad=width=640:height=480:x=0:y=40:color=violet
13286 Pad the input to get an output with dimensions increased by 3/2,
13287 and put the input video at the center of the padded area:
13289 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
13293 Pad the input to get a squared output with size equal to the maximum
13294 value between the input width and height, and put the input video at
13295 the center of the padded area:
13297 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
13301 Pad the input to get a final w/h ratio of 16:9:
13303 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
13307 In case of anamorphic video, in order to set the output display aspect
13308 correctly, it is necessary to use @var{sar} in the expression,
13309 according to the relation:
13311 (ih * X / ih) * sar = output_dar
13312 X = output_dar / sar
13315 Thus the previous example needs to be modified to:
13317 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
13321 Double the output size and put the input video in the bottom-right
13322 corner of the output padded area:
13324 pad="2*iw:2*ih:ow-iw:oh-ih"
13328 @anchor{palettegen}
13329 @section palettegen
13331 Generate one palette for a whole video stream.
13333 It accepts the following options:
13337 Set the maximum number of colors to quantize in the palette.
13338 Note: the palette will still contain 256 colors; the unused palette entries
13341 @item reserve_transparent
13342 Create a palette of 255 colors maximum and reserve the last one for
13343 transparency. Reserving the transparency color is useful for GIF optimization.
13344 If not set, the maximum of colors in the palette will be 256. You probably want
13345 to disable this option for a standalone image.
13348 @item transparency_color
13349 Set the color that will be used as background for transparency.
13352 Set statistics mode.
13354 It accepts the following values:
13357 Compute full frame histograms.
13359 Compute histograms only for the part that differs from previous frame. This
13360 might be relevant to give more importance to the moving part of your input if
13361 the background is static.
13363 Compute new histogram for each frame.
13366 Default value is @var{full}.
13369 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
13370 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
13371 color quantization of the palette. This information is also visible at
13372 @var{info} logging level.
13374 @subsection Examples
13378 Generate a representative palette of a given video using @command{ffmpeg}:
13380 ffmpeg -i input.mkv -vf palettegen palette.png
13384 @section paletteuse
13386 Use a palette to downsample an input video stream.
13388 The filter takes two inputs: one video stream and a palette. The palette must
13389 be a 256 pixels image.
13391 It accepts the following options:
13395 Select dithering mode. Available algorithms are:
13398 Ordered 8x8 bayer dithering (deterministic)
13400 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
13401 Note: this dithering is sometimes considered "wrong" and is included as a
13403 @item floyd_steinberg
13404 Floyd and Steingberg dithering (error diffusion)
13406 Frankie Sierra dithering v2 (error diffusion)
13408 Frankie Sierra dithering v2 "Lite" (error diffusion)
13411 Default is @var{sierra2_4a}.
13414 When @var{bayer} dithering is selected, this option defines the scale of the
13415 pattern (how much the crosshatch pattern is visible). A low value means more
13416 visible pattern for less banding, and higher value means less visible pattern
13417 at the cost of more banding.
13419 The option must be an integer value in the range [0,5]. Default is @var{2}.
13422 If set, define the zone to process
13426 Only the changing rectangle will be reprocessed. This is similar to GIF
13427 cropping/offsetting compression mechanism. This option can be useful for speed
13428 if only a part of the image is changing, and has use cases such as limiting the
13429 scope of the error diffusal @option{dither} to the rectangle that bounds the
13430 moving scene (it leads to more deterministic output if the scene doesn't change
13431 much, and as a result less moving noise and better GIF compression).
13434 Default is @var{none}.
13437 Take new palette for each output frame.
13439 @item alpha_threshold
13440 Sets the alpha threshold for transparency. Alpha values above this threshold
13441 will be treated as completely opaque, and values below this threshold will be
13442 treated as completely transparent.
13444 The option must be an integer value in the range [0,255]. Default is @var{128}.
13447 @subsection Examples
13451 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
13452 using @command{ffmpeg}:
13454 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
13458 @section perspective
13460 Correct perspective of video not recorded perpendicular to the screen.
13462 A description of the accepted parameters follows.
13473 Set coordinates expression for top left, top right, bottom left and bottom right corners.
13474 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
13475 If the @code{sense} option is set to @code{source}, then the specified points will be sent
13476 to the corners of the destination. If the @code{sense} option is set to @code{destination},
13477 then the corners of the source will be sent to the specified coordinates.
13479 The expressions can use the following variables:
13484 the width and height of video frame.
13488 Output frame count.
13491 @item interpolation
13492 Set interpolation for perspective correction.
13494 It accepts the following values:
13500 Default value is @samp{linear}.
13503 Set interpretation of coordinate options.
13505 It accepts the following values:
13509 Send point in the source specified by the given coordinates to
13510 the corners of the destination.
13512 @item 1, destination
13514 Send the corners of the source to the point in the destination specified
13515 by the given coordinates.
13517 Default value is @samp{source}.
13521 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
13523 It accepts the following values:
13526 only evaluate expressions once during the filter initialization or
13527 when a command is processed
13530 evaluate expressions for each incoming frame
13533 Default value is @samp{init}.
13538 Delay interlaced video by one field time so that the field order changes.
13540 The intended use is to fix PAL movies that have been captured with the
13541 opposite field order to the film-to-video transfer.
13543 A description of the accepted parameters follows.
13549 It accepts the following values:
13552 Capture field order top-first, transfer bottom-first.
13553 Filter will delay the bottom field.
13556 Capture field order bottom-first, transfer top-first.
13557 Filter will delay the top field.
13560 Capture and transfer with the same field order. This mode only exists
13561 for the documentation of the other options to refer to, but if you
13562 actually select it, the filter will faithfully do nothing.
13565 Capture field order determined automatically by field flags, transfer
13567 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
13568 basis using field flags. If no field information is available,
13569 then this works just like @samp{u}.
13572 Capture unknown or varying, transfer opposite.
13573 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
13574 analyzing the images and selecting the alternative that produces best
13575 match between the fields.
13578 Capture top-first, transfer unknown or varying.
13579 Filter selects among @samp{t} and @samp{p} using image analysis.
13582 Capture bottom-first, transfer unknown or varying.
13583 Filter selects among @samp{b} and @samp{p} using image analysis.
13586 Capture determined by field flags, transfer unknown or varying.
13587 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
13588 image analysis. If no field information is available, then this works just
13589 like @samp{U}. This is the default mode.
13592 Both capture and transfer unknown or varying.
13593 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
13597 @section pixdesctest
13599 Pixel format descriptor test filter, mainly useful for internal
13600 testing. The output video should be equal to the input video.
13604 format=monow, pixdesctest
13607 can be used to test the monowhite pixel format descriptor definition.
13611 Display sample values of color channels. Mainly useful for checking color
13612 and levels. Minimum supported resolution is 640x480.
13614 The filters accept the following options:
13618 Set scope X position, relative offset on X axis.
13621 Set scope Y position, relative offset on Y axis.
13630 Set window opacity. This window also holds statistics about pixel area.
13633 Set window X position, relative offset on X axis.
13636 Set window Y position, relative offset on Y axis.
13641 Enable the specified chain of postprocessing subfilters using libpostproc. This
13642 library should be automatically selected with a GPL build (@code{--enable-gpl}).
13643 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
13644 Each subfilter and some options have a short and a long name that can be used
13645 interchangeably, i.e. dr/dering are the same.
13647 The filters accept the following options:
13651 Set postprocessing subfilters string.
13654 All subfilters share common options to determine their scope:
13658 Honor the quality commands for this subfilter.
13661 Do chrominance filtering, too (default).
13664 Do luminance filtering only (no chrominance).
13667 Do chrominance filtering only (no luminance).
13670 These options can be appended after the subfilter name, separated by a '|'.
13672 Available subfilters are:
13675 @item hb/hdeblock[|difference[|flatness]]
13676 Horizontal deblocking filter
13679 Difference factor where higher values mean more deblocking (default: @code{32}).
13681 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13684 @item vb/vdeblock[|difference[|flatness]]
13685 Vertical deblocking filter
13688 Difference factor where higher values mean more deblocking (default: @code{32}).
13690 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13693 @item ha/hadeblock[|difference[|flatness]]
13694 Accurate horizontal deblocking filter
13697 Difference factor where higher values mean more deblocking (default: @code{32}).
13699 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13702 @item va/vadeblock[|difference[|flatness]]
13703 Accurate vertical deblocking filter
13706 Difference factor where higher values mean more deblocking (default: @code{32}).
13708 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13712 The horizontal and vertical deblocking filters share the difference and
13713 flatness values so you cannot set different horizontal and vertical
13717 @item h1/x1hdeblock
13718 Experimental horizontal deblocking filter
13720 @item v1/x1vdeblock
13721 Experimental vertical deblocking filter
13726 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
13729 larger -> stronger filtering
13731 larger -> stronger filtering
13733 larger -> stronger filtering
13736 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
13739 Stretch luminance to @code{0-255}.
13742 @item lb/linblenddeint
13743 Linear blend deinterlacing filter that deinterlaces the given block by
13744 filtering all lines with a @code{(1 2 1)} filter.
13746 @item li/linipoldeint
13747 Linear interpolating deinterlacing filter that deinterlaces the given block by
13748 linearly interpolating every second line.
13750 @item ci/cubicipoldeint
13751 Cubic interpolating deinterlacing filter deinterlaces the given block by
13752 cubically interpolating every second line.
13754 @item md/mediandeint
13755 Median deinterlacing filter that deinterlaces the given block by applying a
13756 median filter to every second line.
13758 @item fd/ffmpegdeint
13759 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
13760 second line with a @code{(-1 4 2 4 -1)} filter.
13763 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
13764 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
13766 @item fq/forceQuant[|quantizer]
13767 Overrides the quantizer table from the input with the constant quantizer you
13775 Default pp filter combination (@code{hb|a,vb|a,dr|a})
13778 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
13781 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
13784 @subsection Examples
13788 Apply horizontal and vertical deblocking, deringing and automatic
13789 brightness/contrast:
13795 Apply default filters without brightness/contrast correction:
13801 Apply default filters and temporal denoiser:
13803 pp=default/tmpnoise|1|2|3
13807 Apply deblocking on luminance only, and switch vertical deblocking on or off
13808 automatically depending on available CPU time:
13815 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
13816 similar to spp = 6 with 7 point DCT, where only the center sample is
13819 The filter accepts the following options:
13823 Force a constant quantization parameter. It accepts an integer in range
13824 0 to 63. If not set, the filter will use the QP from the video stream
13828 Set thresholding mode. Available modes are:
13832 Set hard thresholding.
13834 Set soft thresholding (better de-ringing effect, but likely blurrier).
13836 Set medium thresholding (good results, default).
13840 @section premultiply
13841 Apply alpha premultiply effect to input video stream using first plane
13842 of second stream as alpha.
13844 Both streams must have same dimensions and same pixel format.
13846 The filter accepts the following option:
13850 Set which planes will be processed, unprocessed planes will be copied.
13851 By default value 0xf, all planes will be processed.
13854 Do not require 2nd input for processing, instead use alpha plane from input stream.
13858 Apply prewitt operator to input video stream.
13860 The filter accepts the following option:
13864 Set which planes will be processed, unprocessed planes will be copied.
13865 By default value 0xf, all planes will be processed.
13868 Set value which will be multiplied with filtered result.
13871 Set value which will be added to filtered result.
13874 @anchor{program_opencl}
13875 @section program_opencl
13877 Filter video using an OpenCL program.
13882 OpenCL program source file.
13885 Kernel name in program.
13888 Number of inputs to the filter. Defaults to 1.
13891 Size of output frames. Defaults to the same as the first input.
13895 The program source file must contain a kernel function with the given name,
13896 which will be run once for each plane of the output. Each run on a plane
13897 gets enqueued as a separate 2D global NDRange with one work-item for each
13898 pixel to be generated. The global ID offset for each work-item is therefore
13899 the coordinates of a pixel in the destination image.
13901 The kernel function needs to take the following arguments:
13904 Destination image, @var{__write_only image2d_t}.
13906 This image will become the output; the kernel should write all of it.
13908 Frame index, @var{unsigned int}.
13910 This is a counter starting from zero and increasing by one for each frame.
13912 Source images, @var{__read_only image2d_t}.
13914 These are the most recent images on each input. The kernel may read from
13915 them to generate the output, but they can't be written to.
13922 Copy the input to the output (output must be the same size as the input).
13924 __kernel void copy(__write_only image2d_t destination,
13925 unsigned int index,
13926 __read_only image2d_t source)
13928 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
13930 int2 location = (int2)(get_global_id(0), get_global_id(1));
13932 float4 value = read_imagef(source, sampler, location);
13934 write_imagef(destination, location, value);
13939 Apply a simple transformation, rotating the input by an amount increasing
13940 with the index counter. Pixel values are linearly interpolated by the
13941 sampler, and the output need not have the same dimensions as the input.
13943 __kernel void rotate_image(__write_only image2d_t dst,
13944 unsigned int index,
13945 __read_only image2d_t src)
13947 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
13948 CLK_FILTER_LINEAR);
13950 float angle = (float)index / 100.0f;
13952 float2 dst_dim = convert_float2(get_image_dim(dst));
13953 float2 src_dim = convert_float2(get_image_dim(src));
13955 float2 dst_cen = dst_dim / 2.0f;
13956 float2 src_cen = src_dim / 2.0f;
13958 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
13960 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
13962 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
13963 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
13965 src_pos = src_pos * src_dim / dst_dim;
13967 float2 src_loc = src_pos + src_cen;
13969 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
13970 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
13971 write_imagef(dst, dst_loc, 0.5f);
13973 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
13978 Blend two inputs together, with the amount of each input used varying
13979 with the index counter.
13981 __kernel void blend_images(__write_only image2d_t dst,
13982 unsigned int index,
13983 __read_only image2d_t src1,
13984 __read_only image2d_t src2)
13986 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
13987 CLK_FILTER_LINEAR);
13989 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
13991 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
13992 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
13993 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
13995 float4 val1 = read_imagef(src1, sampler, src1_loc);
13996 float4 val2 = read_imagef(src2, sampler, src2_loc);
13998 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
14004 @section pseudocolor
14006 Alter frame colors in video with pseudocolors.
14008 This filter accept the following options:
14012 set pixel first component expression
14015 set pixel second component expression
14018 set pixel third component expression
14021 set pixel fourth component expression, corresponds to the alpha component
14024 set component to use as base for altering colors
14027 Each of them specifies the expression to use for computing the lookup table for
14028 the corresponding pixel component values.
14030 The expressions can contain the following constants and functions:
14035 The input width and height.
14038 The input value for the pixel component.
14040 @item ymin, umin, vmin, amin
14041 The minimum allowed component value.
14043 @item ymax, umax, vmax, amax
14044 The maximum allowed component value.
14047 All expressions default to "val".
14049 @subsection Examples
14053 Change too high luma values to gradient:
14055 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'"
14061 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
14062 Ratio) between two input videos.
14064 This filter takes in input two input videos, the first input is
14065 considered the "main" source and is passed unchanged to the
14066 output. The second input is used as a "reference" video for computing
14069 Both video inputs must have the same resolution and pixel format for
14070 this filter to work correctly. Also it assumes that both inputs
14071 have the same number of frames, which are compared one by one.
14073 The obtained average PSNR is printed through the logging system.
14075 The filter stores the accumulated MSE (mean squared error) of each
14076 frame, and at the end of the processing it is averaged across all frames
14077 equally, and the following formula is applied to obtain the PSNR:
14080 PSNR = 10*log10(MAX^2/MSE)
14083 Where MAX is the average of the maximum values of each component of the
14086 The description of the accepted parameters follows.
14089 @item stats_file, f
14090 If specified the filter will use the named file to save the PSNR of
14091 each individual frame. When filename equals "-" the data is sent to
14094 @item stats_version
14095 Specifies which version of the stats file format to use. Details of
14096 each format are written below.
14097 Default value is 1.
14099 @item stats_add_max
14100 Determines whether the max value is output to the stats log.
14101 Default value is 0.
14102 Requires stats_version >= 2. If this is set and stats_version < 2,
14103 the filter will return an error.
14106 This filter also supports the @ref{framesync} options.
14108 The file printed if @var{stats_file} is selected, contains a sequence of
14109 key/value pairs of the form @var{key}:@var{value} for each compared
14112 If a @var{stats_version} greater than 1 is specified, a header line precedes
14113 the list of per-frame-pair stats, with key value pairs following the frame
14114 format with the following parameters:
14117 @item psnr_log_version
14118 The version of the log file format. Will match @var{stats_version}.
14121 A comma separated list of the per-frame-pair parameters included in
14125 A description of each shown per-frame-pair parameter follows:
14129 sequential number of the input frame, starting from 1
14132 Mean Square Error pixel-by-pixel average difference of the compared
14133 frames, averaged over all the image components.
14135 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
14136 Mean Square Error pixel-by-pixel average difference of the compared
14137 frames for the component specified by the suffix.
14139 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
14140 Peak Signal to Noise ratio of the compared frames for the component
14141 specified by the suffix.
14143 @item max_avg, max_y, max_u, max_v
14144 Maximum allowed value for each channel, and average over all
14150 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
14151 [main][ref] psnr="stats_file=stats.log" [out]
14154 On this example the input file being processed is compared with the
14155 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
14156 is stored in @file{stats.log}.
14161 Pulldown reversal (inverse telecine) filter, capable of handling mixed
14162 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
14165 The pullup filter is designed to take advantage of future context in making
14166 its decisions. This filter is stateless in the sense that it does not lock
14167 onto a pattern to follow, but it instead looks forward to the following
14168 fields in order to identify matches and rebuild progressive frames.
14170 To produce content with an even framerate, insert the fps filter after
14171 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
14172 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
14174 The filter accepts the following options:
14181 These options set the amount of "junk" to ignore at the left, right, top, and
14182 bottom of the image, respectively. Left and right are in units of 8 pixels,
14183 while top and bottom are in units of 2 lines.
14184 The default is 8 pixels on each side.
14187 Set the strict breaks. Setting this option to 1 will reduce the chances of
14188 filter generating an occasional mismatched frame, but it may also cause an
14189 excessive number of frames to be dropped during high motion sequences.
14190 Conversely, setting it to -1 will make filter match fields more easily.
14191 This may help processing of video where there is slight blurring between
14192 the fields, but may also cause there to be interlaced frames in the output.
14193 Default value is @code{0}.
14196 Set the metric plane to use. It accepts the following values:
14202 Use chroma blue plane.
14205 Use chroma red plane.
14208 This option may be set to use chroma plane instead of the default luma plane
14209 for doing filter's computations. This may improve accuracy on very clean
14210 source material, but more likely will decrease accuracy, especially if there
14211 is chroma noise (rainbow effect) or any grayscale video.
14212 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
14213 load and make pullup usable in realtime on slow machines.
14216 For best results (without duplicated frames in the output file) it is
14217 necessary to change the output frame rate. For example, to inverse
14218 telecine NTSC input:
14220 ffmpeg -i input -vf pullup -r 24000/1001 ...
14225 Change video quantization parameters (QP).
14227 The filter accepts the following option:
14231 Set expression for quantization parameter.
14234 The expression is evaluated through the eval API and can contain, among others,
14235 the following constants:
14239 1 if index is not 129, 0 otherwise.
14242 Sequential index starting from -129 to 128.
14245 @subsection Examples
14249 Some equation like:
14257 Flush video frames from internal cache of frames into a random order.
14258 No frame is discarded.
14259 Inspired by @ref{frei0r} nervous filter.
14263 Set size in number of frames of internal cache, in range from @code{2} to
14264 @code{512}. Default is @code{30}.
14267 Set seed for random number generator, must be an integer included between
14268 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
14269 less than @code{0}, the filter will try to use a good random seed on a
14273 @section readeia608
14275 Read closed captioning (EIA-608) information from the top lines of a video frame.
14277 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
14278 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
14279 with EIA-608 data (starting from 0). A description of each metadata value follows:
14282 @item lavfi.readeia608.X.cc
14283 The two bytes stored as EIA-608 data (printed in hexadecimal).
14285 @item lavfi.readeia608.X.line
14286 The number of the line on which the EIA-608 data was identified and read.
14289 This filter accepts the following options:
14293 Set the line to start scanning for EIA-608 data. Default is @code{0}.
14296 Set the line to end scanning for EIA-608 data. Default is @code{29}.
14299 Set minimal acceptable amplitude change for sync codes detection.
14300 Default is @code{0.2}. Allowed range is @code{[0.001 - 1]}.
14303 Set the ratio of width reserved for sync code detection.
14304 Default is @code{0.27}. Allowed range is @code{[0.01 - 0.7]}.
14307 Set the max peaks height difference for sync code detection.
14308 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
14311 Set max peaks period difference for sync code detection.
14312 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
14315 Set the first two max start code bits differences.
14316 Default is @code{0.02}. Allowed range is @code{[0.0 - 0.5]}.
14319 Set the minimum ratio of bits height compared to 3rd start code bit.
14320 Default is @code{0.75}. Allowed range is @code{[0.01 - 1]}.
14323 Set the white color threshold. Default is @code{0.35}. Allowed range is @code{[0.1 - 1]}.
14326 Set the black color threshold. Default is @code{0.15}. Allowed range is @code{[0.0 - 0.5]}.
14329 Enable checking the parity bit. In the event of a parity error, the filter will output
14330 @code{0x00} for that character. Default is false.
14333 @subsection Examples
14337 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
14339 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
14345 Read vertical interval timecode (VITC) information from the top lines of a
14348 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
14349 timecode value, if a valid timecode has been detected. Further metadata key
14350 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
14351 timecode data has been found or not.
14353 This filter accepts the following options:
14357 Set the maximum number of lines to scan for VITC data. If the value is set to
14358 @code{-1} the full video frame is scanned. Default is @code{45}.
14361 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
14362 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
14365 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
14366 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
14369 @subsection Examples
14373 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
14374 draw @code{--:--:--:--} as a placeholder:
14376 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
14382 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
14384 Destination pixel at position (X, Y) will be picked from source (x, y) position
14385 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
14386 value for pixel will be used for destination pixel.
14388 Xmap and Ymap input video streams must be of same dimensions. Output video stream
14389 will have Xmap/Ymap video stream dimensions.
14390 Xmap and Ymap input video streams are 16bit depth, single channel.
14392 @section removegrain
14394 The removegrain filter is a spatial denoiser for progressive video.
14398 Set mode for the first plane.
14401 Set mode for the second plane.
14404 Set mode for the third plane.
14407 Set mode for the fourth plane.
14410 Range of mode is from 0 to 24. Description of each mode follows:
14414 Leave input plane unchanged. Default.
14417 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
14420 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
14423 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
14426 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
14427 This is equivalent to a median filter.
14430 Line-sensitive clipping giving the minimal change.
14433 Line-sensitive clipping, intermediate.
14436 Line-sensitive clipping, intermediate.
14439 Line-sensitive clipping, intermediate.
14442 Line-sensitive clipping on a line where the neighbours pixels are the closest.
14445 Replaces the target pixel with the closest neighbour.
14448 [1 2 1] horizontal and vertical kernel blur.
14454 Bob mode, interpolates top field from the line where the neighbours
14455 pixels are the closest.
14458 Bob mode, interpolates bottom field from the line where the neighbours
14459 pixels are the closest.
14462 Bob mode, interpolates top field. Same as 13 but with a more complicated
14463 interpolation formula.
14466 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
14467 interpolation formula.
14470 Clips the pixel with the minimum and maximum of respectively the maximum and
14471 minimum of each pair of opposite neighbour pixels.
14474 Line-sensitive clipping using opposite neighbours whose greatest distance from
14475 the current pixel is minimal.
14478 Replaces the pixel with the average of its 8 neighbours.
14481 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
14484 Clips pixels using the averages of opposite neighbour.
14487 Same as mode 21 but simpler and faster.
14490 Small edge and halo removal, but reputed useless.
14496 @section removelogo
14498 Suppress a TV station logo, using an image file to determine which
14499 pixels comprise the logo. It works by filling in the pixels that
14500 comprise the logo with neighboring pixels.
14502 The filter accepts the following options:
14506 Set the filter bitmap file, which can be any image format supported by
14507 libavformat. The width and height of the image file must match those of the
14508 video stream being processed.
14511 Pixels in the provided bitmap image with a value of zero are not
14512 considered part of the logo, non-zero pixels are considered part of
14513 the logo. If you use white (255) for the logo and black (0) for the
14514 rest, you will be safe. For making the filter bitmap, it is
14515 recommended to take a screen capture of a black frame with the logo
14516 visible, and then using a threshold filter followed by the erode
14517 filter once or twice.
14519 If needed, little splotches can be fixed manually. Remember that if
14520 logo pixels are not covered, the filter quality will be much
14521 reduced. Marking too many pixels as part of the logo does not hurt as
14522 much, but it will increase the amount of blurring needed to cover over
14523 the image and will destroy more information than necessary, and extra
14524 pixels will slow things down on a large logo.
14526 @section repeatfields
14528 This filter uses the repeat_field flag from the Video ES headers and hard repeats
14529 fields based on its value.
14533 Reverse a video clip.
14535 Warning: This filter requires memory to buffer the entire clip, so trimming
14538 @subsection Examples
14542 Take the first 5 seconds of a clip, and reverse it.
14549 Shift R/G/B/A pixels horizontally and/or vertically.
14551 The filter accepts the following options:
14554 Set amount to shift red horizontally.
14556 Set amount to shift red vertically.
14558 Set amount to shift green horizontally.
14560 Set amount to shift green vertically.
14562 Set amount to shift blue horizontally.
14564 Set amount to shift blue vertically.
14566 Set amount to shift alpha horizontally.
14568 Set amount to shift alpha vertically.
14570 Set edge mode, can be @var{smear}, default, or @var{warp}.
14574 Apply roberts cross operator to input video stream.
14576 The filter accepts the following option:
14580 Set which planes will be processed, unprocessed planes will be copied.
14581 By default value 0xf, all planes will be processed.
14584 Set value which will be multiplied with filtered result.
14587 Set value which will be added to filtered result.
14592 Rotate video by an arbitrary angle expressed in radians.
14594 The filter accepts the following options:
14596 A description of the optional parameters follows.
14599 Set an expression for the angle by which to rotate the input video
14600 clockwise, expressed as a number of radians. A negative value will
14601 result in a counter-clockwise rotation. By default it is set to "0".
14603 This expression is evaluated for each frame.
14606 Set the output width expression, default value is "iw".
14607 This expression is evaluated just once during configuration.
14610 Set the output height expression, default value is "ih".
14611 This expression is evaluated just once during configuration.
14614 Enable bilinear interpolation if set to 1, a value of 0 disables
14615 it. Default value is 1.
14618 Set the color used to fill the output area not covered by the rotated
14619 image. For the general syntax of this option, check the
14620 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
14621 If the special value "none" is selected then no
14622 background is printed (useful for example if the background is never shown).
14624 Default value is "black".
14627 The expressions for the angle and the output size can contain the
14628 following constants and functions:
14632 sequential number of the input frame, starting from 0. It is always NAN
14633 before the first frame is filtered.
14636 time in seconds of the input frame, it is set to 0 when the filter is
14637 configured. It is always NAN before the first frame is filtered.
14641 horizontal and vertical chroma subsample values. For example for the
14642 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14646 the input video width and height
14650 the output width and height, that is the size of the padded area as
14651 specified by the @var{width} and @var{height} expressions
14655 the minimal width/height required for completely containing the input
14656 video rotated by @var{a} radians.
14658 These are only available when computing the @option{out_w} and
14659 @option{out_h} expressions.
14662 @subsection Examples
14666 Rotate the input by PI/6 radians clockwise:
14672 Rotate the input by PI/6 radians counter-clockwise:
14678 Rotate the input by 45 degrees clockwise:
14684 Apply a constant rotation with period T, starting from an angle of PI/3:
14686 rotate=PI/3+2*PI*t/T
14690 Make the input video rotation oscillating with a period of T
14691 seconds and an amplitude of A radians:
14693 rotate=A*sin(2*PI/T*t)
14697 Rotate the video, output size is chosen so that the whole rotating
14698 input video is always completely contained in the output:
14700 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
14704 Rotate the video, reduce the output size so that no background is ever
14707 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
14711 @subsection Commands
14713 The filter supports the following commands:
14717 Set the angle expression.
14718 The command accepts the same syntax of the corresponding option.
14720 If the specified expression is not valid, it is kept at its current
14726 Apply Shape Adaptive Blur.
14728 The filter accepts the following options:
14731 @item luma_radius, lr
14732 Set luma blur filter strength, must be a value in range 0.1-4.0, default
14733 value is 1.0. A greater value will result in a more blurred image, and
14734 in slower processing.
14736 @item luma_pre_filter_radius, lpfr
14737 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
14740 @item luma_strength, ls
14741 Set luma maximum difference between pixels to still be considered, must
14742 be a value in the 0.1-100.0 range, default value is 1.0.
14744 @item chroma_radius, cr
14745 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
14746 greater value will result in a more blurred image, and in slower
14749 @item chroma_pre_filter_radius, cpfr
14750 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
14752 @item chroma_strength, cs
14753 Set chroma maximum difference between pixels to still be considered,
14754 must be a value in the -0.9-100.0 range.
14757 Each chroma option value, if not explicitly specified, is set to the
14758 corresponding luma option value.
14763 Scale (resize) the input video, using the libswscale library.
14765 The scale filter forces the output display aspect ratio to be the same
14766 of the input, by changing the output sample aspect ratio.
14768 If the input image format is different from the format requested by
14769 the next filter, the scale filter will convert the input to the
14772 @subsection Options
14773 The filter accepts the following options, or any of the options
14774 supported by the libswscale scaler.
14776 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
14777 the complete list of scaler options.
14782 Set the output video dimension expression. Default value is the input
14785 If the @var{width} or @var{w} value is 0, the input width is used for
14786 the output. If the @var{height} or @var{h} value is 0, the input height
14787 is used for the output.
14789 If one and only one of the values is -n with n >= 1, the scale filter
14790 will use a value that maintains the aspect ratio of the input image,
14791 calculated from the other specified dimension. After that it will,
14792 however, make sure that the calculated dimension is divisible by n and
14793 adjust the value if necessary.
14795 If both values are -n with n >= 1, the behavior will be identical to
14796 both values being set to 0 as previously detailed.
14798 See below for the list of accepted constants for use in the dimension
14802 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
14806 Only evaluate expressions once during the filter initialization or when a command is processed.
14809 Evaluate expressions for each incoming frame.
14813 Default value is @samp{init}.
14817 Set the interlacing mode. It accepts the following values:
14821 Force interlaced aware scaling.
14824 Do not apply interlaced scaling.
14827 Select interlaced aware scaling depending on whether the source frames
14828 are flagged as interlaced or not.
14831 Default value is @samp{0}.
14834 Set libswscale scaling flags. See
14835 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
14836 complete list of values. If not explicitly specified the filter applies
14840 @item param0, param1
14841 Set libswscale input parameters for scaling algorithms that need them. See
14842 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
14843 complete documentation. If not explicitly specified the filter applies
14849 Set the video size. For the syntax of this option, check the
14850 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
14852 @item in_color_matrix
14853 @item out_color_matrix
14854 Set in/output YCbCr color space type.
14856 This allows the autodetected value to be overridden as well as allows forcing
14857 a specific value used for the output and encoder.
14859 If not specified, the color space type depends on the pixel format.
14865 Choose automatically.
14868 Format conforming to International Telecommunication Union (ITU)
14869 Recommendation BT.709.
14872 Set color space conforming to the United States Federal Communications
14873 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
14876 Set color space conforming to:
14880 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
14883 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
14886 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
14891 Set color space conforming to SMPTE ST 240:1999.
14896 Set in/output YCbCr sample range.
14898 This allows the autodetected value to be overridden as well as allows forcing
14899 a specific value used for the output and encoder. If not specified, the
14900 range depends on the pixel format. Possible values:
14904 Choose automatically.
14907 Set full range (0-255 in case of 8-bit luma).
14909 @item mpeg/limited/tv
14910 Set "MPEG" range (16-235 in case of 8-bit luma).
14913 @item force_original_aspect_ratio
14914 Enable decreasing or increasing output video width or height if necessary to
14915 keep the original aspect ratio. Possible values:
14919 Scale the video as specified and disable this feature.
14922 The output video dimensions will automatically be decreased if needed.
14925 The output video dimensions will automatically be increased if needed.
14929 One useful instance of this option is that when you know a specific device's
14930 maximum allowed resolution, you can use this to limit the output video to
14931 that, while retaining the aspect ratio. For example, device A allows
14932 1280x720 playback, and your video is 1920x800. Using this option (set it to
14933 decrease) and specifying 1280x720 to the command line makes the output
14936 Please note that this is a different thing than specifying -1 for @option{w}
14937 or @option{h}, you still need to specify the output resolution for this option
14942 The values of the @option{w} and @option{h} options are expressions
14943 containing the following constants:
14948 The input width and height
14952 These are the same as @var{in_w} and @var{in_h}.
14956 The output (scaled) width and height
14960 These are the same as @var{out_w} and @var{out_h}
14963 The same as @var{iw} / @var{ih}
14966 input sample aspect ratio
14969 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
14973 horizontal and vertical input chroma subsample values. For example for the
14974 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14978 horizontal and vertical output chroma subsample values. For example for the
14979 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14982 @subsection Examples
14986 Scale the input video to a size of 200x100
14991 This is equivalent to:
15002 Specify a size abbreviation for the output size:
15007 which can also be written as:
15013 Scale the input to 2x:
15015 scale=w=2*iw:h=2*ih
15019 The above is the same as:
15021 scale=2*in_w:2*in_h
15025 Scale the input to 2x with forced interlaced scaling:
15027 scale=2*iw:2*ih:interl=1
15031 Scale the input to half size:
15033 scale=w=iw/2:h=ih/2
15037 Increase the width, and set the height to the same size:
15043 Seek Greek harmony:
15050 Increase the height, and set the width to 3/2 of the height:
15052 scale=w=3/2*oh:h=3/5*ih
15056 Increase the size, making the size a multiple of the chroma
15059 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
15063 Increase the width to a maximum of 500 pixels,
15064 keeping the same aspect ratio as the input:
15066 scale=w='min(500\, iw*3/2):h=-1'
15070 Make pixels square by combining scale and setsar:
15072 scale='trunc(ih*dar):ih',setsar=1/1
15076 Make pixels square by combining scale and setsar,
15077 making sure the resulting resolution is even (required by some codecs):
15079 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
15083 @subsection Commands
15085 This filter supports the following commands:
15089 Set the output video dimension expression.
15090 The command accepts the same syntax of the corresponding option.
15092 If the specified expression is not valid, it is kept at its current
15098 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
15099 format conversion on CUDA video frames. Setting the output width and height
15100 works in the same way as for the @var{scale} filter.
15102 The following additional options are accepted:
15105 The pixel format of the output CUDA frames. If set to the string "same" (the
15106 default), the input format will be kept. Note that automatic format negotiation
15107 and conversion is not yet supported for hardware frames
15110 The interpolation algorithm used for resizing. One of the following:
15117 @item cubic2p_bspline
15118 2-parameter cubic (B=1, C=0)
15120 @item cubic2p_catmullrom
15121 2-parameter cubic (B=0, C=1/2)
15123 @item cubic2p_b05c03
15124 2-parameter cubic (B=1/2, C=3/10)
15136 Scale (resize) the input video, based on a reference video.
15138 See the scale filter for available options, scale2ref supports the same but
15139 uses the reference video instead of the main input as basis. scale2ref also
15140 supports the following additional constants for the @option{w} and
15141 @option{h} options:
15146 The main input video's width and height
15149 The same as @var{main_w} / @var{main_h}
15152 The main input video's sample aspect ratio
15154 @item main_dar, mdar
15155 The main input video's display aspect ratio. Calculated from
15156 @code{(main_w / main_h) * main_sar}.
15160 The main input video's horizontal and vertical chroma subsample values.
15161 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
15165 @subsection Examples
15169 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
15171 'scale2ref[b][a];[a][b]overlay'
15175 @anchor{selectivecolor}
15176 @section selectivecolor
15178 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
15179 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
15180 by the "purity" of the color (that is, how saturated it already is).
15182 This filter is similar to the Adobe Photoshop Selective Color tool.
15184 The filter accepts the following options:
15187 @item correction_method
15188 Select color correction method.
15190 Available values are:
15193 Specified adjustments are applied "as-is" (added/subtracted to original pixel
15196 Specified adjustments are relative to the original component value.
15198 Default is @code{absolute}.
15200 Adjustments for red pixels (pixels where the red component is the maximum)
15202 Adjustments for yellow pixels (pixels where the blue component is the minimum)
15204 Adjustments for green pixels (pixels where the green component is the maximum)
15206 Adjustments for cyan pixels (pixels where the red component is the minimum)
15208 Adjustments for blue pixels (pixels where the blue component is the maximum)
15210 Adjustments for magenta pixels (pixels where the green component is the minimum)
15212 Adjustments for white pixels (pixels where all components are greater than 128)
15214 Adjustments for all pixels except pure black and pure white
15216 Adjustments for black pixels (pixels where all components are lesser than 128)
15218 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
15221 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
15222 4 space separated floating point adjustment values in the [-1,1] range,
15223 respectively to adjust the amount of cyan, magenta, yellow and black for the
15224 pixels of its range.
15226 @subsection Examples
15230 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
15231 increase magenta by 27% in blue areas:
15233 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
15237 Use a Photoshop selective color preset:
15239 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
15243 @anchor{separatefields}
15244 @section separatefields
15246 The @code{separatefields} takes a frame-based video input and splits
15247 each frame into its components fields, producing a new half height clip
15248 with twice the frame rate and twice the frame count.
15250 This filter use field-dominance information in frame to decide which
15251 of each pair of fields to place first in the output.
15252 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
15254 @section setdar, setsar
15256 The @code{setdar} filter sets the Display Aspect Ratio for the filter
15259 This is done by changing the specified Sample (aka Pixel) Aspect
15260 Ratio, according to the following equation:
15262 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
15265 Keep in mind that the @code{setdar} filter does not modify the pixel
15266 dimensions of the video frame. Also, the display aspect ratio set by
15267 this filter may be changed by later filters in the filterchain,
15268 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
15271 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
15272 the filter output video.
15274 Note that as a consequence of the application of this filter, the
15275 output display aspect ratio will change according to the equation
15278 Keep in mind that the sample aspect ratio set by the @code{setsar}
15279 filter may be changed by later filters in the filterchain, e.g. if
15280 another "setsar" or a "setdar" filter is applied.
15282 It accepts the following parameters:
15285 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
15286 Set the aspect ratio used by the filter.
15288 The parameter can be a floating point number string, an expression, or
15289 a string of the form @var{num}:@var{den}, where @var{num} and
15290 @var{den} are the numerator and denominator of the aspect ratio. If
15291 the parameter is not specified, it is assumed the value "0".
15292 In case the form "@var{num}:@var{den}" is used, the @code{:} character
15296 Set the maximum integer value to use for expressing numerator and
15297 denominator when reducing the expressed aspect ratio to a rational.
15298 Default value is @code{100}.
15302 The parameter @var{sar} is an expression containing
15303 the following constants:
15307 These are approximated values for the mathematical constants e
15308 (Euler's number), pi (Greek pi), and phi (the golden ratio).
15311 The input width and height.
15314 These are the same as @var{w} / @var{h}.
15317 The input sample aspect ratio.
15320 The input display aspect ratio. It is the same as
15321 (@var{w} / @var{h}) * @var{sar}.
15324 Horizontal and vertical chroma subsample values. For example, for the
15325 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15328 @subsection Examples
15333 To change the display aspect ratio to 16:9, specify one of the following:
15340 To change the sample aspect ratio to 10:11, specify:
15346 To set a display aspect ratio of 16:9, and specify a maximum integer value of
15347 1000 in the aspect ratio reduction, use the command:
15349 setdar=ratio=16/9:max=1000
15357 Force field for the output video frame.
15359 The @code{setfield} filter marks the interlace type field for the
15360 output frames. It does not change the input frame, but only sets the
15361 corresponding property, which affects how the frame is treated by
15362 following filters (e.g. @code{fieldorder} or @code{yadif}).
15364 The filter accepts the following options:
15369 Available values are:
15373 Keep the same field property.
15376 Mark the frame as bottom-field-first.
15379 Mark the frame as top-field-first.
15382 Mark the frame as progressive.
15389 Force frame parameter for the output video frame.
15391 The @code{setparams} filter marks interlace and color range for the
15392 output frames. It does not change the input frame, but only sets the
15393 corresponding property, which affects how the frame is treated by
15398 Available values are:
15402 Keep the same field property (default).
15405 Mark the frame as bottom-field-first.
15408 Mark the frame as top-field-first.
15411 Mark the frame as progressive.
15415 Available values are:
15419 Keep the same color range property (default).
15421 @item unspecified, unknown
15422 Mark the frame as unspecified color range.
15424 @item limited, tv, mpeg
15425 Mark the frame as limited range.
15427 @item full, pc, jpeg
15428 Mark the frame as full range.
15431 @item color_primaries
15432 Set the color primaries.
15433 Available values are:
15437 Keep the same color primaries property (default).
15454 Set the color transfer.
15455 Available values are:
15459 Keep the same color trc property (default).
15481 Set the colorspace.
15482 Available values are:
15486 Keep the same colorspace property (default).
15499 @item chroma-derived-nc
15500 @item chroma-derived-c
15507 Show a line containing various information for each input video frame.
15508 The input video is not modified.
15510 This filter supports the following options:
15514 Calculate checksums of each plane. By default enabled.
15517 The shown line contains a sequence of key/value pairs of the form
15518 @var{key}:@var{value}.
15520 The following values are shown in the output:
15524 The (sequential) number of the input frame, starting from 0.
15527 The Presentation TimeStamp of the input frame, expressed as a number of
15528 time base units. The time base unit depends on the filter input pad.
15531 The Presentation TimeStamp of the input frame, expressed as a number of
15535 The position of the frame in the input stream, or -1 if this information is
15536 unavailable and/or meaningless (for example in case of synthetic video).
15539 The pixel format name.
15542 The sample aspect ratio of the input frame, expressed in the form
15543 @var{num}/@var{den}.
15546 The size of the input frame. For the syntax of this option, check the
15547 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15550 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
15551 for bottom field first).
15554 This is 1 if the frame is a key frame, 0 otherwise.
15557 The picture type of the input frame ("I" for an I-frame, "P" for a
15558 P-frame, "B" for a B-frame, or "?" for an unknown type).
15559 Also refer to the documentation of the @code{AVPictureType} enum and of
15560 the @code{av_get_picture_type_char} function defined in
15561 @file{libavutil/avutil.h}.
15564 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
15566 @item plane_checksum
15567 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
15568 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
15571 @section showpalette
15573 Displays the 256 colors palette of each frame. This filter is only relevant for
15574 @var{pal8} pixel format frames.
15576 It accepts the following option:
15580 Set the size of the box used to represent one palette color entry. Default is
15581 @code{30} (for a @code{30x30} pixel box).
15584 @section shuffleframes
15586 Reorder and/or duplicate and/or drop video frames.
15588 It accepts the following parameters:
15592 Set the destination indexes of input frames.
15593 This is space or '|' separated list of indexes that maps input frames to output
15594 frames. Number of indexes also sets maximal value that each index may have.
15595 '-1' index have special meaning and that is to drop frame.
15598 The first frame has the index 0. The default is to keep the input unchanged.
15600 @subsection Examples
15604 Swap second and third frame of every three frames of the input:
15606 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
15610 Swap 10th and 1st frame of every ten frames of the input:
15612 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
15616 @section shuffleplanes
15618 Reorder and/or duplicate video planes.
15620 It accepts the following parameters:
15625 The index of the input plane to be used as the first output plane.
15628 The index of the input plane to be used as the second output plane.
15631 The index of the input plane to be used as the third output plane.
15634 The index of the input plane to be used as the fourth output plane.
15638 The first plane has the index 0. The default is to keep the input unchanged.
15640 @subsection Examples
15644 Swap the second and third planes of the input:
15646 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
15650 @anchor{signalstats}
15651 @section signalstats
15652 Evaluate various visual metrics that assist in determining issues associated
15653 with the digitization of analog video media.
15655 By default the filter will log these metadata values:
15659 Display the minimal Y value contained within the input frame. Expressed in
15663 Display the Y value at the 10% percentile within the input frame. Expressed in
15667 Display the average Y value within the input frame. Expressed in range of
15671 Display the Y value at the 90% percentile within the input frame. Expressed in
15675 Display the maximum Y value contained within the input frame. Expressed in
15679 Display the minimal U value contained within the input frame. Expressed in
15683 Display the U value at the 10% percentile within the input frame. Expressed in
15687 Display the average U value within the input frame. Expressed in range of
15691 Display the U value at the 90% percentile within the input frame. Expressed in
15695 Display the maximum U value contained within the input frame. Expressed in
15699 Display the minimal V value contained within the input frame. Expressed in
15703 Display the V value at the 10% percentile within the input frame. Expressed in
15707 Display the average V value within the input frame. Expressed in range of
15711 Display the V value at the 90% percentile within the input frame. Expressed in
15715 Display the maximum V value contained within the input frame. Expressed in
15719 Display the minimal saturation value contained within the input frame.
15720 Expressed in range of [0-~181.02].
15723 Display the saturation value at the 10% percentile within the input frame.
15724 Expressed in range of [0-~181.02].
15727 Display the average saturation value within the input frame. Expressed in range
15731 Display the saturation value at the 90% percentile within the input frame.
15732 Expressed in range of [0-~181.02].
15735 Display the maximum saturation value contained within the input frame.
15736 Expressed in range of [0-~181.02].
15739 Display the median value for hue within the input frame. Expressed in range of
15743 Display the average value for hue within the input frame. Expressed in range of
15747 Display the average of sample value difference between all values of the Y
15748 plane in the current frame and corresponding values of the previous input frame.
15749 Expressed in range of [0-255].
15752 Display the average of sample value difference between all values of the U
15753 plane in the current frame and corresponding values of the previous input frame.
15754 Expressed in range of [0-255].
15757 Display the average of sample value difference between all values of the V
15758 plane in the current frame and corresponding values of the previous input frame.
15759 Expressed in range of [0-255].
15762 Display bit depth of Y plane in current frame.
15763 Expressed in range of [0-16].
15766 Display bit depth of U plane in current frame.
15767 Expressed in range of [0-16].
15770 Display bit depth of V plane in current frame.
15771 Expressed in range of [0-16].
15774 The filter accepts the following options:
15780 @option{stat} specify an additional form of image analysis.
15781 @option{out} output video with the specified type of pixel highlighted.
15783 Both options accept the following values:
15787 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
15788 unlike the neighboring pixels of the same field. Examples of temporal outliers
15789 include the results of video dropouts, head clogs, or tape tracking issues.
15792 Identify @var{vertical line repetition}. Vertical line repetition includes
15793 similar rows of pixels within a frame. In born-digital video vertical line
15794 repetition is common, but this pattern is uncommon in video digitized from an
15795 analog source. When it occurs in video that results from the digitization of an
15796 analog source it can indicate concealment from a dropout compensator.
15799 Identify pixels that fall outside of legal broadcast range.
15803 Set the highlight color for the @option{out} option. The default color is
15807 @subsection Examples
15811 Output data of various video metrics:
15813 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
15817 Output specific data about the minimum and maximum values of the Y plane per frame:
15819 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
15823 Playback video while highlighting pixels that are outside of broadcast range in red.
15825 ffplay example.mov -vf signalstats="out=brng:color=red"
15829 Playback video with signalstats metadata drawn over the frame.
15831 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
15834 The contents of signalstat_drawtext.txt used in the command are:
15837 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
15838 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
15839 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
15840 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
15848 Calculates the MPEG-7 Video Signature. The filter can handle more than one
15849 input. In this case the matching between the inputs can be calculated additionally.
15850 The filter always passes through the first input. The signature of each stream can
15851 be written into a file.
15853 It accepts the following options:
15857 Enable or disable the matching process.
15859 Available values are:
15863 Disable the calculation of a matching (default).
15865 Calculate the matching for the whole video and output whether the whole video
15866 matches or only parts.
15868 Calculate only until a matching is found or the video ends. Should be faster in
15873 Set the number of inputs. The option value must be a non negative integer.
15874 Default value is 1.
15877 Set the path to which the output is written. If there is more than one input,
15878 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
15879 integer), that will be replaced with the input number. If no filename is
15880 specified, no output will be written. This is the default.
15883 Choose the output format.
15885 Available values are:
15889 Use the specified binary representation (default).
15891 Use the specified xml representation.
15895 Set threshold to detect one word as similar. The option value must be an integer
15896 greater than zero. The default value is 9000.
15899 Set threshold to detect all words as similar. The option value must be an integer
15900 greater than zero. The default value is 60000.
15903 Set threshold to detect frames as similar. The option value must be an integer
15904 greater than zero. The default value is 116.
15907 Set the minimum length of a sequence in frames to recognize it as matching
15908 sequence. The option value must be a non negative integer value.
15909 The default value is 0.
15912 Set the minimum relation, that matching frames to all frames must have.
15913 The option value must be a double value between 0 and 1. The default value is 0.5.
15916 @subsection Examples
15920 To calculate the signature of an input video and store it in signature.bin:
15922 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
15926 To detect whether two videos match and store the signatures in XML format in
15927 signature0.xml and signature1.xml:
15929 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 -
15937 Blur the input video without impacting the outlines.
15939 It accepts the following options:
15942 @item luma_radius, lr
15943 Set the luma radius. The option value must be a float number in
15944 the range [0.1,5.0] that specifies the variance of the gaussian filter
15945 used to blur the image (slower if larger). Default value is 1.0.
15947 @item luma_strength, ls
15948 Set the luma strength. The option value must be a float number
15949 in the range [-1.0,1.0] that configures the blurring. A value included
15950 in [0.0,1.0] will blur the image whereas a value included in
15951 [-1.0,0.0] will sharpen the image. Default value is 1.0.
15953 @item luma_threshold, lt
15954 Set the luma threshold used as a coefficient to determine
15955 whether a pixel should be blurred or not. The option value must be an
15956 integer in the range [-30,30]. A value of 0 will filter all the image,
15957 a value included in [0,30] will filter flat areas and a value included
15958 in [-30,0] will filter edges. Default value is 0.
15960 @item chroma_radius, cr
15961 Set the chroma radius. The option value must be a float number in
15962 the range [0.1,5.0] that specifies the variance of the gaussian filter
15963 used to blur the image (slower if larger). Default value is @option{luma_radius}.
15965 @item chroma_strength, cs
15966 Set the chroma strength. The option value must be a float number
15967 in the range [-1.0,1.0] that configures the blurring. A value included
15968 in [0.0,1.0] will blur the image whereas a value included in
15969 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
15971 @item chroma_threshold, ct
15972 Set the chroma threshold used as a coefficient to determine
15973 whether a pixel should be blurred or not. The option value must be an
15974 integer in the range [-30,30]. A value of 0 will filter all the image,
15975 a value included in [0,30] will filter flat areas and a value included
15976 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
15979 If a chroma option is not explicitly set, the corresponding luma value
15984 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
15986 This filter takes in input two input videos, the first input is
15987 considered the "main" source and is passed unchanged to the
15988 output. The second input is used as a "reference" video for computing
15991 Both video inputs must have the same resolution and pixel format for
15992 this filter to work correctly. Also it assumes that both inputs
15993 have the same number of frames, which are compared one by one.
15995 The filter stores the calculated SSIM of each frame.
15997 The description of the accepted parameters follows.
16000 @item stats_file, f
16001 If specified the filter will use the named file to save the SSIM of
16002 each individual frame. When filename equals "-" the data is sent to
16006 The file printed if @var{stats_file} is selected, contains a sequence of
16007 key/value pairs of the form @var{key}:@var{value} for each compared
16010 A description of each shown parameter follows:
16014 sequential number of the input frame, starting from 1
16016 @item Y, U, V, R, G, B
16017 SSIM of the compared frames for the component specified by the suffix.
16020 SSIM of the compared frames for the whole frame.
16023 Same as above but in dB representation.
16026 This filter also supports the @ref{framesync} options.
16030 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16031 [main][ref] ssim="stats_file=stats.log" [out]
16034 On this example the input file being processed is compared with the
16035 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
16036 is stored in @file{stats.log}.
16038 Another example with both psnr and ssim at same time:
16040 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
16045 Convert between different stereoscopic image formats.
16047 The filters accept the following options:
16051 Set stereoscopic image format of input.
16053 Available values for input image formats are:
16056 side by side parallel (left eye left, right eye right)
16059 side by side crosseye (right eye left, left eye right)
16062 side by side parallel with half width resolution
16063 (left eye left, right eye right)
16066 side by side crosseye with half width resolution
16067 (right eye left, left eye right)
16070 above-below (left eye above, right eye below)
16073 above-below (right eye above, left eye below)
16076 above-below with half height resolution
16077 (left eye above, right eye below)
16080 above-below with half height resolution
16081 (right eye above, left eye below)
16084 alternating frames (left eye first, right eye second)
16087 alternating frames (right eye first, left eye second)
16090 interleaved rows (left eye has top row, right eye starts on next row)
16093 interleaved rows (right eye has top row, left eye starts on next row)
16096 interleaved columns, left eye first
16099 interleaved columns, right eye first
16101 Default value is @samp{sbsl}.
16105 Set stereoscopic image format of output.
16109 side by side parallel (left eye left, right eye right)
16112 side by side crosseye (right eye left, left eye right)
16115 side by side parallel with half width resolution
16116 (left eye left, right eye right)
16119 side by side crosseye with half width resolution
16120 (right eye left, left eye right)
16123 above-below (left eye above, right eye below)
16126 above-below (right eye above, left eye below)
16129 above-below with half height resolution
16130 (left eye above, right eye below)
16133 above-below with half height resolution
16134 (right eye above, left eye below)
16137 alternating frames (left eye first, right eye second)
16140 alternating frames (right eye first, left eye second)
16143 interleaved rows (left eye has top row, right eye starts on next row)
16146 interleaved rows (right eye has top row, left eye starts on next row)
16149 anaglyph red/blue gray
16150 (red filter on left eye, blue filter on right eye)
16153 anaglyph red/green gray
16154 (red filter on left eye, green filter on right eye)
16157 anaglyph red/cyan gray
16158 (red filter on left eye, cyan filter on right eye)
16161 anaglyph red/cyan half colored
16162 (red filter on left eye, cyan filter on right eye)
16165 anaglyph red/cyan color
16166 (red filter on left eye, cyan filter on right eye)
16169 anaglyph red/cyan color optimized with the least squares projection of dubois
16170 (red filter on left eye, cyan filter on right eye)
16173 anaglyph green/magenta gray
16174 (green filter on left eye, magenta filter on right eye)
16177 anaglyph green/magenta half colored
16178 (green filter on left eye, magenta filter on right eye)
16181 anaglyph green/magenta colored
16182 (green filter on left eye, magenta filter on right eye)
16185 anaglyph green/magenta color optimized with the least squares projection of dubois
16186 (green filter on left eye, magenta filter on right eye)
16189 anaglyph yellow/blue gray
16190 (yellow filter on left eye, blue filter on right eye)
16193 anaglyph yellow/blue half colored
16194 (yellow filter on left eye, blue filter on right eye)
16197 anaglyph yellow/blue colored
16198 (yellow filter on left eye, blue filter on right eye)
16201 anaglyph yellow/blue color optimized with the least squares projection of dubois
16202 (yellow filter on left eye, blue filter on right eye)
16205 mono output (left eye only)
16208 mono output (right eye only)
16211 checkerboard, left eye first
16214 checkerboard, right eye first
16217 interleaved columns, left eye first
16220 interleaved columns, right eye first
16226 Default value is @samp{arcd}.
16229 @subsection Examples
16233 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
16239 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
16245 @section streamselect, astreamselect
16246 Select video or audio streams.
16248 The filter accepts the following options:
16252 Set number of inputs. Default is 2.
16255 Set input indexes to remap to outputs.
16258 @subsection Commands
16260 The @code{streamselect} and @code{astreamselect} filter supports the following
16265 Set input indexes to remap to outputs.
16268 @subsection Examples
16272 Select first 5 seconds 1st stream and rest of time 2nd stream:
16274 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
16278 Same as above, but for audio:
16280 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
16285 Apply sobel operator to input video stream.
16287 The filter accepts the following option:
16291 Set which planes will be processed, unprocessed planes will be copied.
16292 By default value 0xf, all planes will be processed.
16295 Set value which will be multiplied with filtered result.
16298 Set value which will be added to filtered result.
16304 Apply a simple postprocessing filter that compresses and decompresses the image
16305 at several (or - in the case of @option{quality} level @code{6} - all) shifts
16306 and average the results.
16308 The filter accepts the following options:
16312 Set quality. This option defines the number of levels for averaging. It accepts
16313 an integer in the range 0-6. If set to @code{0}, the filter will have no
16314 effect. A value of @code{6} means the higher quality. For each increment of
16315 that value the speed drops by a factor of approximately 2. Default value is
16319 Force a constant quantization parameter. If not set, the filter will use the QP
16320 from the video stream (if available).
16323 Set thresholding mode. Available modes are:
16327 Set hard thresholding (default).
16329 Set soft thresholding (better de-ringing effect, but likely blurrier).
16332 @item use_bframe_qp
16333 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
16334 option may cause flicker since the B-Frames have often larger QP. Default is
16335 @code{0} (not enabled).
16340 Scale the input by applying one of the super-resolution methods based on
16341 convolutional neural networks. Supported models:
16345 Super-Resolution Convolutional Neural Network model (SRCNN).
16346 See @url{https://arxiv.org/abs/1501.00092}.
16349 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
16350 See @url{https://arxiv.org/abs/1609.05158}.
16353 Training scripts as well as scripts for model generation are provided in
16354 the repository at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
16356 The filter accepts the following options:
16360 Specify which DNN backend to use for model loading and execution. This option accepts
16361 the following values:
16365 Native implementation of DNN loading and execution.
16368 TensorFlow backend. To enable this backend you
16369 need to install the TensorFlow for C library (see
16370 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
16371 @code{--enable-libtensorflow}
16374 Default value is @samp{native}.
16377 Set path to model file specifying network architecture and its parameters.
16378 Note that different backends use different file formats. TensorFlow backend
16379 can load files for both formats, while native backend can load files for only
16383 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
16384 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
16385 input upscaled using bicubic upscaling with proper scale factor.
16391 Draw subtitles on top of input video using the libass library.
16393 To enable compilation of this filter you need to configure FFmpeg with
16394 @code{--enable-libass}. This filter also requires a build with libavcodec and
16395 libavformat to convert the passed subtitles file to ASS (Advanced Substation
16396 Alpha) subtitles format.
16398 The filter accepts the following options:
16402 Set the filename of the subtitle file to read. It must be specified.
16404 @item original_size
16405 Specify the size of the original video, the video for which the ASS file
16406 was composed. For the syntax of this option, check the
16407 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16408 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
16409 correctly scale the fonts if the aspect ratio has been changed.
16412 Set a directory path containing fonts that can be used by the filter.
16413 These fonts will be used in addition to whatever the font provider uses.
16416 Process alpha channel, by default alpha channel is untouched.
16419 Set subtitles input character encoding. @code{subtitles} filter only. Only
16420 useful if not UTF-8.
16422 @item stream_index, si
16423 Set subtitles stream index. @code{subtitles} filter only.
16426 Override default style or script info parameters of the subtitles. It accepts a
16427 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
16430 If the first key is not specified, it is assumed that the first value
16431 specifies the @option{filename}.
16433 For example, to render the file @file{sub.srt} on top of the input
16434 video, use the command:
16439 which is equivalent to:
16441 subtitles=filename=sub.srt
16444 To render the default subtitles stream from file @file{video.mkv}, use:
16446 subtitles=video.mkv
16449 To render the second subtitles stream from that file, use:
16451 subtitles=video.mkv:si=1
16454 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
16455 @code{DejaVu Serif}, use:
16457 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
16460 @section super2xsai
16462 Scale the input by 2x and smooth using the Super2xSaI (Scale and
16463 Interpolate) pixel art scaling algorithm.
16465 Useful for enlarging pixel art images without reducing sharpness.
16469 Swap two rectangular objects in video.
16471 This filter accepts the following options:
16481 Set 1st rect x coordinate.
16484 Set 1st rect y coordinate.
16487 Set 2nd rect x coordinate.
16490 Set 2nd rect y coordinate.
16492 All expressions are evaluated once for each frame.
16495 The all options are expressions containing the following constants:
16500 The input width and height.
16503 same as @var{w} / @var{h}
16506 input sample aspect ratio
16509 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
16512 The number of the input frame, starting from 0.
16515 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
16518 the position in the file of the input frame, NAN if unknown
16526 Apply telecine process to the video.
16528 This filter accepts the following options:
16537 The default value is @code{top}.
16541 A string of numbers representing the pulldown pattern you wish to apply.
16542 The default value is @code{23}.
16546 Some typical patterns:
16551 24p: 2332 (preferred)
16558 24p: 222222222223 ("Euro pulldown")
16565 Apply threshold effect to video stream.
16567 This filter needs four video streams to perform thresholding.
16568 First stream is stream we are filtering.
16569 Second stream is holding threshold values, third stream is holding min values,
16570 and last, fourth stream is holding max values.
16572 The filter accepts the following option:
16576 Set which planes will be processed, unprocessed planes will be copied.
16577 By default value 0xf, all planes will be processed.
16580 For example if first stream pixel's component value is less then threshold value
16581 of pixel component from 2nd threshold stream, third stream value will picked,
16582 otherwise fourth stream pixel component value will be picked.
16584 Using color source filter one can perform various types of thresholding:
16586 @subsection Examples
16590 Binary threshold, using gray color as threshold:
16592 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
16596 Inverted binary threshold, using gray color as threshold:
16598 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
16602 Truncate binary threshold, using gray color as threshold:
16604 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
16608 Threshold to zero, using gray color as threshold:
16610 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
16614 Inverted threshold to zero, using gray color as threshold:
16616 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
16621 Select the most representative frame in a given sequence of consecutive frames.
16623 The filter accepts the following options:
16627 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
16628 will pick one of them, and then handle the next batch of @var{n} frames until
16629 the end. Default is @code{100}.
16632 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
16633 value will result in a higher memory usage, so a high value is not recommended.
16635 @subsection Examples
16639 Extract one picture each 50 frames:
16645 Complete example of a thumbnail creation with @command{ffmpeg}:
16647 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
16653 Tile several successive frames together.
16655 The filter accepts the following options:
16660 Set the grid size (i.e. the number of lines and columns). For the syntax of
16661 this option, check the
16662 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16665 Set the maximum number of frames to render in the given area. It must be less
16666 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
16667 the area will be used.
16670 Set the outer border margin in pixels.
16673 Set the inner border thickness (i.e. the number of pixels between frames). For
16674 more advanced padding options (such as having different values for the edges),
16675 refer to the pad video filter.
16678 Specify the color of the unused area. For the syntax of this option, check the
16679 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16680 The default value of @var{color} is "black".
16683 Set the number of frames to overlap when tiling several successive frames together.
16684 The value must be between @code{0} and @var{nb_frames - 1}.
16687 Set the number of frames to initially be empty before displaying first output frame.
16688 This controls how soon will one get first output frame.
16689 The value must be between @code{0} and @var{nb_frames - 1}.
16692 @subsection Examples
16696 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
16698 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
16700 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
16701 duplicating each output frame to accommodate the originally detected frame
16705 Display @code{5} pictures in an area of @code{3x2} frames,
16706 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
16707 mixed flat and named options:
16709 tile=3x2:nb_frames=5:padding=7:margin=2
16713 @section tinterlace
16715 Perform various types of temporal field interlacing.
16717 Frames are counted starting from 1, so the first input frame is
16720 The filter accepts the following options:
16725 Specify the mode of the interlacing. This option can also be specified
16726 as a value alone. See below for a list of values for this option.
16728 Available values are:
16732 Move odd frames into the upper field, even into the lower field,
16733 generating a double height frame at half frame rate.
16737 Frame 1 Frame 2 Frame 3 Frame 4
16739 11111 22222 33333 44444
16740 11111 22222 33333 44444
16741 11111 22222 33333 44444
16742 11111 22222 33333 44444
16756 Only output odd frames, even frames are dropped, generating a frame with
16757 unchanged height at half frame rate.
16762 Frame 1 Frame 2 Frame 3 Frame 4
16764 11111 22222 33333 44444
16765 11111 22222 33333 44444
16766 11111 22222 33333 44444
16767 11111 22222 33333 44444
16777 Only output even frames, odd frames are dropped, generating a frame with
16778 unchanged height at half frame rate.
16783 Frame 1 Frame 2 Frame 3 Frame 4
16785 11111 22222 33333 44444
16786 11111 22222 33333 44444
16787 11111 22222 33333 44444
16788 11111 22222 33333 44444
16798 Expand each frame to full height, but pad alternate lines with black,
16799 generating a frame with double height at the same input frame rate.
16804 Frame 1 Frame 2 Frame 3 Frame 4
16806 11111 22222 33333 44444
16807 11111 22222 33333 44444
16808 11111 22222 33333 44444
16809 11111 22222 33333 44444
16812 11111 ..... 33333 .....
16813 ..... 22222 ..... 44444
16814 11111 ..... 33333 .....
16815 ..... 22222 ..... 44444
16816 11111 ..... 33333 .....
16817 ..... 22222 ..... 44444
16818 11111 ..... 33333 .....
16819 ..... 22222 ..... 44444
16823 @item interleave_top, 4
16824 Interleave the upper field from odd frames with the lower field from
16825 even frames, generating a frame with unchanged height at half frame rate.
16830 Frame 1 Frame 2 Frame 3 Frame 4
16832 11111<- 22222 33333<- 44444
16833 11111 22222<- 33333 44444<-
16834 11111<- 22222 33333<- 44444
16835 11111 22222<- 33333 44444<-
16845 @item interleave_bottom, 5
16846 Interleave the lower field from odd frames with the upper field from
16847 even frames, generating a frame with unchanged height at half frame rate.
16852 Frame 1 Frame 2 Frame 3 Frame 4
16854 11111 22222<- 33333 44444<-
16855 11111<- 22222 33333<- 44444
16856 11111 22222<- 33333 44444<-
16857 11111<- 22222 33333<- 44444
16867 @item interlacex2, 6
16868 Double frame rate with unchanged height. Frames are inserted each
16869 containing the second temporal field from the previous input frame and
16870 the first temporal field from the next input frame. This mode relies on
16871 the top_field_first flag. Useful for interlaced video displays with no
16872 field synchronisation.
16877 Frame 1 Frame 2 Frame 3 Frame 4
16879 11111 22222 33333 44444
16880 11111 22222 33333 44444
16881 11111 22222 33333 44444
16882 11111 22222 33333 44444
16885 11111 22222 22222 33333 33333 44444 44444
16886 11111 11111 22222 22222 33333 33333 44444
16887 11111 22222 22222 33333 33333 44444 44444
16888 11111 11111 22222 22222 33333 33333 44444
16893 Move odd frames into the upper field, even into the lower field,
16894 generating a double height frame at same frame rate.
16899 Frame 1 Frame 2 Frame 3 Frame 4
16901 11111 22222 33333 44444
16902 11111 22222 33333 44444
16903 11111 22222 33333 44444
16904 11111 22222 33333 44444
16907 11111 33333 33333 55555
16908 22222 22222 44444 44444
16909 11111 33333 33333 55555
16910 22222 22222 44444 44444
16911 11111 33333 33333 55555
16912 22222 22222 44444 44444
16913 11111 33333 33333 55555
16914 22222 22222 44444 44444
16919 Numeric values are deprecated but are accepted for backward
16920 compatibility reasons.
16922 Default mode is @code{merge}.
16925 Specify flags influencing the filter process.
16927 Available value for @var{flags} is:
16930 @item low_pass_filter, vlfp
16931 Enable linear vertical low-pass filtering in the filter.
16932 Vertical low-pass filtering is required when creating an interlaced
16933 destination from a progressive source which contains high-frequency
16934 vertical detail. Filtering will reduce interlace 'twitter' and Moire
16937 @item complex_filter, cvlfp
16938 Enable complex vertical low-pass filtering.
16939 This will slightly less reduce interlace 'twitter' and Moire
16940 patterning but better retain detail and subjective sharpness impression.
16944 Vertical low-pass filtering can only be enabled for @option{mode}
16945 @var{interleave_top} and @var{interleave_bottom}.
16951 Mix successive video frames.
16953 A description of the accepted options follows.
16957 The number of successive frames to mix. If unspecified, it defaults to 3.
16960 Specify weight of each input video frame.
16961 Each weight is separated by space. If number of weights is smaller than
16962 number of @var{frames} last specified weight will be used for all remaining
16966 Specify scale, if it is set it will be multiplied with sum
16967 of each weight multiplied with pixel values to give final destination
16968 pixel value. By default @var{scale} is auto scaled to sum of weights.
16971 @subsection Examples
16975 Average 7 successive frames:
16977 tmix=frames=7:weights="1 1 1 1 1 1 1"
16981 Apply simple temporal convolution:
16983 tmix=frames=3:weights="-1 3 -1"
16987 Similar as above but only showing temporal differences:
16989 tmix=frames=3:weights="-1 2 -1":scale=1
16995 Tone map colors from different dynamic ranges.
16997 This filter expects data in single precision floating point, as it needs to
16998 operate on (and can output) out-of-range values. Another filter, such as
16999 @ref{zscale}, is needed to convert the resulting frame to a usable format.
17001 The tonemapping algorithms implemented only work on linear light, so input
17002 data should be linearized beforehand (and possibly correctly tagged).
17005 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
17008 @subsection Options
17009 The filter accepts the following options.
17013 Set the tone map algorithm to use.
17015 Possible values are:
17018 Do not apply any tone map, only desaturate overbright pixels.
17021 Hard-clip any out-of-range values. Use it for perfect color accuracy for
17022 in-range values, while distorting out-of-range values.
17025 Stretch the entire reference gamut to a linear multiple of the display.
17028 Fit a logarithmic transfer between the tone curves.
17031 Preserve overall image brightness with a simple curve, using nonlinear
17032 contrast, which results in flattening details and degrading color accuracy.
17035 Preserve both dark and bright details better than @var{reinhard}, at the cost
17036 of slightly darkening everything. Use it when detail preservation is more
17037 important than color and brightness accuracy.
17040 Smoothly map out-of-range values, while retaining contrast and colors for
17041 in-range material as much as possible. Use it when color accuracy is more
17042 important than detail preservation.
17048 Tune the tone mapping algorithm.
17050 This affects the following algorithms:
17056 Specifies the scale factor to use while stretching.
17060 Specifies the exponent of the function.
17064 Specify an extra linear coefficient to multiply into the signal before clipping.
17068 Specify the local contrast coefficient at the display peak.
17069 Default to 0.5, which means that in-gamut values will be about half as bright
17076 Specify the transition point from linear to mobius transform. Every value
17077 below this point is guaranteed to be mapped 1:1. The higher the value, the
17078 more accurate the result will be, at the cost of losing bright details.
17079 Default to 0.3, which due to the steep initial slope still preserves in-range
17080 colors fairly accurately.
17084 Apply desaturation for highlights that exceed this level of brightness. The
17085 higher the parameter, the more color information will be preserved. This
17086 setting helps prevent unnaturally blown-out colors for super-highlights, by
17087 (smoothly) turning into white instead. This makes images feel more natural,
17088 at the cost of reducing information about out-of-range colors.
17090 The default of 2.0 is somewhat conservative and will mostly just apply to
17091 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
17093 This option works only if the input frame has a supported color tag.
17096 Override signal/nominal/reference peak with this value. Useful when the
17097 embedded peak information in display metadata is not reliable or when tone
17098 mapping from a lower range to a higher range.
17103 Temporarily pad video frames.
17105 The filter accepts the following options:
17109 Specify number of delay frames before input video stream.
17112 Specify number of padding frames after input video stream.
17113 Set to -1 to pad indefinitely.
17116 Set kind of frames added to beginning of stream.
17117 Can be either @var{add} or @var{clone}.
17118 With @var{add} frames of solid-color are added.
17119 With @var{clone} frames are clones of first frame.
17122 Set kind of frames added to end of stream.
17123 Can be either @var{add} or @var{clone}.
17124 With @var{add} frames of solid-color are added.
17125 With @var{clone} frames are clones of last frame.
17127 @item start_duration, stop_duration
17128 Specify the duration of the start/stop delay. See
17129 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17130 for the accepted syntax.
17131 These options override @var{start} and @var{stop}.
17134 Specify the color of the padded area. For the syntax of this option,
17135 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
17136 manual,ffmpeg-utils}.
17138 The default value of @var{color} is "black".
17144 Transpose rows with columns in the input video and optionally flip it.
17146 It accepts the following parameters:
17151 Specify the transposition direction.
17153 Can assume the following values:
17155 @item 0, 4, cclock_flip
17156 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
17164 Rotate by 90 degrees clockwise, that is:
17172 Rotate by 90 degrees counterclockwise, that is:
17179 @item 3, 7, clock_flip
17180 Rotate by 90 degrees clockwise and vertically flip, that is:
17188 For values between 4-7, the transposition is only done if the input
17189 video geometry is portrait and not landscape. These values are
17190 deprecated, the @code{passthrough} option should be used instead.
17192 Numerical values are deprecated, and should be dropped in favor of
17193 symbolic constants.
17196 Do not apply the transposition if the input geometry matches the one
17197 specified by the specified value. It accepts the following values:
17200 Always apply transposition.
17202 Preserve portrait geometry (when @var{height} >= @var{width}).
17204 Preserve landscape geometry (when @var{width} >= @var{height}).
17207 Default value is @code{none}.
17210 For example to rotate by 90 degrees clockwise and preserve portrait
17213 transpose=dir=1:passthrough=portrait
17216 The command above can also be specified as:
17218 transpose=1:portrait
17221 @section transpose_npp
17223 Transpose rows with columns in the input video and optionally flip it.
17224 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
17226 It accepts the following parameters:
17231 Specify the transposition direction.
17233 Can assume the following values:
17236 Rotate by 90 degrees counterclockwise and vertically flip. (default)
17239 Rotate by 90 degrees clockwise.
17242 Rotate by 90 degrees counterclockwise.
17245 Rotate by 90 degrees clockwise and vertically flip.
17249 Do not apply the transposition if the input geometry matches the one
17250 specified by the specified value. It accepts the following values:
17253 Always apply transposition. (default)
17255 Preserve portrait geometry (when @var{height} >= @var{width}).
17257 Preserve landscape geometry (when @var{width} >= @var{height}).
17263 Trim the input so that the output contains one continuous subpart of the input.
17265 It accepts the following parameters:
17268 Specify the time of the start of the kept section, i.e. the frame with the
17269 timestamp @var{start} will be the first frame in the output.
17272 Specify the time of the first frame that will be dropped, i.e. the frame
17273 immediately preceding the one with the timestamp @var{end} will be the last
17274 frame in the output.
17277 This is the same as @var{start}, except this option sets the start timestamp
17278 in timebase units instead of seconds.
17281 This is the same as @var{end}, except this option sets the end timestamp
17282 in timebase units instead of seconds.
17285 The maximum duration of the output in seconds.
17288 The number of the first frame that should be passed to the output.
17291 The number of the first frame that should be dropped.
17294 @option{start}, @option{end}, and @option{duration} are expressed as time
17295 duration specifications; see
17296 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17297 for the accepted syntax.
17299 Note that the first two sets of the start/end options and the @option{duration}
17300 option look at the frame timestamp, while the _frame variants simply count the
17301 frames that pass through the filter. Also note that this filter does not modify
17302 the timestamps. If you wish for the output timestamps to start at zero, insert a
17303 setpts filter after the trim filter.
17305 If multiple start or end options are set, this filter tries to be greedy and
17306 keep all the frames that match at least one of the specified constraints. To keep
17307 only the part that matches all the constraints at once, chain multiple trim
17310 The defaults are such that all the input is kept. So it is possible to set e.g.
17311 just the end values to keep everything before the specified time.
17316 Drop everything except the second minute of input:
17318 ffmpeg -i INPUT -vf trim=60:120
17322 Keep only the first second:
17324 ffmpeg -i INPUT -vf trim=duration=1
17329 @section unpremultiply
17330 Apply alpha unpremultiply effect to input video stream using first plane
17331 of second stream as alpha.
17333 Both streams must have same dimensions and same pixel format.
17335 The filter accepts the following option:
17339 Set which planes will be processed, unprocessed planes will be copied.
17340 By default value 0xf, all planes will be processed.
17342 If the format has 1 or 2 components, then luma is bit 0.
17343 If the format has 3 or 4 components:
17344 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
17345 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
17346 If present, the alpha channel is always the last bit.
17349 Do not require 2nd input for processing, instead use alpha plane from input stream.
17355 Sharpen or blur the input video.
17357 It accepts the following parameters:
17360 @item luma_msize_x, lx
17361 Set the luma matrix horizontal size. It must be an odd integer between
17362 3 and 23. The default value is 5.
17364 @item luma_msize_y, ly
17365 Set the luma matrix vertical size. It must be an odd integer between 3
17366 and 23. The default value is 5.
17368 @item luma_amount, la
17369 Set the luma effect strength. It must be a floating point number, reasonable
17370 values lay between -1.5 and 1.5.
17372 Negative values will blur the input video, while positive values will
17373 sharpen it, a value of zero will disable the effect.
17375 Default value is 1.0.
17377 @item chroma_msize_x, cx
17378 Set the chroma matrix horizontal size. It must be an odd integer
17379 between 3 and 23. The default value is 5.
17381 @item chroma_msize_y, cy
17382 Set the chroma matrix vertical size. It must be an odd integer
17383 between 3 and 23. The default value is 5.
17385 @item chroma_amount, ca
17386 Set the chroma effect strength. It must be a floating point number, reasonable
17387 values lay between -1.5 and 1.5.
17389 Negative values will blur the input video, while positive values will
17390 sharpen it, a value of zero will disable the effect.
17392 Default value is 0.0.
17396 All parameters are optional and default to the equivalent of the
17397 string '5:5:1.0:5:5:0.0'.
17399 @subsection Examples
17403 Apply strong luma sharpen effect:
17405 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
17409 Apply a strong blur of both luma and chroma parameters:
17411 unsharp=7:7:-2:7:7:-2
17417 Apply ultra slow/simple postprocessing filter that compresses and decompresses
17418 the image at several (or - in the case of @option{quality} level @code{8} - all)
17419 shifts and average the results.
17421 The way this differs from the behavior of spp is that uspp actually encodes &
17422 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
17423 DCT similar to MJPEG.
17425 The filter accepts the following options:
17429 Set quality. This option defines the number of levels for averaging. It accepts
17430 an integer in the range 0-8. If set to @code{0}, the filter will have no
17431 effect. A value of @code{8} means the higher quality. For each increment of
17432 that value the speed drops by a factor of approximately 2. Default value is
17436 Force a constant quantization parameter. If not set, the filter will use the QP
17437 from the video stream (if available).
17440 @section vaguedenoiser
17442 Apply a wavelet based denoiser.
17444 It transforms each frame from the video input into the wavelet domain,
17445 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
17446 the obtained coefficients. It does an inverse wavelet transform after.
17447 Due to wavelet properties, it should give a nice smoothed result, and
17448 reduced noise, without blurring picture features.
17450 This filter accepts the following options:
17454 The filtering strength. The higher, the more filtered the video will be.
17455 Hard thresholding can use a higher threshold than soft thresholding
17456 before the video looks overfiltered. Default value is 2.
17459 The filtering method the filter will use.
17461 It accepts the following values:
17464 All values under the threshold will be zeroed.
17467 All values under the threshold will be zeroed. All values above will be
17468 reduced by the threshold.
17471 Scales or nullifies coefficients - intermediary between (more) soft and
17472 (less) hard thresholding.
17475 Default is garrote.
17478 Number of times, the wavelet will decompose the picture. Picture can't
17479 be decomposed beyond a particular point (typically, 8 for a 640x480
17480 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
17483 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
17486 A list of the planes to process. By default all planes are processed.
17489 @section vectorscope
17491 Display 2 color component values in the two dimensional graph (which is called
17494 This filter accepts the following options:
17498 Set vectorscope mode.
17500 It accepts the following values:
17503 Gray values are displayed on graph, higher brightness means more pixels have
17504 same component color value on location in graph. This is the default mode.
17507 Gray values are displayed on graph. Surrounding pixels values which are not
17508 present in video frame are drawn in gradient of 2 color components which are
17509 set by option @code{x} and @code{y}. The 3rd color component is static.
17512 Actual color components values present in video frame are displayed on graph.
17515 Similar as color2 but higher frequency of same values @code{x} and @code{y}
17516 on graph increases value of another color component, which is luminance by
17517 default values of @code{x} and @code{y}.
17520 Actual colors present in video frame are displayed on graph. If two different
17521 colors map to same position on graph then color with higher value of component
17522 not present in graph is picked.
17525 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
17526 component picked from radial gradient.
17530 Set which color component will be represented on X-axis. Default is @code{1}.
17533 Set which color component will be represented on Y-axis. Default is @code{2}.
17536 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
17537 of color component which represents frequency of (X, Y) location in graph.
17542 No envelope, this is default.
17545 Instant envelope, even darkest single pixel will be clearly highlighted.
17548 Hold maximum and minimum values presented in graph over time. This way you
17549 can still spot out of range values without constantly looking at vectorscope.
17552 Peak and instant envelope combined together.
17556 Set what kind of graticule to draw.
17564 Set graticule opacity.
17567 Set graticule flags.
17571 Draw graticule for white point.
17574 Draw graticule for black point.
17577 Draw color points short names.
17581 Set background opacity.
17583 @item lthreshold, l
17584 Set low threshold for color component not represented on X or Y axis.
17585 Values lower than this value will be ignored. Default is 0.
17586 Note this value is multiplied with actual max possible value one pixel component
17587 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
17590 @item hthreshold, h
17591 Set high threshold for color component not represented on X or Y axis.
17592 Values higher than this value will be ignored. Default is 1.
17593 Note this value is multiplied with actual max possible value one pixel component
17594 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
17595 is 0.9 * 255 = 230.
17597 @item colorspace, c
17598 Set what kind of colorspace to use when drawing graticule.
17607 @anchor{vidstabdetect}
17608 @section vidstabdetect
17610 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
17611 @ref{vidstabtransform} for pass 2.
17613 This filter generates a file with relative translation and rotation
17614 transform information about subsequent frames, which is then used by
17615 the @ref{vidstabtransform} filter.
17617 To enable compilation of this filter you need to configure FFmpeg with
17618 @code{--enable-libvidstab}.
17620 This filter accepts the following options:
17624 Set the path to the file used to write the transforms information.
17625 Default value is @file{transforms.trf}.
17628 Set how shaky the video is and how quick the camera is. It accepts an
17629 integer in the range 1-10, a value of 1 means little shakiness, a
17630 value of 10 means strong shakiness. Default value is 5.
17633 Set the accuracy of the detection process. It must be a value in the
17634 range 1-15. A value of 1 means low accuracy, a value of 15 means high
17635 accuracy. Default value is 15.
17638 Set stepsize of the search process. The region around minimum is
17639 scanned with 1 pixel resolution. Default value is 6.
17642 Set minimum contrast. Below this value a local measurement field is
17643 discarded. Must be a floating point value in the range 0-1. Default
17647 Set reference frame number for tripod mode.
17649 If enabled, the motion of the frames is compared to a reference frame
17650 in the filtered stream, identified by the specified number. The idea
17651 is to compensate all movements in a more-or-less static scene and keep
17652 the camera view absolutely still.
17654 If set to 0, it is disabled. The frames are counted starting from 1.
17657 Show fields and transforms in the resulting frames. It accepts an
17658 integer in the range 0-2. Default value is 0, which disables any
17662 @subsection Examples
17666 Use default values:
17672 Analyze strongly shaky movie and put the results in file
17673 @file{mytransforms.trf}:
17675 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
17679 Visualize the result of internal transformations in the resulting
17682 vidstabdetect=show=1
17686 Analyze a video with medium shakiness using @command{ffmpeg}:
17688 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
17692 @anchor{vidstabtransform}
17693 @section vidstabtransform
17695 Video stabilization/deshaking: pass 2 of 2,
17696 see @ref{vidstabdetect} for pass 1.
17698 Read a file with transform information for each frame and
17699 apply/compensate them. Together with the @ref{vidstabdetect}
17700 filter this can be used to deshake videos. See also
17701 @url{http://public.hronopik.de/vid.stab}. It is important to also use
17702 the @ref{unsharp} filter, see below.
17704 To enable compilation of this filter you need to configure FFmpeg with
17705 @code{--enable-libvidstab}.
17707 @subsection Options
17711 Set path to the file used to read the transforms. Default value is
17712 @file{transforms.trf}.
17715 Set the number of frames (value*2 + 1) used for lowpass filtering the
17716 camera movements. Default value is 10.
17718 For example a number of 10 means that 21 frames are used (10 in the
17719 past and 10 in the future) to smoothen the motion in the video. A
17720 larger value leads to a smoother video, but limits the acceleration of
17721 the camera (pan/tilt movements). 0 is a special case where a static
17722 camera is simulated.
17725 Set the camera path optimization algorithm.
17727 Accepted values are:
17730 gaussian kernel low-pass filter on camera motion (default)
17732 averaging on transformations
17736 Set maximal number of pixels to translate frames. Default value is -1,
17740 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
17741 value is -1, meaning no limit.
17744 Specify how to deal with borders that may be visible due to movement
17747 Available values are:
17750 keep image information from previous frame (default)
17752 fill the border black
17756 Invert transforms if set to 1. Default value is 0.
17759 Consider transforms as relative to previous frame if set to 1,
17760 absolute if set to 0. Default value is 0.
17763 Set percentage to zoom. A positive value will result in a zoom-in
17764 effect, a negative value in a zoom-out effect. Default value is 0 (no
17768 Set optimal zooming to avoid borders.
17770 Accepted values are:
17775 optimal static zoom value is determined (only very strong movements
17776 will lead to visible borders) (default)
17778 optimal adaptive zoom value is determined (no borders will be
17779 visible), see @option{zoomspeed}
17782 Note that the value given at zoom is added to the one calculated here.
17785 Set percent to zoom maximally each frame (enabled when
17786 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
17790 Specify type of interpolation.
17792 Available values are:
17797 linear only horizontal
17799 linear in both directions (default)
17801 cubic in both directions (slow)
17805 Enable virtual tripod mode if set to 1, which is equivalent to
17806 @code{relative=0:smoothing=0}. Default value is 0.
17808 Use also @code{tripod} option of @ref{vidstabdetect}.
17811 Increase log verbosity if set to 1. Also the detected global motions
17812 are written to the temporary file @file{global_motions.trf}. Default
17816 @subsection Examples
17820 Use @command{ffmpeg} for a typical stabilization with default values:
17822 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
17825 Note the use of the @ref{unsharp} filter which is always recommended.
17828 Zoom in a bit more and load transform data from a given file:
17830 vidstabtransform=zoom=5:input="mytransforms.trf"
17834 Smoothen the video even more:
17836 vidstabtransform=smoothing=30
17842 Flip the input video vertically.
17844 For example, to vertically flip a video with @command{ffmpeg}:
17846 ffmpeg -i in.avi -vf "vflip" out.avi
17851 Detect variable frame rate video.
17853 This filter tries to detect if the input is variable or constant frame rate.
17855 At end it will output number of frames detected as having variable delta pts,
17856 and ones with constant delta pts.
17857 If there was frames with variable delta, than it will also show min and max delta
17862 Boost or alter saturation.
17864 The filter accepts the following options:
17867 Set strength of boost if positive value or strength of alter if negative value.
17868 Default is 0. Allowed range is from -2 to 2.
17871 Set the red balance. Default is 1. Allowed range is from -10 to 10.
17874 Set the green balance. Default is 1. Allowed range is from -10 to 10.
17877 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
17880 Set the red luma coefficient.
17883 Set the green luma coefficient.
17886 Set the blue luma coefficient.
17892 Make or reverse a natural vignetting effect.
17894 The filter accepts the following options:
17898 Set lens angle expression as a number of radians.
17900 The value is clipped in the @code{[0,PI/2]} range.
17902 Default value: @code{"PI/5"}
17906 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
17910 Set forward/backward mode.
17912 Available modes are:
17915 The larger the distance from the central point, the darker the image becomes.
17918 The larger the distance from the central point, the brighter the image becomes.
17919 This can be used to reverse a vignette effect, though there is no automatic
17920 detection to extract the lens @option{angle} and other settings (yet). It can
17921 also be used to create a burning effect.
17924 Default value is @samp{forward}.
17927 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
17929 It accepts the following values:
17932 Evaluate expressions only once during the filter initialization.
17935 Evaluate expressions for each incoming frame. This is way slower than the
17936 @samp{init} mode since it requires all the scalers to be re-computed, but it
17937 allows advanced dynamic expressions.
17940 Default value is @samp{init}.
17943 Set dithering to reduce the circular banding effects. Default is @code{1}
17947 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
17948 Setting this value to the SAR of the input will make a rectangular vignetting
17949 following the dimensions of the video.
17951 Default is @code{1/1}.
17954 @subsection Expressions
17956 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
17957 following parameters.
17962 input width and height
17965 the number of input frame, starting from 0
17968 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
17969 @var{TB} units, NAN if undefined
17972 frame rate of the input video, NAN if the input frame rate is unknown
17975 the PTS (Presentation TimeStamp) of the filtered video frame,
17976 expressed in seconds, NAN if undefined
17979 time base of the input video
17983 @subsection Examples
17987 Apply simple strong vignetting effect:
17993 Make a flickering vignetting:
17995 vignette='PI/4+random(1)*PI/50':eval=frame
18000 @section vmafmotion
18002 Obtain the average vmaf motion score of a video.
18003 It is one of the component filters of VMAF.
18005 The obtained average motion score is printed through the logging system.
18007 In the below example the input file @file{ref.mpg} is being processed and score
18011 ffmpeg -i ref.mpg -lavfi vmafmotion -f null -
18015 Stack input videos vertically.
18017 All streams must be of same pixel format and of same width.
18019 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
18020 to create same output.
18022 The filter accept the following option:
18026 Set number of input streams. Default is 2.
18029 If set to 1, force the output to terminate when the shortest input
18030 terminates. Default value is 0.
18035 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
18036 Deinterlacing Filter").
18038 Based on the process described by Martin Weston for BBC R&D, and
18039 implemented based on the de-interlace algorithm written by Jim
18040 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
18041 uses filter coefficients calculated by BBC R&D.
18043 There are two sets of filter coefficients, so called "simple":
18044 and "complex". Which set of filter coefficients is used can
18045 be set by passing an optional parameter:
18049 Set the interlacing filter coefficients. Accepts one of the following values:
18053 Simple filter coefficient set.
18055 More-complex filter coefficient set.
18057 Default value is @samp{complex}.
18060 Specify which frames to deinterlace. Accept one of the following values:
18064 Deinterlace all frames,
18066 Only deinterlace frames marked as interlaced.
18069 Default value is @samp{all}.
18073 Video waveform monitor.
18075 The waveform monitor plots color component intensity. By default luminance
18076 only. Each column of the waveform corresponds to a column of pixels in the
18079 It accepts the following options:
18083 Can be either @code{row}, or @code{column}. Default is @code{column}.
18084 In row mode, the graph on the left side represents color component value 0 and
18085 the right side represents value = 255. In column mode, the top side represents
18086 color component value = 0 and bottom side represents value = 255.
18089 Set intensity. Smaller values are useful to find out how many values of the same
18090 luminance are distributed across input rows/columns.
18091 Default value is @code{0.04}. Allowed range is [0, 1].
18094 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
18095 In mirrored mode, higher values will be represented on the left
18096 side for @code{row} mode and at the top for @code{column} mode. Default is
18097 @code{1} (mirrored).
18101 It accepts the following values:
18104 Presents information identical to that in the @code{parade}, except
18105 that the graphs representing color components are superimposed directly
18108 This display mode makes it easier to spot relative differences or similarities
18109 in overlapping areas of the color components that are supposed to be identical,
18110 such as neutral whites, grays, or blacks.
18113 Display separate graph for the color components side by side in
18114 @code{row} mode or one below the other in @code{column} mode.
18117 Display separate graph for the color components side by side in
18118 @code{column} mode or one below the other in @code{row} mode.
18120 Using this display mode makes it easy to spot color casts in the highlights
18121 and shadows of an image, by comparing the contours of the top and the bottom
18122 graphs of each waveform. Since whites, grays, and blacks are characterized
18123 by exactly equal amounts of red, green, and blue, neutral areas of the picture
18124 should display three waveforms of roughly equal width/height. If not, the
18125 correction is easy to perform by making level adjustments the three waveforms.
18127 Default is @code{stack}.
18129 @item components, c
18130 Set which color components to display. Default is 1, which means only luminance
18131 or red color component if input is in RGB colorspace. If is set for example to
18132 7 it will display all 3 (if) available color components.
18137 No envelope, this is default.
18140 Instant envelope, minimum and maximum values presented in graph will be easily
18141 visible even with small @code{step} value.
18144 Hold minimum and maximum values presented in graph across time. This way you
18145 can still spot out of range values without constantly looking at waveforms.
18148 Peak and instant envelope combined together.
18154 No filtering, this is default.
18157 Luma and chroma combined together.
18160 Similar as above, but shows difference between blue and red chroma.
18163 Similar as above, but use different colors.
18166 Displays only chroma.
18169 Displays actual color value on waveform.
18172 Similar as above, but with luma showing frequency of chroma values.
18176 Set which graticule to display.
18180 Do not display graticule.
18183 Display green graticule showing legal broadcast ranges.
18186 Display orange graticule showing legal broadcast ranges.
18190 Set graticule opacity.
18193 Set graticule flags.
18197 Draw numbers above lines. By default enabled.
18200 Draw dots instead of lines.
18204 Set scale used for displaying graticule.
18211 Default is digital.
18214 Set background opacity.
18217 @section weave, doubleweave
18219 The @code{weave} takes a field-based video input and join
18220 each two sequential fields into single frame, producing a new double
18221 height clip with half the frame rate and half the frame count.
18223 The @code{doubleweave} works same as @code{weave} but without
18224 halving frame rate and frame count.
18226 It accepts the following option:
18230 Set first field. Available values are:
18234 Set the frame as top-field-first.
18237 Set the frame as bottom-field-first.
18241 @subsection Examples
18245 Interlace video using @ref{select} and @ref{separatefields} filter:
18247 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
18252 Apply the xBR high-quality magnification filter which is designed for pixel
18253 art. It follows a set of edge-detection rules, see
18254 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
18256 It accepts the following option:
18260 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
18261 @code{3xBR} and @code{4} for @code{4xBR}.
18262 Default is @code{3}.
18266 Stack video inputs into custom layout.
18268 All streams must be of same pixel format.
18270 The filter accept the following option:
18274 Set number of input streams. Default is 2.
18277 Specify layout of inputs.
18278 This option requires the desired layout configuration to be explicitly set by the user.
18279 This sets position of each video input in output. Each input
18280 is separated by '|'.
18281 The first number represents the column, and the second number represents the row.
18282 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
18283 where X is video input from which to take width or height.
18284 Multiple values can be used when separated by '+'. In such
18285 case values are summed together.
18288 If set to 1, force the output to terminate when the shortest input
18289 terminates. Default value is 0.
18292 @subsection Examples
18296 Display 4 inputs into 2x2 grid,
18297 note that if inputs are of different sizes unused gaps might appear,
18298 as not all of output video is used.
18300 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
18304 Display 4 inputs into 1x4 grid,
18305 note that if inputs are of different sizes unused gaps might appear,
18306 as not all of output video is used.
18308 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
18312 Display 9 inputs into 3x3 grid,
18313 note that if inputs are of different sizes unused gaps might appear,
18314 as not all of output video is used.
18316 xstack=inputs=9:layout=w3_0|w3_h0+h2|w3_h0|0_h4|0_0|w3+w1_0|0_h1+h2|w3+w1_h0|w3+w1_h1+h2
18323 Deinterlace the input video ("yadif" means "yet another deinterlacing
18326 It accepts the following parameters:
18332 The interlacing mode to adopt. It accepts one of the following values:
18335 @item 0, send_frame
18336 Output one frame for each frame.
18337 @item 1, send_field
18338 Output one frame for each field.
18339 @item 2, send_frame_nospatial
18340 Like @code{send_frame}, but it skips the spatial interlacing check.
18341 @item 3, send_field_nospatial
18342 Like @code{send_field}, but it skips the spatial interlacing check.
18345 The default value is @code{send_frame}.
18348 The picture field parity assumed for the input interlaced video. It accepts one
18349 of the following values:
18353 Assume the top field is first.
18355 Assume the bottom field is first.
18357 Enable automatic detection of field parity.
18360 The default value is @code{auto}.
18361 If the interlacing is unknown or the decoder does not export this information,
18362 top field first will be assumed.
18365 Specify which frames to deinterlace. Accept one of the following
18370 Deinterlace all frames.
18371 @item 1, interlaced
18372 Only deinterlace frames marked as interlaced.
18375 The default value is @code{all}.
18378 @section yadif_cuda
18380 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
18381 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
18384 It accepts the following parameters:
18390 The interlacing mode to adopt. It accepts one of the following values:
18393 @item 0, send_frame
18394 Output one frame for each frame.
18395 @item 1, send_field
18396 Output one frame for each field.
18397 @item 2, send_frame_nospatial
18398 Like @code{send_frame}, but it skips the spatial interlacing check.
18399 @item 3, send_field_nospatial
18400 Like @code{send_field}, but it skips the spatial interlacing check.
18403 The default value is @code{send_frame}.
18406 The picture field parity assumed for the input interlaced video. It accepts one
18407 of the following values:
18411 Assume the top field is first.
18413 Assume the bottom field is first.
18415 Enable automatic detection of field parity.
18418 The default value is @code{auto}.
18419 If the interlacing is unknown or the decoder does not export this information,
18420 top field first will be assumed.
18423 Specify which frames to deinterlace. Accept one of the following
18428 Deinterlace all frames.
18429 @item 1, interlaced
18430 Only deinterlace frames marked as interlaced.
18433 The default value is @code{all}.
18438 Apply Zoom & Pan effect.
18440 This filter accepts the following options:
18444 Set the zoom expression. Range is 1-10. Default is 1.
18448 Set the x and y expression. Default is 0.
18451 Set the duration expression in number of frames.
18452 This sets for how many number of frames effect will last for
18453 single input image.
18456 Set the output image size, default is 'hd720'.
18459 Set the output frame rate, default is '25'.
18462 Each expression can contain the following constants:
18481 Output frame count.
18485 Last calculated 'x' and 'y' position from 'x' and 'y' expression
18486 for current input frame.
18490 'x' and 'y' of last output frame of previous input frame or 0 when there was
18491 not yet such frame (first input frame).
18494 Last calculated zoom from 'z' expression for current input frame.
18497 Last calculated zoom of last output frame of previous input frame.
18500 Number of output frames for current input frame. Calculated from 'd' expression
18501 for each input frame.
18504 number of output frames created for previous input frame
18507 Rational number: input width / input height
18510 sample aspect ratio
18513 display aspect ratio
18517 @subsection Examples
18521 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
18523 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
18527 Zoom-in up to 1.5 and pan always at center of picture:
18529 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
18533 Same as above but without pausing:
18535 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
18541 Scale (resize) the input video, using the z.lib library:
18542 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
18543 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
18545 The zscale filter forces the output display aspect ratio to be the same
18546 as the input, by changing the output sample aspect ratio.
18548 If the input image format is different from the format requested by
18549 the next filter, the zscale filter will convert the input to the
18552 @subsection Options
18553 The filter accepts the following options.
18558 Set the output video dimension expression. Default value is the input
18561 If the @var{width} or @var{w} value is 0, the input width is used for
18562 the output. If the @var{height} or @var{h} value is 0, the input height
18563 is used for the output.
18565 If one and only one of the values is -n with n >= 1, the zscale filter
18566 will use a value that maintains the aspect ratio of the input image,
18567 calculated from the other specified dimension. After that it will,
18568 however, make sure that the calculated dimension is divisible by n and
18569 adjust the value if necessary.
18571 If both values are -n with n >= 1, the behavior will be identical to
18572 both values being set to 0 as previously detailed.
18574 See below for the list of accepted constants for use in the dimension
18578 Set the video size. For the syntax of this option, check the
18579 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18582 Set the dither type.
18584 Possible values are:
18589 @item error_diffusion
18595 Set the resize filter type.
18597 Possible values are:
18607 Default is bilinear.
18610 Set the color range.
18612 Possible values are:
18619 Default is same as input.
18622 Set the color primaries.
18624 Possible values are:
18634 Default is same as input.
18637 Set the transfer characteristics.
18639 Possible values are:
18653 Default is same as input.
18656 Set the colorspace matrix.
18658 Possible value are:
18669 Default is same as input.
18672 Set the input color range.
18674 Possible values are:
18681 Default is same as input.
18683 @item primariesin, pin
18684 Set the input color primaries.
18686 Possible values are:
18696 Default is same as input.
18698 @item transferin, tin
18699 Set the input transfer characteristics.
18701 Possible values are:
18712 Default is same as input.
18714 @item matrixin, min
18715 Set the input colorspace matrix.
18717 Possible value are:
18729 Set the output chroma location.
18731 Possible values are:
18742 @item chromalin, cin
18743 Set the input chroma location.
18745 Possible values are:
18757 Set the nominal peak luminance.
18760 The values of the @option{w} and @option{h} options are expressions
18761 containing the following constants:
18766 The input width and height
18770 These are the same as @var{in_w} and @var{in_h}.
18774 The output (scaled) width and height
18778 These are the same as @var{out_w} and @var{out_h}
18781 The same as @var{iw} / @var{ih}
18784 input sample aspect ratio
18787 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
18791 horizontal and vertical input chroma subsample values. For example for the
18792 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
18796 horizontal and vertical output chroma subsample values. For example for the
18797 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
18803 @c man end VIDEO FILTERS
18805 @chapter OpenCL Video Filters
18806 @c man begin OPENCL VIDEO FILTERS
18808 Below is a description of the currently available OpenCL video filters.
18810 To enable compilation of these filters you need to configure FFmpeg with
18811 @code{--enable-opencl}.
18813 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
18816 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
18817 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
18818 given device parameters.
18820 @item -filter_hw_device @var{name}
18821 Pass the hardware device called @var{name} to all filters in any filter graph.
18825 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
18829 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
18831 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
18835 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.
18837 @section avgblur_opencl
18839 Apply average blur filter.
18841 The filter accepts the following options:
18845 Set horizontal radius size.
18846 Range is @code{[1, 1024]} and default value is @code{1}.
18849 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
18852 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
18855 @subsection Example
18859 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.
18861 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
18865 @section boxblur_opencl
18867 Apply a boxblur algorithm to the input video.
18869 It accepts the following parameters:
18873 @item luma_radius, lr
18874 @item luma_power, lp
18875 @item chroma_radius, cr
18876 @item chroma_power, cp
18877 @item alpha_radius, ar
18878 @item alpha_power, ap
18882 A description of the accepted options follows.
18885 @item luma_radius, lr
18886 @item chroma_radius, cr
18887 @item alpha_radius, ar
18888 Set an expression for the box radius in pixels used for blurring the
18889 corresponding input plane.
18891 The radius value must be a non-negative number, and must not be
18892 greater than the value of the expression @code{min(w,h)/2} for the
18893 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
18896 Default value for @option{luma_radius} is "2". If not specified,
18897 @option{chroma_radius} and @option{alpha_radius} default to the
18898 corresponding value set for @option{luma_radius}.
18900 The expressions can contain the following constants:
18904 The input width and height in pixels.
18908 The input chroma image width and height in pixels.
18912 The horizontal and vertical chroma subsample values. For example, for the
18913 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
18916 @item luma_power, lp
18917 @item chroma_power, cp
18918 @item alpha_power, ap
18919 Specify how many times the boxblur filter is applied to the
18920 corresponding plane.
18922 Default value for @option{luma_power} is 2. If not specified,
18923 @option{chroma_power} and @option{alpha_power} default to the
18924 corresponding value set for @option{luma_power}.
18926 A value of 0 will disable the effect.
18929 @subsection Examples
18931 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.
18935 Apply a boxblur filter with the luma, chroma, and alpha radius
18936 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.
18938 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
18939 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
18943 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.
18945 For the luma plane, a 2x2 box radius will be run once.
18947 For the chroma plane, a 4x4 box radius will be run 5 times.
18949 For the alpha plane, a 3x3 box radius will be run 7 times.
18951 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
18955 @section convolution_opencl
18957 Apply convolution of 3x3, 5x5, 7x7 matrix.
18959 The filter accepts the following options:
18966 Set matrix for each plane.
18967 Matrix is sequence of 9, 25 or 49 signed numbers.
18968 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
18974 Set multiplier for calculated value for each plane.
18975 If unset or 0, it will be sum of all matrix elements.
18976 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
18982 Set bias for each plane. This value is added to the result of the multiplication.
18983 Useful for making the overall image brighter or darker.
18984 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
18988 @subsection Examples
18994 -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
19000 -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
19004 Apply edge enhance:
19006 -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
19012 -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
19016 Apply laplacian edge detector which includes diagonals:
19018 -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
19024 -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
19028 @section dilation_opencl
19030 Apply dilation effect to the video.
19032 This filter replaces the pixel by the local(3x3) maximum.
19034 It accepts the following options:
19041 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
19042 If @code{0}, plane will remain unchanged.
19045 Flag which specifies the pixel to refer to.
19046 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
19048 Flags to local 3x3 coordinates region centered on @code{x}:
19057 @subsection Example
19061 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.
19063 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
19067 @section erosion_opencl
19069 Apply erosion effect to the video.
19071 This filter replaces the pixel by the local(3x3) minimum.
19073 It accepts the following options:
19080 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
19081 If @code{0}, plane will remain unchanged.
19084 Flag which specifies the pixel to refer to.
19085 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
19087 Flags to local 3x3 coordinates region centered on @code{x}:
19096 @subsection Example
19100 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.
19102 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
19106 @section overlay_opencl
19108 Overlay one video on top of another.
19110 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
19111 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
19113 The filter accepts the following options:
19118 Set the x coordinate of the overlaid video on the main video.
19119 Default value is @code{0}.
19122 Set the x coordinate of the overlaid video on the main video.
19123 Default value is @code{0}.
19127 @subsection Examples
19131 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
19133 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
19136 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
19138 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
19143 @section prewitt_opencl
19145 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
19147 The filter accepts the following option:
19151 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19154 Set value which will be multiplied with filtered result.
19155 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
19158 Set value which will be added to filtered result.
19159 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
19162 @subsection Example
19166 Apply the Prewitt operator with scale set to 2 and delta set to 10.
19168 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
19172 @section roberts_opencl
19173 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
19175 The filter accepts the following option:
19179 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19182 Set value which will be multiplied with filtered result.
19183 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
19186 Set value which will be added to filtered result.
19187 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
19190 @subsection Example
19194 Apply the Roberts cross operator with scale set to 2 and delta set to 10
19196 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
19200 @section sobel_opencl
19202 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
19204 The filter accepts the following option:
19208 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19211 Set value which will be multiplied with filtered result.
19212 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
19215 Set value which will be added to filtered result.
19216 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
19219 @subsection Example
19223 Apply sobel operator with scale set to 2 and delta set to 10
19225 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
19229 @section tonemap_opencl
19231 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
19233 It accepts the following parameters:
19237 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
19240 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
19243 Apply desaturation for highlights that exceed this level of brightness. The
19244 higher the parameter, the more color information will be preserved. This
19245 setting helps prevent unnaturally blown-out colors for super-highlights, by
19246 (smoothly) turning into white instead. This makes images feel more natural,
19247 at the cost of reducing information about out-of-range colors.
19249 The default value is 0.5, and the algorithm here is a little different from
19250 the cpu version tonemap currently. A setting of 0.0 disables this option.
19253 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
19254 is used to detect whether the scene has changed or not. If the distance between
19255 the current frame average brightness and the current running average exceeds
19256 a threshold value, we would re-calculate scene average and peak brightness.
19257 The default value is 0.2.
19260 Specify the output pixel format.
19262 Currently supported formats are:
19269 Set the output color range.
19271 Possible values are:
19277 Default is same as input.
19280 Set the output color primaries.
19282 Possible values are:
19288 Default is same as input.
19291 Set the output transfer characteristics.
19293 Possible values are:
19302 Set the output colorspace matrix.
19304 Possible value are:
19310 Default is same as input.
19314 @subsection Example
19318 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
19320 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
19324 @section unsharp_opencl
19326 Sharpen or blur the input video.
19328 It accepts the following parameters:
19331 @item luma_msize_x, lx
19332 Set the luma matrix horizontal size.
19333 Range is @code{[1, 23]} and default value is @code{5}.
19335 @item luma_msize_y, ly
19336 Set the luma matrix vertical size.
19337 Range is @code{[1, 23]} and default value is @code{5}.
19339 @item luma_amount, la
19340 Set the luma effect strength.
19341 Range is @code{[-10, 10]} and default value is @code{1.0}.
19343 Negative values will blur the input video, while positive values will
19344 sharpen it, a value of zero will disable the effect.
19346 @item chroma_msize_x, cx
19347 Set the chroma matrix horizontal size.
19348 Range is @code{[1, 23]} and default value is @code{5}.
19350 @item chroma_msize_y, cy
19351 Set the chroma matrix vertical size.
19352 Range is @code{[1, 23]} and default value is @code{5}.
19354 @item chroma_amount, ca
19355 Set the chroma effect strength.
19356 Range is @code{[-10, 10]} and default value is @code{0.0}.
19358 Negative values will blur the input video, while positive values will
19359 sharpen it, a value of zero will disable the effect.
19363 All parameters are optional and default to the equivalent of the
19364 string '5:5:1.0:5:5:0.0'.
19366 @subsection Examples
19370 Apply strong luma sharpen effect:
19372 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
19376 Apply a strong blur of both luma and chroma parameters:
19378 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
19382 @c man end OPENCL VIDEO FILTERS
19384 @chapter Video Sources
19385 @c man begin VIDEO SOURCES
19387 Below is a description of the currently available video sources.
19391 Buffer video frames, and make them available to the filter chain.
19393 This source is mainly intended for a programmatic use, in particular
19394 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
19396 It accepts the following parameters:
19401 Specify the size (width and height) of the buffered video frames. For the
19402 syntax of this option, check the
19403 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19406 The input video width.
19409 The input video height.
19412 A string representing the pixel format of the buffered video frames.
19413 It may be a number corresponding to a pixel format, or a pixel format
19417 Specify the timebase assumed by the timestamps of the buffered frames.
19420 Specify the frame rate expected for the video stream.
19422 @item pixel_aspect, sar
19423 The sample (pixel) aspect ratio of the input video.
19426 Specify the optional parameters to be used for the scale filter which
19427 is automatically inserted when an input change is detected in the
19428 input size or format.
19430 @item hw_frames_ctx
19431 When using a hardware pixel format, this should be a reference to an
19432 AVHWFramesContext describing input frames.
19437 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
19440 will instruct the source to accept video frames with size 320x240 and
19441 with format "yuv410p", assuming 1/24 as the timestamps timebase and
19442 square pixels (1:1 sample aspect ratio).
19443 Since the pixel format with name "yuv410p" corresponds to the number 6
19444 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
19445 this example corresponds to:
19447 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
19450 Alternatively, the options can be specified as a flat string, but this
19451 syntax is deprecated:
19453 @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}]
19457 Create a pattern generated by an elementary cellular automaton.
19459 The initial state of the cellular automaton can be defined through the
19460 @option{filename} and @option{pattern} options. If such options are
19461 not specified an initial state is created randomly.
19463 At each new frame a new row in the video is filled with the result of
19464 the cellular automaton next generation. The behavior when the whole
19465 frame is filled is defined by the @option{scroll} option.
19467 This source accepts the following options:
19471 Read the initial cellular automaton state, i.e. the starting row, from
19472 the specified file.
19473 In the file, each non-whitespace character is considered an alive
19474 cell, a newline will terminate the row, and further characters in the
19475 file will be ignored.
19478 Read the initial cellular automaton state, i.e. the starting row, from
19479 the specified string.
19481 Each non-whitespace character in the string is considered an alive
19482 cell, a newline will terminate the row, and further characters in the
19483 string will be ignored.
19486 Set the video rate, that is the number of frames generated per second.
19489 @item random_fill_ratio, ratio
19490 Set the random fill ratio for the initial cellular automaton row. It
19491 is a floating point number value ranging from 0 to 1, defaults to
19494 This option is ignored when a file or a pattern is specified.
19496 @item random_seed, seed
19497 Set the seed for filling randomly the initial row, must be an integer
19498 included between 0 and UINT32_MAX. If not specified, or if explicitly
19499 set to -1, the filter will try to use a good random seed on a best
19503 Set the cellular automaton rule, it is a number ranging from 0 to 255.
19504 Default value is 110.
19507 Set the size of the output video. For the syntax of this option, check the
19508 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19510 If @option{filename} or @option{pattern} is specified, the size is set
19511 by default to the width of the specified initial state row, and the
19512 height is set to @var{width} * PHI.
19514 If @option{size} is set, it must contain the width of the specified
19515 pattern string, and the specified pattern will be centered in the
19518 If a filename or a pattern string is not specified, the size value
19519 defaults to "320x518" (used for a randomly generated initial state).
19522 If set to 1, scroll the output upward when all the rows in the output
19523 have been already filled. If set to 0, the new generated row will be
19524 written over the top row just after the bottom row is filled.
19527 @item start_full, full
19528 If set to 1, completely fill the output with generated rows before
19529 outputting the first frame.
19530 This is the default behavior, for disabling set the value to 0.
19533 If set to 1, stitch the left and right row edges together.
19534 This is the default behavior, for disabling set the value to 0.
19537 @subsection Examples
19541 Read the initial state from @file{pattern}, and specify an output of
19544 cellauto=f=pattern:s=200x400
19548 Generate a random initial row with a width of 200 cells, with a fill
19551 cellauto=ratio=2/3:s=200x200
19555 Create a pattern generated by rule 18 starting by a single alive cell
19556 centered on an initial row with width 100:
19558 cellauto=p=@@:s=100x400:full=0:rule=18
19562 Specify a more elaborated initial pattern:
19564 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
19569 @anchor{coreimagesrc}
19570 @section coreimagesrc
19571 Video source generated on GPU using Apple's CoreImage API on OSX.
19573 This video source is a specialized version of the @ref{coreimage} video filter.
19574 Use a core image generator at the beginning of the applied filterchain to
19575 generate the content.
19577 The coreimagesrc video source accepts the following options:
19579 @item list_generators
19580 List all available generators along with all their respective options as well as
19581 possible minimum and maximum values along with the default values.
19583 list_generators=true
19587 Specify the size of the sourced video. For the syntax of this option, check the
19588 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19589 The default value is @code{320x240}.
19592 Specify the frame rate of the sourced video, as the number of frames
19593 generated per second. It has to be a string in the format
19594 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
19595 number or a valid video frame rate abbreviation. The default value is
19599 Set the sample aspect ratio of the sourced video.
19602 Set the duration of the sourced video. See
19603 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19604 for the accepted syntax.
19606 If not specified, or the expressed duration is negative, the video is
19607 supposed to be generated forever.
19610 Additionally, all options of the @ref{coreimage} video filter are accepted.
19611 A complete filterchain can be used for further processing of the
19612 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
19613 and examples for details.
19615 @subsection Examples
19620 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
19621 given as complete and escaped command-line for Apple's standard bash shell:
19623 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
19625 This example is equivalent to the QRCode example of @ref{coreimage} without the
19626 need for a nullsrc video source.
19630 @section mandelbrot
19632 Generate a Mandelbrot set fractal, and progressively zoom towards the
19633 point specified with @var{start_x} and @var{start_y}.
19635 This source accepts the following options:
19640 Set the terminal pts value. Default value is 400.
19643 Set the terminal scale value.
19644 Must be a floating point value. Default value is 0.3.
19647 Set the inner coloring mode, that is the algorithm used to draw the
19648 Mandelbrot fractal internal region.
19650 It shall assume one of the following values:
19655 Show time until convergence.
19657 Set color based on point closest to the origin of the iterations.
19662 Default value is @var{mincol}.
19665 Set the bailout value. Default value is 10.0.
19668 Set the maximum of iterations performed by the rendering
19669 algorithm. Default value is 7189.
19672 Set outer coloring mode.
19673 It shall assume one of following values:
19675 @item iteration_count
19676 Set iteration count mode.
19677 @item normalized_iteration_count
19678 set normalized iteration count mode.
19680 Default value is @var{normalized_iteration_count}.
19683 Set frame rate, expressed as number of frames per second. Default
19687 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
19688 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
19691 Set the initial scale value. Default value is 3.0.
19694 Set the initial x position. Must be a floating point value between
19695 -100 and 100. Default value is -0.743643887037158704752191506114774.
19698 Set the initial y position. Must be a floating point value between
19699 -100 and 100. Default value is -0.131825904205311970493132056385139.
19704 Generate various test patterns, as generated by the MPlayer test filter.
19706 The size of the generated video is fixed, and is 256x256.
19707 This source is useful in particular for testing encoding features.
19709 This source accepts the following options:
19714 Specify the frame rate of the sourced video, as the number of frames
19715 generated per second. It has to be a string in the format
19716 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
19717 number or a valid video frame rate abbreviation. The default value is
19721 Set the duration of the sourced video. See
19722 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19723 for the accepted syntax.
19725 If not specified, or the expressed duration is negative, the video is
19726 supposed to be generated forever.
19730 Set the number or the name of the test to perform. Supported tests are:
19746 Default value is "all", which will cycle through the list of all tests.
19751 mptestsrc=t=dc_luma
19754 will generate a "dc_luma" test pattern.
19756 @section frei0r_src
19758 Provide a frei0r source.
19760 To enable compilation of this filter you need to install the frei0r
19761 header and configure FFmpeg with @code{--enable-frei0r}.
19763 This source accepts the following parameters:
19768 The size of the video to generate. For the syntax of this option, check the
19769 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19772 The framerate of the generated video. It may be a string of the form
19773 @var{num}/@var{den} or a frame rate abbreviation.
19776 The name to the frei0r source to load. For more information regarding frei0r and
19777 how to set the parameters, read the @ref{frei0r} section in the video filters
19780 @item filter_params
19781 A '|'-separated list of parameters to pass to the frei0r source.
19785 For example, to generate a frei0r partik0l source with size 200x200
19786 and frame rate 10 which is overlaid on the overlay filter main input:
19788 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
19793 Generate a life pattern.
19795 This source is based on a generalization of John Conway's life game.
19797 The sourced input represents a life grid, each pixel represents a cell
19798 which can be in one of two possible states, alive or dead. Every cell
19799 interacts with its eight neighbours, which are the cells that are
19800 horizontally, vertically, or diagonally adjacent.
19802 At each interaction the grid evolves according to the adopted rule,
19803 which specifies the number of neighbor alive cells which will make a
19804 cell stay alive or born. The @option{rule} option allows one to specify
19807 This source accepts the following options:
19811 Set the file from which to read the initial grid state. In the file,
19812 each non-whitespace character is considered an alive cell, and newline
19813 is used to delimit the end of each row.
19815 If this option is not specified, the initial grid is generated
19819 Set the video rate, that is the number of frames generated per second.
19822 @item random_fill_ratio, ratio
19823 Set the random fill ratio for the initial random grid. It is a
19824 floating point number value ranging from 0 to 1, defaults to 1/PHI.
19825 It is ignored when a file is specified.
19827 @item random_seed, seed
19828 Set the seed for filling the initial random grid, must be an integer
19829 included between 0 and UINT32_MAX. If not specified, or if explicitly
19830 set to -1, the filter will try to use a good random seed on a best
19836 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
19837 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
19838 @var{NS} specifies the number of alive neighbor cells which make a
19839 live cell stay alive, and @var{NB} the number of alive neighbor cells
19840 which make a dead cell to become alive (i.e. to "born").
19841 "s" and "b" can be used in place of "S" and "B", respectively.
19843 Alternatively a rule can be specified by an 18-bits integer. The 9
19844 high order bits are used to encode the next cell state if it is alive
19845 for each number of neighbor alive cells, the low order bits specify
19846 the rule for "borning" new cells. Higher order bits encode for an
19847 higher number of neighbor cells.
19848 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
19849 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
19851 Default value is "S23/B3", which is the original Conway's game of life
19852 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
19853 cells, and will born a new cell if there are three alive cells around
19857 Set the size of the output video. For the syntax of this option, check the
19858 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19860 If @option{filename} is specified, the size is set by default to the
19861 same size of the input file. If @option{size} is set, it must contain
19862 the size specified in the input file, and the initial grid defined in
19863 that file is centered in the larger resulting area.
19865 If a filename is not specified, the size value defaults to "320x240"
19866 (used for a randomly generated initial grid).
19869 If set to 1, stitch the left and right grid edges together, and the
19870 top and bottom edges also. Defaults to 1.
19873 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
19874 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
19875 value from 0 to 255.
19878 Set the color of living (or new born) cells.
19881 Set the color of dead cells. If @option{mold} is set, this is the first color
19882 used to represent a dead cell.
19885 Set mold color, for definitely dead and moldy cells.
19887 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
19888 ffmpeg-utils manual,ffmpeg-utils}.
19891 @subsection Examples
19895 Read a grid from @file{pattern}, and center it on a grid of size
19898 life=f=pattern:s=300x300
19902 Generate a random grid of size 200x200, with a fill ratio of 2/3:
19904 life=ratio=2/3:s=200x200
19908 Specify a custom rule for evolving a randomly generated grid:
19914 Full example with slow death effect (mold) using @command{ffplay}:
19916 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
19923 @anchor{haldclutsrc}
19926 @anchor{pal100bars}
19927 @anchor{rgbtestsrc}
19929 @anchor{smptehdbars}
19932 @anchor{yuvtestsrc}
19933 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
19935 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
19937 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
19939 The @code{color} source provides an uniformly colored input.
19941 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
19942 @ref{haldclut} filter.
19944 The @code{nullsrc} source returns unprocessed video frames. It is
19945 mainly useful to be employed in analysis / debugging tools, or as the
19946 source for filters which ignore the input data.
19948 The @code{pal75bars} source generates a color bars pattern, based on
19949 EBU PAL recommendations with 75% color levels.
19951 The @code{pal100bars} source generates a color bars pattern, based on
19952 EBU PAL recommendations with 100% color levels.
19954 The @code{rgbtestsrc} source generates an RGB test pattern useful for
19955 detecting RGB vs BGR issues. You should see a red, green and blue
19956 stripe from top to bottom.
19958 The @code{smptebars} source generates a color bars pattern, based on
19959 the SMPTE Engineering Guideline EG 1-1990.
19961 The @code{smptehdbars} source generates a color bars pattern, based on
19962 the SMPTE RP 219-2002.
19964 The @code{testsrc} source generates a test video pattern, showing a
19965 color pattern, a scrolling gradient and a timestamp. This is mainly
19966 intended for testing purposes.
19968 The @code{testsrc2} source is similar to testsrc, but supports more
19969 pixel formats instead of just @code{rgb24}. This allows using it as an
19970 input for other tests without requiring a format conversion.
19972 The @code{yuvtestsrc} source generates an YUV test pattern. You should
19973 see a y, cb and cr stripe from top to bottom.
19975 The sources accept the following parameters:
19980 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
19981 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
19982 pixels to be used as identity matrix for 3D lookup tables. Each component is
19983 coded on a @code{1/(N*N)} scale.
19986 Specify the color of the source, only available in the @code{color}
19987 source. For the syntax of this option, check the
19988 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19991 Specify the size of the sourced video. For the syntax of this option, check the
19992 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19993 The default value is @code{320x240}.
19995 This option is not available with the @code{allrgb}, @code{allyuv}, and
19996 @code{haldclutsrc} filters.
19999 Specify the frame rate of the sourced video, as the number of frames
20000 generated per second. It has to be a string in the format
20001 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20002 number or a valid video frame rate abbreviation. The default value is
20006 Set the duration of the sourced video. See
20007 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20008 for the accepted syntax.
20010 If not specified, or the expressed duration is negative, the video is
20011 supposed to be generated forever.
20014 Set the sample aspect ratio of the sourced video.
20017 Specify the alpha (opacity) of the background, only available in the
20018 @code{testsrc2} source. The value must be between 0 (fully transparent) and
20019 255 (fully opaque, the default).
20022 Set the number of decimals to show in the timestamp, only available in the
20023 @code{testsrc} source.
20025 The displayed timestamp value will correspond to the original
20026 timestamp value multiplied by the power of 10 of the specified
20027 value. Default value is 0.
20030 @subsection Examples
20034 Generate a video with a duration of 5.3 seconds, with size
20035 176x144 and a frame rate of 10 frames per second:
20037 testsrc=duration=5.3:size=qcif:rate=10
20041 The following graph description will generate a red source
20042 with an opacity of 0.2, with size "qcif" and a frame rate of 10
20045 color=c=red@@0.2:s=qcif:r=10
20049 If the input content is to be ignored, @code{nullsrc} can be used. The
20050 following command generates noise in the luminance plane by employing
20051 the @code{geq} filter:
20053 nullsrc=s=256x256, geq=random(1)*255:128:128
20057 @subsection Commands
20059 The @code{color} source supports the following commands:
20063 Set the color of the created image. Accepts the same syntax of the
20064 corresponding @option{color} option.
20069 Generate video using an OpenCL program.
20074 OpenCL program source file.
20077 Kernel name in program.
20080 Size of frames to generate. This must be set.
20083 Pixel format to use for the generated frames. This must be set.
20086 Number of frames generated every second. Default value is '25'.
20090 For details of how the program loading works, see the @ref{program_opencl}
20097 Generate a colour ramp by setting pixel values from the position of the pixel
20098 in the output image. (Note that this will work with all pixel formats, but
20099 the generated output will not be the same.)
20101 __kernel void ramp(__write_only image2d_t dst,
20102 unsigned int index)
20104 int2 loc = (int2)(get_global_id(0), get_global_id(1));
20107 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
20109 write_imagef(dst, loc, val);
20114 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
20116 __kernel void sierpinski_carpet(__write_only image2d_t dst,
20117 unsigned int index)
20119 int2 loc = (int2)(get_global_id(0), get_global_id(1));
20121 float4 value = 0.0f;
20122 int x = loc.x + index;
20123 int y = loc.y + index;
20124 while (x > 0 || y > 0) {
20125 if (x % 3 == 1 && y % 3 == 1) {
20133 write_imagef(dst, loc, value);
20139 @c man end VIDEO SOURCES
20141 @chapter Video Sinks
20142 @c man begin VIDEO SINKS
20144 Below is a description of the currently available video sinks.
20146 @section buffersink
20148 Buffer video frames, and make them available to the end of the filter
20151 This sink is mainly intended for programmatic use, in particular
20152 through the interface defined in @file{libavfilter/buffersink.h}
20153 or the options system.
20155 It accepts a pointer to an AVBufferSinkContext structure, which
20156 defines the incoming buffers' formats, to be passed as the opaque
20157 parameter to @code{avfilter_init_filter} for initialization.
20161 Null video sink: do absolutely nothing with the input video. It is
20162 mainly useful as a template and for use in analysis / debugging
20165 @c man end VIDEO SINKS
20167 @chapter Multimedia Filters
20168 @c man begin MULTIMEDIA FILTERS
20170 Below is a description of the currently available multimedia filters.
20174 Convert input audio to a video output, displaying the audio bit scope.
20176 The filter accepts the following options:
20180 Set frame rate, expressed as number of frames per second. Default
20184 Specify the video size for the output. For the syntax of this option, check the
20185 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20186 Default value is @code{1024x256}.
20189 Specify list of colors separated by space or by '|' which will be used to
20190 draw channels. Unrecognized or missing colors will be replaced
20194 @section ahistogram
20196 Convert input audio to a video output, displaying the volume histogram.
20198 The filter accepts the following options:
20202 Specify how histogram is calculated.
20204 It accepts the following values:
20207 Use single histogram for all channels.
20209 Use separate histogram for each channel.
20211 Default is @code{single}.
20214 Set frame rate, expressed as number of frames per second. Default
20218 Specify the video size for the output. For the syntax of this option, check the
20219 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20220 Default value is @code{hd720}.
20225 It accepts the following values:
20236 reverse logarithmic
20238 Default is @code{log}.
20241 Set amplitude scale.
20243 It accepts the following values:
20250 Default is @code{log}.
20253 Set how much frames to accumulate in histogram.
20254 Default is 1. Setting this to -1 accumulates all frames.
20257 Set histogram ratio of window height.
20260 Set sonogram sliding.
20262 It accepts the following values:
20265 replace old rows with new ones.
20267 scroll from top to bottom.
20269 Default is @code{replace}.
20272 @section aphasemeter
20274 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
20275 representing mean phase of current audio frame. A video output can also be produced and is
20276 enabled by default. The audio is passed through as first output.
20278 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
20279 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
20280 and @code{1} means channels are in phase.
20282 The filter accepts the following options, all related to its video output:
20286 Set the output frame rate. Default value is @code{25}.
20289 Set the video size for the output. For the syntax of this option, check the
20290 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20291 Default value is @code{800x400}.
20296 Specify the red, green, blue contrast. Default values are @code{2},
20297 @code{7} and @code{1}.
20298 Allowed range is @code{[0, 255]}.
20301 Set color which will be used for drawing median phase. If color is
20302 @code{none} which is default, no median phase value will be drawn.
20305 Enable video output. Default is enabled.
20308 @section avectorscope
20310 Convert input audio to a video output, representing the audio vector
20313 The filter is used to measure the difference between channels of stereo
20314 audio stream. A monoaural signal, consisting of identical left and right
20315 signal, results in straight vertical line. Any stereo separation is visible
20316 as a deviation from this line, creating a Lissajous figure.
20317 If the straight (or deviation from it) but horizontal line appears this
20318 indicates that the left and right channels are out of phase.
20320 The filter accepts the following options:
20324 Set the vectorscope mode.
20326 Available values are:
20329 Lissajous rotated by 45 degrees.
20332 Same as above but not rotated.
20335 Shape resembling half of circle.
20338 Default value is @samp{lissajous}.
20341 Set the video size for the output. For the syntax of this option, check the
20342 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20343 Default value is @code{400x400}.
20346 Set the output frame rate. Default value is @code{25}.
20352 Specify the red, green, blue and alpha contrast. Default values are @code{40},
20353 @code{160}, @code{80} and @code{255}.
20354 Allowed range is @code{[0, 255]}.
20360 Specify the red, green, blue and alpha fade. Default values are @code{15},
20361 @code{10}, @code{5} and @code{5}.
20362 Allowed range is @code{[0, 255]}.
20365 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
20366 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
20369 Set the vectorscope drawing mode.
20371 Available values are:
20374 Draw dot for each sample.
20377 Draw line between previous and current sample.
20380 Default value is @samp{dot}.
20383 Specify amplitude scale of audio samples.
20385 Available values are:
20401 Swap left channel axis with right channel axis.
20411 Mirror only x axis.
20414 Mirror only y axis.
20422 @subsection Examples
20426 Complete example using @command{ffplay}:
20428 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
20429 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
20433 @section bench, abench
20435 Benchmark part of a filtergraph.
20437 The filter accepts the following options:
20441 Start or stop a timer.
20443 Available values are:
20446 Get the current time, set it as frame metadata (using the key
20447 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
20450 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
20451 the input frame metadata to get the time difference. Time difference, average,
20452 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
20453 @code{min}) are then printed. The timestamps are expressed in seconds.
20457 @subsection Examples
20461 Benchmark @ref{selectivecolor} filter:
20463 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
20469 Concatenate audio and video streams, joining them together one after the
20472 The filter works on segments of synchronized video and audio streams. All
20473 segments must have the same number of streams of each type, and that will
20474 also be the number of streams at output.
20476 The filter accepts the following options:
20481 Set the number of segments. Default is 2.
20484 Set the number of output video streams, that is also the number of video
20485 streams in each segment. Default is 1.
20488 Set the number of output audio streams, that is also the number of audio
20489 streams in each segment. Default is 0.
20492 Activate unsafe mode: do not fail if segments have a different format.
20496 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
20497 @var{a} audio outputs.
20499 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
20500 segment, in the same order as the outputs, then the inputs for the second
20503 Related streams do not always have exactly the same duration, for various
20504 reasons including codec frame size or sloppy authoring. For that reason,
20505 related synchronized streams (e.g. a video and its audio track) should be
20506 concatenated at once. The concat filter will use the duration of the longest
20507 stream in each segment (except the last one), and if necessary pad shorter
20508 audio streams with silence.
20510 For this filter to work correctly, all segments must start at timestamp 0.
20512 All corresponding streams must have the same parameters in all segments; the
20513 filtering system will automatically select a common pixel format for video
20514 streams, and a common sample format, sample rate and channel layout for
20515 audio streams, but other settings, such as resolution, must be converted
20516 explicitly by the user.
20518 Different frame rates are acceptable but will result in variable frame rate
20519 at output; be sure to configure the output file to handle it.
20521 @subsection Examples
20525 Concatenate an opening, an episode and an ending, all in bilingual version
20526 (video in stream 0, audio in streams 1 and 2):
20528 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
20529 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
20530 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
20531 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
20535 Concatenate two parts, handling audio and video separately, using the
20536 (a)movie sources, and adjusting the resolution:
20538 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
20539 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
20540 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
20542 Note that a desync will happen at the stitch if the audio and video streams
20543 do not have exactly the same duration in the first file.
20547 @subsection Commands
20549 This filter supports the following commands:
20552 Close the current segment and step to the next one
20555 @section drawgraph, adrawgraph
20557 Draw a graph using input video or audio metadata.
20559 It accepts the following parameters:
20563 Set 1st frame metadata key from which metadata values will be used to draw a graph.
20566 Set 1st foreground color expression.
20569 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
20572 Set 2nd foreground color expression.
20575 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
20578 Set 3rd foreground color expression.
20581 Set 4th frame metadata key from which metadata values will be used to draw a graph.
20584 Set 4th foreground color expression.
20587 Set minimal value of metadata value.
20590 Set maximal value of metadata value.
20593 Set graph background color. Default is white.
20598 Available values for mode is:
20605 Default is @code{line}.
20610 Available values for slide is:
20613 Draw new frame when right border is reached.
20616 Replace old columns with new ones.
20619 Scroll from right to left.
20622 Scroll from left to right.
20625 Draw single picture.
20628 Default is @code{frame}.
20631 Set size of graph video. For the syntax of this option, check the
20632 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20633 The default value is @code{900x256}.
20635 The foreground color expressions can use the following variables:
20638 Minimal value of metadata value.
20641 Maximal value of metadata value.
20644 Current metadata key value.
20647 The color is defined as 0xAABBGGRR.
20650 Example using metadata from @ref{signalstats} filter:
20652 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
20655 Example using metadata from @ref{ebur128} filter:
20657 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
20663 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
20664 level. By default, it logs a message at a frequency of 10Hz with the
20665 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
20666 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
20668 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
20669 sample format is double-precision floating point. The input stream will be converted to
20670 this specification, if needed. Users may need to insert aformat and/or aresample filters
20671 after this filter to obtain the original parameters.
20673 The filter also has a video output (see the @var{video} option) with a real
20674 time graph to observe the loudness evolution. The graphic contains the logged
20675 message mentioned above, so it is not printed anymore when this option is set,
20676 unless the verbose logging is set. The main graphing area contains the
20677 short-term loudness (3 seconds of analysis), and the gauge on the right is for
20678 the momentary loudness (400 milliseconds), but can optionally be configured
20679 to instead display short-term loudness (see @var{gauge}).
20681 The green area marks a +/- 1LU target range around the target loudness
20682 (-23LUFS by default, unless modified through @var{target}).
20684 More information about the Loudness Recommendation EBU R128 on
20685 @url{http://tech.ebu.ch/loudness}.
20687 The filter accepts the following options:
20692 Activate the video output. The audio stream is passed unchanged whether this
20693 option is set or no. The video stream will be the first output stream if
20694 activated. Default is @code{0}.
20697 Set the video size. This option is for video only. For the syntax of this
20699 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20700 Default and minimum resolution is @code{640x480}.
20703 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
20704 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
20705 other integer value between this range is allowed.
20708 Set metadata injection. If set to @code{1}, the audio input will be segmented
20709 into 100ms output frames, each of them containing various loudness information
20710 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
20712 Default is @code{0}.
20715 Force the frame logging level.
20717 Available values are:
20720 information logging level
20722 verbose logging level
20725 By default, the logging level is set to @var{info}. If the @option{video} or
20726 the @option{metadata} options are set, it switches to @var{verbose}.
20731 Available modes can be cumulated (the option is a @code{flag} type). Possible
20735 Disable any peak mode (default).
20737 Enable sample-peak mode.
20739 Simple peak mode looking for the higher sample value. It logs a message
20740 for sample-peak (identified by @code{SPK}).
20742 Enable true-peak mode.
20744 If enabled, the peak lookup is done on an over-sampled version of the input
20745 stream for better peak accuracy. It logs a message for true-peak.
20746 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
20747 This mode requires a build with @code{libswresample}.
20751 Treat mono input files as "dual mono". If a mono file is intended for playback
20752 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
20753 If set to @code{true}, this option will compensate for this effect.
20754 Multi-channel input files are not affected by this option.
20757 Set a specific pan law to be used for the measurement of dual mono files.
20758 This parameter is optional, and has a default value of -3.01dB.
20761 Set a specific target level (in LUFS) used as relative zero in the visualization.
20762 This parameter is optional and has a default value of -23LUFS as specified
20763 by EBU R128. However, material published online may prefer a level of -16LUFS
20764 (e.g. for use with podcasts or video platforms).
20767 Set the value displayed by the gauge. Valid values are @code{momentary} and s
20768 @code{shortterm}. By default the momentary value will be used, but in certain
20769 scenarios it may be more useful to observe the short term value instead (e.g.
20773 Sets the display scale for the loudness. Valid parameters are @code{absolute}
20774 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
20775 video output, not the summary or continuous log output.
20778 @subsection Examples
20782 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
20784 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
20788 Run an analysis with @command{ffmpeg}:
20790 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
20794 @section interleave, ainterleave
20796 Temporally interleave frames from several inputs.
20798 @code{interleave} works with video inputs, @code{ainterleave} with audio.
20800 These filters read frames from several inputs and send the oldest
20801 queued frame to the output.
20803 Input streams must have well defined, monotonically increasing frame
20806 In order to submit one frame to output, these filters need to enqueue
20807 at least one frame for each input, so they cannot work in case one
20808 input is not yet terminated and will not receive incoming frames.
20810 For example consider the case when one input is a @code{select} filter
20811 which always drops input frames. The @code{interleave} filter will keep
20812 reading from that input, but it will never be able to send new frames
20813 to output until the input sends an end-of-stream signal.
20815 Also, depending on inputs synchronization, the filters will drop
20816 frames in case one input receives more frames than the other ones, and
20817 the queue is already filled.
20819 These filters accept the following options:
20823 Set the number of different inputs, it is 2 by default.
20826 @subsection Examples
20830 Interleave frames belonging to different streams using @command{ffmpeg}:
20832 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
20836 Add flickering blur effect:
20838 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
20842 @section metadata, ametadata
20844 Manipulate frame metadata.
20846 This filter accepts the following options:
20850 Set mode of operation of the filter.
20852 Can be one of the following:
20856 If both @code{value} and @code{key} is set, select frames
20857 which have such metadata. If only @code{key} is set, select
20858 every frame that has such key in metadata.
20861 Add new metadata @code{key} and @code{value}. If key is already available
20865 Modify value of already present key.
20868 If @code{value} is set, delete only keys that have such value.
20869 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
20873 Print key and its value if metadata was found. If @code{key} is not set print all
20874 metadata values available in frame.
20878 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
20881 Set metadata value which will be used. This option is mandatory for
20882 @code{modify} and @code{add} mode.
20885 Which function to use when comparing metadata value and @code{value}.
20887 Can be one of following:
20891 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
20894 Values are interpreted as strings, returns true if metadata value starts with
20895 the @code{value} option string.
20898 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
20901 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
20904 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
20907 Values are interpreted as floats, returns true if expression from option @code{expr}
20912 Set expression which is used when @code{function} is set to @code{expr}.
20913 The expression is evaluated through the eval API and can contain the following
20918 Float representation of @code{value} from metadata key.
20921 Float representation of @code{value} as supplied by user in @code{value} option.
20925 If specified in @code{print} mode, output is written to the named file. Instead of
20926 plain filename any writable url can be specified. Filename ``-'' is a shorthand
20927 for standard output. If @code{file} option is not set, output is written to the log
20928 with AV_LOG_INFO loglevel.
20932 @subsection Examples
20936 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
20939 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
20942 Print silencedetect output to file @file{metadata.txt}.
20944 silencedetect,ametadata=mode=print:file=metadata.txt
20947 Direct all metadata to a pipe with file descriptor 4.
20949 metadata=mode=print:file='pipe\:4'
20953 @section perms, aperms
20955 Set read/write permissions for the output frames.
20957 These filters are mainly aimed at developers to test direct path in the
20958 following filter in the filtergraph.
20960 The filters accept the following options:
20964 Select the permissions mode.
20966 It accepts the following values:
20969 Do nothing. This is the default.
20971 Set all the output frames read-only.
20973 Set all the output frames directly writable.
20975 Make the frame read-only if writable, and writable if read-only.
20977 Set each output frame read-only or writable randomly.
20981 Set the seed for the @var{random} mode, must be an integer included between
20982 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
20983 @code{-1}, the filter will try to use a good random seed on a best effort
20987 Note: in case of auto-inserted filter between the permission filter and the
20988 following one, the permission might not be received as expected in that
20989 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
20990 perms/aperms filter can avoid this problem.
20992 @section realtime, arealtime
20994 Slow down filtering to match real time approximately.
20996 These filters will pause the filtering for a variable amount of time to
20997 match the output rate with the input timestamps.
20998 They are similar to the @option{re} option to @code{ffmpeg}.
21000 They accept the following options:
21004 Time limit for the pauses. Any pause longer than that will be considered
21005 a timestamp discontinuity and reset the timer. Default is 2 seconds.
21009 @section select, aselect
21011 Select frames to pass in output.
21013 This filter accepts the following options:
21018 Set expression, which is evaluated for each input frame.
21020 If the expression is evaluated to zero, the frame is discarded.
21022 If the evaluation result is negative or NaN, the frame is sent to the
21023 first output; otherwise it is sent to the output with index
21024 @code{ceil(val)-1}, assuming that the input index starts from 0.
21026 For example a value of @code{1.2} corresponds to the output with index
21027 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
21030 Set the number of outputs. The output to which to send the selected
21031 frame is based on the result of the evaluation. Default value is 1.
21034 The expression can contain the following constants:
21038 The (sequential) number of the filtered frame, starting from 0.
21041 The (sequential) number of the selected frame, starting from 0.
21043 @item prev_selected_n
21044 The sequential number of the last selected frame. It's NAN if undefined.
21047 The timebase of the input timestamps.
21050 The PTS (Presentation TimeStamp) of the filtered video frame,
21051 expressed in @var{TB} units. It's NAN if undefined.
21054 The PTS of the filtered video frame,
21055 expressed in seconds. It's NAN if undefined.
21058 The PTS of the previously filtered video frame. It's NAN if undefined.
21060 @item prev_selected_pts
21061 The PTS of the last previously filtered video frame. It's NAN if undefined.
21063 @item prev_selected_t
21064 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
21067 The PTS of the first video frame in the video. It's NAN if undefined.
21070 The time of the first video frame in the video. It's NAN if undefined.
21072 @item pict_type @emph{(video only)}
21073 The type of the filtered frame. It can assume one of the following
21085 @item interlace_type @emph{(video only)}
21086 The frame interlace type. It can assume one of the following values:
21089 The frame is progressive (not interlaced).
21091 The frame is top-field-first.
21093 The frame is bottom-field-first.
21096 @item consumed_sample_n @emph{(audio only)}
21097 the number of selected samples before the current frame
21099 @item samples_n @emph{(audio only)}
21100 the number of samples in the current frame
21102 @item sample_rate @emph{(audio only)}
21103 the input sample rate
21106 This is 1 if the filtered frame is a key-frame, 0 otherwise.
21109 the position in the file of the filtered frame, -1 if the information
21110 is not available (e.g. for synthetic video)
21112 @item scene @emph{(video only)}
21113 value between 0 and 1 to indicate a new scene; a low value reflects a low
21114 probability for the current frame to introduce a new scene, while a higher
21115 value means the current frame is more likely to be one (see the example below)
21117 @item concatdec_select
21118 The concat demuxer can select only part of a concat input file by setting an
21119 inpoint and an outpoint, but the output packets may not be entirely contained
21120 in the selected interval. By using this variable, it is possible to skip frames
21121 generated by the concat demuxer which are not exactly contained in the selected
21124 This works by comparing the frame pts against the @var{lavf.concat.start_time}
21125 and the @var{lavf.concat.duration} packet metadata values which are also
21126 present in the decoded frames.
21128 The @var{concatdec_select} variable is -1 if the frame pts is at least
21129 start_time and either the duration metadata is missing or the frame pts is less
21130 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
21133 That basically means that an input frame is selected if its pts is within the
21134 interval set by the concat demuxer.
21138 The default value of the select expression is "1".
21140 @subsection Examples
21144 Select all frames in input:
21149 The example above is the same as:
21161 Select only I-frames:
21163 select='eq(pict_type\,I)'
21167 Select one frame every 100:
21169 select='not(mod(n\,100))'
21173 Select only frames contained in the 10-20 time interval:
21175 select=between(t\,10\,20)
21179 Select only I-frames contained in the 10-20 time interval:
21181 select=between(t\,10\,20)*eq(pict_type\,I)
21185 Select frames with a minimum distance of 10 seconds:
21187 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
21191 Use aselect to select only audio frames with samples number > 100:
21193 aselect='gt(samples_n\,100)'
21197 Create a mosaic of the first scenes:
21199 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
21202 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
21206 Send even and odd frames to separate outputs, and compose them:
21208 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
21212 Select useful frames from an ffconcat file which is using inpoints and
21213 outpoints but where the source files are not intra frame only.
21215 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
21219 @section sendcmd, asendcmd
21221 Send commands to filters in the filtergraph.
21223 These filters read commands to be sent to other filters in the
21226 @code{sendcmd} must be inserted between two video filters,
21227 @code{asendcmd} must be inserted between two audio filters, but apart
21228 from that they act the same way.
21230 The specification of commands can be provided in the filter arguments
21231 with the @var{commands} option, or in a file specified by the
21232 @var{filename} option.
21234 These filters accept the following options:
21237 Set the commands to be read and sent to the other filters.
21239 Set the filename of the commands to be read and sent to the other
21243 @subsection Commands syntax
21245 A commands description consists of a sequence of interval
21246 specifications, comprising a list of commands to be executed when a
21247 particular event related to that interval occurs. The occurring event
21248 is typically the current frame time entering or leaving a given time
21251 An interval is specified by the following syntax:
21253 @var{START}[-@var{END}] @var{COMMANDS};
21256 The time interval is specified by the @var{START} and @var{END} times.
21257 @var{END} is optional and defaults to the maximum time.
21259 The current frame time is considered within the specified interval if
21260 it is included in the interval [@var{START}, @var{END}), that is when
21261 the time is greater or equal to @var{START} and is lesser than
21264 @var{COMMANDS} consists of a sequence of one or more command
21265 specifications, separated by ",", relating to that interval. The
21266 syntax of a command specification is given by:
21268 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
21271 @var{FLAGS} is optional and specifies the type of events relating to
21272 the time interval which enable sending the specified command, and must
21273 be a non-null sequence of identifier flags separated by "+" or "|" and
21274 enclosed between "[" and "]".
21276 The following flags are recognized:
21279 The command is sent when the current frame timestamp enters the
21280 specified interval. In other words, the command is sent when the
21281 previous frame timestamp was not in the given interval, and the
21285 The command is sent when the current frame timestamp leaves the
21286 specified interval. In other words, the command is sent when the
21287 previous frame timestamp was in the given interval, and the
21291 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
21294 @var{TARGET} specifies the target of the command, usually the name of
21295 the filter class or a specific filter instance name.
21297 @var{COMMAND} specifies the name of the command for the target filter.
21299 @var{ARG} is optional and specifies the optional list of argument for
21300 the given @var{COMMAND}.
21302 Between one interval specification and another, whitespaces, or
21303 sequences of characters starting with @code{#} until the end of line,
21304 are ignored and can be used to annotate comments.
21306 A simplified BNF description of the commands specification syntax
21309 @var{COMMAND_FLAG} ::= "enter" | "leave"
21310 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
21311 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
21312 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
21313 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
21314 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
21317 @subsection Examples
21321 Specify audio tempo change at second 4:
21323 asendcmd=c='4.0 atempo tempo 1.5',atempo
21327 Target a specific filter instance:
21329 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
21333 Specify a list of drawtext and hue commands in a file.
21335 # show text in the interval 5-10
21336 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
21337 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
21339 # desaturate the image in the interval 15-20
21340 15.0-20.0 [enter] hue s 0,
21341 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
21343 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
21345 # apply an exponential saturation fade-out effect, starting from time 25
21346 25 [enter] hue s exp(25-t)
21349 A filtergraph allowing to read and process the above command list
21350 stored in a file @file{test.cmd}, can be specified with:
21352 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
21357 @section setpts, asetpts
21359 Change the PTS (presentation timestamp) of the input frames.
21361 @code{setpts} works on video frames, @code{asetpts} on audio frames.
21363 This filter accepts the following options:
21368 The expression which is evaluated for each frame to construct its timestamp.
21372 The expression is evaluated through the eval API and can contain the following
21376 @item FRAME_RATE, FR
21377 frame rate, only defined for constant frame-rate video
21380 The presentation timestamp in input
21383 The count of the input frame for video or the number of consumed samples,
21384 not including the current frame for audio, starting from 0.
21386 @item NB_CONSUMED_SAMPLES
21387 The number of consumed samples, not including the current frame (only
21390 @item NB_SAMPLES, S
21391 The number of samples in the current frame (only audio)
21393 @item SAMPLE_RATE, SR
21394 The audio sample rate.
21397 The PTS of the first frame.
21400 the time in seconds of the first frame
21403 State whether the current frame is interlaced.
21406 the time in seconds of the current frame
21409 original position in the file of the frame, or undefined if undefined
21410 for the current frame
21413 The previous input PTS.
21416 previous input time in seconds
21419 The previous output PTS.
21422 previous output time in seconds
21425 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
21429 The wallclock (RTC) time at the start of the movie in microseconds.
21432 The timebase of the input timestamps.
21436 @subsection Examples
21440 Start counting PTS from zero
21442 setpts=PTS-STARTPTS
21446 Apply fast motion effect:
21452 Apply slow motion effect:
21458 Set fixed rate of 25 frames per second:
21464 Set fixed rate 25 fps with some jitter:
21466 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
21470 Apply an offset of 10 seconds to the input PTS:
21476 Generate timestamps from a "live source" and rebase onto the current timebase:
21478 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
21482 Generate timestamps by counting samples:
21491 Force color range for the output video frame.
21493 The @code{setrange} filter marks the color range property for the
21494 output frames. It does not change the input frame, but only sets the
21495 corresponding property, which affects how the frame is treated by
21498 The filter accepts the following options:
21503 Available values are:
21507 Keep the same color range property.
21509 @item unspecified, unknown
21510 Set the color range as unspecified.
21512 @item limited, tv, mpeg
21513 Set the color range as limited.
21515 @item full, pc, jpeg
21516 Set the color range as full.
21520 @section settb, asettb
21522 Set the timebase to use for the output frames timestamps.
21523 It is mainly useful for testing timebase configuration.
21525 It accepts the following parameters:
21530 The expression which is evaluated into the output timebase.
21534 The value for @option{tb} is an arithmetic expression representing a
21535 rational. The expression can contain the constants "AVTB" (the default
21536 timebase), "intb" (the input timebase) and "sr" (the sample rate,
21537 audio only). Default value is "intb".
21539 @subsection Examples
21543 Set the timebase to 1/25:
21549 Set the timebase to 1/10:
21555 Set the timebase to 1001/1000:
21561 Set the timebase to 2*intb:
21567 Set the default timebase value:
21574 Convert input audio to a video output representing frequency spectrum
21575 logarithmically using Brown-Puckette constant Q transform algorithm with
21576 direct frequency domain coefficient calculation (but the transform itself
21577 is not really constant Q, instead the Q factor is actually variable/clamped),
21578 with musical tone scale, from E0 to D#10.
21580 The filter accepts the following options:
21584 Specify the video size for the output. It must be even. For the syntax of this option,
21585 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21586 Default value is @code{1920x1080}.
21589 Set the output frame rate. Default value is @code{25}.
21592 Set the bargraph height. It must be even. Default value is @code{-1} which
21593 computes the bargraph height automatically.
21596 Set the axis height. It must be even. Default value is @code{-1} which computes
21597 the axis height automatically.
21600 Set the sonogram height. It must be even. Default value is @code{-1} which
21601 computes the sonogram height automatically.
21604 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
21605 instead. Default value is @code{1}.
21607 @item sono_v, volume
21608 Specify the sonogram volume expression. It can contain variables:
21611 the @var{bar_v} evaluated expression
21612 @item frequency, freq, f
21613 the frequency where it is evaluated
21614 @item timeclamp, tc
21615 the value of @var{timeclamp} option
21619 @item a_weighting(f)
21620 A-weighting of equal loudness
21621 @item b_weighting(f)
21622 B-weighting of equal loudness
21623 @item c_weighting(f)
21624 C-weighting of equal loudness.
21626 Default value is @code{16}.
21628 @item bar_v, volume2
21629 Specify the bargraph volume expression. It can contain variables:
21632 the @var{sono_v} evaluated expression
21633 @item frequency, freq, f
21634 the frequency where it is evaluated
21635 @item timeclamp, tc
21636 the value of @var{timeclamp} option
21640 @item a_weighting(f)
21641 A-weighting of equal loudness
21642 @item b_weighting(f)
21643 B-weighting of equal loudness
21644 @item c_weighting(f)
21645 C-weighting of equal loudness.
21647 Default value is @code{sono_v}.
21649 @item sono_g, gamma
21650 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
21651 higher gamma makes the spectrum having more range. Default value is @code{3}.
21652 Acceptable range is @code{[1, 7]}.
21654 @item bar_g, gamma2
21655 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
21659 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
21660 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
21662 @item timeclamp, tc
21663 Specify the transform timeclamp. At low frequency, there is trade-off between
21664 accuracy in time domain and frequency domain. If timeclamp is lower,
21665 event in time domain is represented more accurately (such as fast bass drum),
21666 otherwise event in frequency domain is represented more accurately
21667 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
21670 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
21671 limits future samples by applying asymmetric windowing in time domain, useful
21672 when low latency is required. Accepted range is @code{[0, 1]}.
21675 Specify the transform base frequency. Default value is @code{20.01523126408007475},
21676 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
21679 Specify the transform end frequency. Default value is @code{20495.59681441799654},
21680 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
21683 This option is deprecated and ignored.
21686 Specify the transform length in time domain. Use this option to control accuracy
21687 trade-off between time domain and frequency domain at every frequency sample.
21688 It can contain variables:
21690 @item frequency, freq, f
21691 the frequency where it is evaluated
21692 @item timeclamp, tc
21693 the value of @var{timeclamp} option.
21695 Default value is @code{384*tc/(384+tc*f)}.
21698 Specify the transform count for every video frame. Default value is @code{6}.
21699 Acceptable range is @code{[1, 30]}.
21702 Specify the transform count for every single pixel. Default value is @code{0},
21703 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
21706 Specify font file for use with freetype to draw the axis. If not specified,
21707 use embedded font. Note that drawing with font file or embedded font is not
21708 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
21712 Specify fontconfig pattern. This has lower priority than @var{fontfile}.
21713 The : in the pattern may be replaced by | to avoid unnecessary escaping.
21716 Specify font color expression. This is arithmetic expression that should return
21717 integer value 0xRRGGBB. It can contain variables:
21719 @item frequency, freq, f
21720 the frequency where it is evaluated
21721 @item timeclamp, tc
21722 the value of @var{timeclamp} option
21727 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
21728 @item r(x), g(x), b(x)
21729 red, green, and blue value of intensity x.
21731 Default value is @code{st(0, (midi(f)-59.5)/12);
21732 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
21733 r(1-ld(1)) + b(ld(1))}.
21736 Specify image file to draw the axis. This option override @var{fontfile} and
21737 @var{fontcolor} option.
21740 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
21741 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
21742 Default value is @code{1}.
21745 Set colorspace. The accepted values are:
21748 Unspecified (default)
21757 BT.470BG or BT.601-6 625
21760 SMPTE-170M or BT.601-6 525
21766 BT.2020 with non-constant luminance
21771 Set spectrogram color scheme. This is list of floating point values with format
21772 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
21773 The default is @code{1|0.5|0|0|0.5|1}.
21777 @subsection Examples
21781 Playing audio while showing the spectrum:
21783 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
21787 Same as above, but with frame rate 30 fps:
21789 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
21793 Playing at 1280x720:
21795 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
21799 Disable sonogram display:
21805 A1 and its harmonics: A1, A2, (near)E3, A3:
21807 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),
21808 asplit[a][out1]; [a] showcqt [out0]'
21812 Same as above, but with more accuracy in frequency domain:
21814 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),
21815 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
21821 bar_v=10:sono_v=bar_v*a_weighting(f)
21825 Custom gamma, now spectrum is linear to the amplitude.
21831 Custom tlength equation:
21833 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)))'
21837 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
21839 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
21843 Custom font using fontconfig:
21845 font='Courier New,Monospace,mono|bold'
21849 Custom frequency range with custom axis using image file:
21851 axisfile=myaxis.png:basefreq=40:endfreq=10000
21857 Convert input audio to video output representing the audio power spectrum.
21858 Audio amplitude is on Y-axis while frequency is on X-axis.
21860 The filter accepts the following options:
21864 Specify size of video. For the syntax of this option, check the
21865 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21866 Default is @code{1024x512}.
21870 This set how each frequency bin will be represented.
21872 It accepts the following values:
21878 Default is @code{bar}.
21881 Set amplitude scale.
21883 It accepts the following values:
21897 Default is @code{log}.
21900 Set frequency scale.
21902 It accepts the following values:
21911 Reverse logarithmic scale.
21913 Default is @code{lin}.
21918 It accepts the following values:
21934 Default is @code{w2048}
21937 Set windowing function.
21939 It accepts the following values:
21962 Default is @code{hanning}.
21965 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
21966 which means optimal overlap for selected window function will be picked.
21969 Set time averaging. Setting this to 0 will display current maximal peaks.
21970 Default is @code{1}, which means time averaging is disabled.
21973 Specify list of colors separated by space or by '|' which will be used to
21974 draw channel frequencies. Unrecognized or missing colors will be replaced
21978 Set channel display mode.
21980 It accepts the following values:
21985 Default is @code{combined}.
21988 Set minimum amplitude used in @code{log} amplitude scaler.
21992 @anchor{showspectrum}
21993 @section showspectrum
21995 Convert input audio to a video output, representing the audio frequency
21998 The filter accepts the following options:
22002 Specify the video size for the output. For the syntax of this option, check the
22003 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22004 Default value is @code{640x512}.
22007 Specify how the spectrum should slide along the window.
22009 It accepts the following values:
22012 the samples start again on the left when they reach the right
22014 the samples scroll from right to left
22016 frames are only produced when the samples reach the right
22018 the samples scroll from left to right
22021 Default value is @code{replace}.
22024 Specify display mode.
22026 It accepts the following values:
22029 all channels are displayed in the same row
22031 all channels are displayed in separate rows
22034 Default value is @samp{combined}.
22037 Specify display color mode.
22039 It accepts the following values:
22042 each channel is displayed in a separate color
22044 each channel is displayed using the same color scheme
22046 each channel is displayed using the rainbow color scheme
22048 each channel is displayed using the moreland color scheme
22050 each channel is displayed using the nebulae color scheme
22052 each channel is displayed using the fire color scheme
22054 each channel is displayed using the fiery color scheme
22056 each channel is displayed using the fruit color scheme
22058 each channel is displayed using the cool color scheme
22060 each channel is displayed using the magma color scheme
22062 each channel is displayed using the green color scheme
22064 each channel is displayed using the viridis color scheme
22066 each channel is displayed using the plasma color scheme
22068 each channel is displayed using the cividis color scheme
22070 each channel is displayed using the terrain color scheme
22073 Default value is @samp{channel}.
22076 Specify scale used for calculating intensity color values.
22078 It accepts the following values:
22083 square root, default
22094 Default value is @samp{sqrt}.
22097 Set saturation modifier for displayed colors. Negative values provide
22098 alternative color scheme. @code{0} is no saturation at all.
22099 Saturation must be in [-10.0, 10.0] range.
22100 Default value is @code{1}.
22103 Set window function.
22105 It accepts the following values:
22130 Default value is @code{hann}.
22133 Set orientation of time vs frequency axis. Can be @code{vertical} or
22134 @code{horizontal}. Default is @code{vertical}.
22137 Set ratio of overlap window. Default value is @code{0}.
22138 When value is @code{1} overlap is set to recommended size for specific
22139 window function currently used.
22142 Set scale gain for calculating intensity color values.
22143 Default value is @code{1}.
22146 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
22149 Set color rotation, must be in [-1.0, 1.0] range.
22150 Default value is @code{0}.
22153 Set start frequency from which to display spectrogram. Default is @code{0}.
22156 Set stop frequency to which to display spectrogram. Default is @code{0}.
22159 Set upper frame rate limit. Default is @code{auto}, unlimited.
22162 Draw time and frequency axes and legends. Default is disabled.
22165 The usage is very similar to the showwaves filter; see the examples in that
22168 @subsection Examples
22172 Large window with logarithmic color scaling:
22174 showspectrum=s=1280x480:scale=log
22178 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
22180 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
22181 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
22185 @section showspectrumpic
22187 Convert input audio to a single video frame, representing the audio frequency
22190 The filter accepts the following options:
22194 Specify the video size for the output. For the syntax of this option, check the
22195 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22196 Default value is @code{4096x2048}.
22199 Specify display mode.
22201 It accepts the following values:
22204 all channels are displayed in the same row
22206 all channels are displayed in separate rows
22208 Default value is @samp{combined}.
22211 Specify display color mode.
22213 It accepts the following values:
22216 each channel is displayed in a separate color
22218 each channel is displayed using the same color scheme
22220 each channel is displayed using the rainbow color scheme
22222 each channel is displayed using the moreland color scheme
22224 each channel is displayed using the nebulae color scheme
22226 each channel is displayed using the fire color scheme
22228 each channel is displayed using the fiery color scheme
22230 each channel is displayed using the fruit color scheme
22232 each channel is displayed using the cool color scheme
22234 each channel is displayed using the magma color scheme
22236 each channel is displayed using the green color scheme
22238 each channel is displayed using the viridis color scheme
22240 each channel is displayed using the plasma color scheme
22242 each channel is displayed using the cividis color scheme
22244 each channel is displayed using the terrain color scheme
22246 Default value is @samp{intensity}.
22249 Specify scale used for calculating intensity color values.
22251 It accepts the following values:
22256 square root, default
22266 Default value is @samp{log}.
22269 Set saturation modifier for displayed colors. Negative values provide
22270 alternative color scheme. @code{0} is no saturation at all.
22271 Saturation must be in [-10.0, 10.0] range.
22272 Default value is @code{1}.
22275 Set window function.
22277 It accepts the following values:
22301 Default value is @code{hann}.
22304 Set orientation of time vs frequency axis. Can be @code{vertical} or
22305 @code{horizontal}. Default is @code{vertical}.
22308 Set scale gain for calculating intensity color values.
22309 Default value is @code{1}.
22312 Draw time and frequency axes and legends. Default is enabled.
22315 Set color rotation, must be in [-1.0, 1.0] range.
22316 Default value is @code{0}.
22319 Set start frequency from which to display spectrogram. Default is @code{0}.
22322 Set stop frequency to which to display spectrogram. Default is @code{0}.
22325 @subsection Examples
22329 Extract an audio spectrogram of a whole audio track
22330 in a 1024x1024 picture using @command{ffmpeg}:
22332 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
22336 @section showvolume
22338 Convert input audio volume to a video output.
22340 The filter accepts the following options:
22347 Set border width, allowed range is [0, 5]. Default is 1.
22350 Set channel width, allowed range is [80, 8192]. Default is 400.
22353 Set channel height, allowed range is [1, 900]. Default is 20.
22356 Set fade, allowed range is [0, 1]. Default is 0.95.
22359 Set volume color expression.
22361 The expression can use the following variables:
22365 Current max volume of channel in dB.
22371 Current channel number, starting from 0.
22375 If set, displays channel names. Default is enabled.
22378 If set, displays volume values. Default is enabled.
22381 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
22382 default is @code{h}.
22385 Set step size, allowed range is [0, 5]. Default is 0, which means
22389 Set background opacity, allowed range is [0, 1]. Default is 0.
22392 Set metering mode, can be peak: @code{p} or rms: @code{r},
22393 default is @code{p}.
22396 Set display scale, can be linear: @code{lin} or log: @code{log},
22397 default is @code{lin}.
22401 If set to > 0., display a line for the max level
22402 in the previous seconds.
22403 default is disabled: @code{0.}
22406 The color of the max line. Use when @code{dm} option is set to > 0.
22407 default is: @code{orange}
22412 Convert input audio to a video output, representing the samples waves.
22414 The filter accepts the following options:
22418 Specify the video size for the output. For the syntax of this option, check the
22419 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22420 Default value is @code{600x240}.
22425 Available values are:
22428 Draw a point for each sample.
22431 Draw a vertical line for each sample.
22434 Draw a point for each sample and a line between them.
22437 Draw a centered vertical line for each sample.
22440 Default value is @code{point}.
22443 Set the number of samples which are printed on the same column. A
22444 larger value will decrease the frame rate. Must be a positive
22445 integer. This option can be set only if the value for @var{rate}
22446 is not explicitly specified.
22449 Set the (approximate) output frame rate. This is done by setting the
22450 option @var{n}. Default value is "25".
22452 @item split_channels
22453 Set if channels should be drawn separately or overlap. Default value is 0.
22456 Set colors separated by '|' which are going to be used for drawing of each channel.
22459 Set amplitude scale.
22461 Available values are:
22479 Set the draw mode. This is mostly useful to set for high @var{n}.
22481 Available values are:
22484 Scale pixel values for each drawn sample.
22487 Draw every sample directly.
22490 Default value is @code{scale}.
22493 @subsection Examples
22497 Output the input file audio and the corresponding video representation
22500 amovie=a.mp3,asplit[out0],showwaves[out1]
22504 Create a synthetic signal and show it with showwaves, forcing a
22505 frame rate of 30 frames per second:
22507 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
22511 @section showwavespic
22513 Convert input audio to a single video frame, representing the samples waves.
22515 The filter accepts the following options:
22519 Specify the video size for the output. For the syntax of this option, check the
22520 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22521 Default value is @code{600x240}.
22523 @item split_channels
22524 Set if channels should be drawn separately or overlap. Default value is 0.
22527 Set colors separated by '|' which are going to be used for drawing of each channel.
22530 Set amplitude scale.
22532 Available values are:
22550 @subsection Examples
22554 Extract a channel split representation of the wave form of a whole audio track
22555 in a 1024x800 picture using @command{ffmpeg}:
22557 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
22561 @section sidedata, asidedata
22563 Delete frame side data, or select frames based on it.
22565 This filter accepts the following options:
22569 Set mode of operation of the filter.
22571 Can be one of the following:
22575 Select every frame with side data of @code{type}.
22578 Delete side data of @code{type}. If @code{type} is not set, delete all side
22584 Set side data type used with all modes. Must be set for @code{select} mode. For
22585 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
22586 in @file{libavutil/frame.h}. For example, to choose
22587 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
22591 @section spectrumsynth
22593 Sythesize audio from 2 input video spectrums, first input stream represents
22594 magnitude across time and second represents phase across time.
22595 The filter will transform from frequency domain as displayed in videos back
22596 to time domain as presented in audio output.
22598 This filter is primarily created for reversing processed @ref{showspectrum}
22599 filter outputs, but can synthesize sound from other spectrograms too.
22600 But in such case results are going to be poor if the phase data is not
22601 available, because in such cases phase data need to be recreated, usually
22602 it's just recreated from random noise.
22603 For best results use gray only output (@code{channel} color mode in
22604 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
22605 @code{lin} scale for phase video. To produce phase, for 2nd video, use
22606 @code{data} option. Inputs videos should generally use @code{fullframe}
22607 slide mode as that saves resources needed for decoding video.
22609 The filter accepts the following options:
22613 Specify sample rate of output audio, the sample rate of audio from which
22614 spectrum was generated may differ.
22617 Set number of channels represented in input video spectrums.
22620 Set scale which was used when generating magnitude input spectrum.
22621 Can be @code{lin} or @code{log}. Default is @code{log}.
22624 Set slide which was used when generating inputs spectrums.
22625 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
22626 Default is @code{fullframe}.
22629 Set window function used for resynthesis.
22632 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
22633 which means optimal overlap for selected window function will be picked.
22636 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
22637 Default is @code{vertical}.
22640 @subsection Examples
22644 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
22645 then resynthesize videos back to audio with spectrumsynth:
22647 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
22648 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
22649 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
22653 @section split, asplit
22655 Split input into several identical outputs.
22657 @code{asplit} works with audio input, @code{split} with video.
22659 The filter accepts a single parameter which specifies the number of outputs. If
22660 unspecified, it defaults to 2.
22662 @subsection Examples
22666 Create two separate outputs from the same input:
22668 [in] split [out0][out1]
22672 To create 3 or more outputs, you need to specify the number of
22675 [in] asplit=3 [out0][out1][out2]
22679 Create two separate outputs from the same input, one cropped and
22682 [in] split [splitout1][splitout2];
22683 [splitout1] crop=100:100:0:0 [cropout];
22684 [splitout2] pad=200:200:100:100 [padout];
22688 Create 5 copies of the input audio with @command{ffmpeg}:
22690 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
22696 Receive commands sent through a libzmq client, and forward them to
22697 filters in the filtergraph.
22699 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
22700 must be inserted between two video filters, @code{azmq} between two
22701 audio filters. Both are capable to send messages to any filter type.
22703 To enable these filters you need to install the libzmq library and
22704 headers and configure FFmpeg with @code{--enable-libzmq}.
22706 For more information about libzmq see:
22707 @url{http://www.zeromq.org/}
22709 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
22710 receives messages sent through a network interface defined by the
22711 @option{bind_address} (or the abbreviation "@option{b}") option.
22712 Default value of this option is @file{tcp://localhost:5555}. You may
22713 want to alter this value to your needs, but do not forget to escape any
22714 ':' signs (see @ref{filtergraph escaping}).
22716 The received message must be in the form:
22718 @var{TARGET} @var{COMMAND} [@var{ARG}]
22721 @var{TARGET} specifies the target of the command, usually the name of
22722 the filter class or a specific filter instance name. The default
22723 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
22724 but you can override this by using the @samp{filter_name@@id} syntax
22725 (see @ref{Filtergraph syntax}).
22727 @var{COMMAND} specifies the name of the command for the target filter.
22729 @var{ARG} is optional and specifies the optional argument list for the
22730 given @var{COMMAND}.
22732 Upon reception, the message is processed and the corresponding command
22733 is injected into the filtergraph. Depending on the result, the filter
22734 will send a reply to the client, adopting the format:
22736 @var{ERROR_CODE} @var{ERROR_REASON}
22740 @var{MESSAGE} is optional.
22742 @subsection Examples
22744 Look at @file{tools/zmqsend} for an example of a zmq client which can
22745 be used to send commands processed by these filters.
22747 Consider the following filtergraph generated by @command{ffplay}.
22748 In this example the last overlay filter has an instance name. All other
22749 filters will have default instance names.
22752 ffplay -dumpgraph 1 -f lavfi "
22753 color=s=100x100:c=red [l];
22754 color=s=100x100:c=blue [r];
22755 nullsrc=s=200x100, zmq [bg];
22756 [bg][l] overlay [bg+l];
22757 [bg+l][r] overlay@@my=x=100 "
22760 To change the color of the left side of the video, the following
22761 command can be used:
22763 echo Parsed_color_0 c yellow | tools/zmqsend
22766 To change the right side:
22768 echo Parsed_color_1 c pink | tools/zmqsend
22771 To change the position of the right side:
22773 echo overlay@@my x 150 | tools/zmqsend
22777 @c man end MULTIMEDIA FILTERS
22779 @chapter Multimedia Sources
22780 @c man begin MULTIMEDIA SOURCES
22782 Below is a description of the currently available multimedia sources.
22786 This is the same as @ref{movie} source, except it selects an audio
22792 Read audio and/or video stream(s) from a movie container.
22794 It accepts the following parameters:
22798 The name of the resource to read (not necessarily a file; it can also be a
22799 device or a stream accessed through some protocol).
22801 @item format_name, f
22802 Specifies the format assumed for the movie to read, and can be either
22803 the name of a container or an input device. If not specified, the
22804 format is guessed from @var{movie_name} or by probing.
22806 @item seek_point, sp
22807 Specifies the seek point in seconds. The frames will be output
22808 starting from this seek point. The parameter is evaluated with
22809 @code{av_strtod}, so the numerical value may be suffixed by an IS
22810 postfix. The default value is "0".
22813 Specifies the streams to read. Several streams can be specified,
22814 separated by "+". The source will then have as many outputs, in the
22815 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
22816 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
22817 respectively the default (best suited) video and audio stream. Default
22818 is "dv", or "da" if the filter is called as "amovie".
22820 @item stream_index, si
22821 Specifies the index of the video stream to read. If the value is -1,
22822 the most suitable video stream will be automatically selected. The default
22823 value is "-1". Deprecated. If the filter is called "amovie", it will select
22824 audio instead of video.
22827 Specifies how many times to read the stream in sequence.
22828 If the value is 0, the stream will be looped infinitely.
22829 Default value is "1".
22831 Note that when the movie is looped the source timestamps are not
22832 changed, so it will generate non monotonically increasing timestamps.
22834 @item discontinuity
22835 Specifies the time difference between frames above which the point is
22836 considered a timestamp discontinuity which is removed by adjusting the later
22840 It allows overlaying a second video on top of the main input of
22841 a filtergraph, as shown in this graph:
22843 input -----------> deltapts0 --> overlay --> output
22846 movie --> scale--> deltapts1 -------+
22848 @subsection Examples
22852 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
22853 on top of the input labelled "in":
22855 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
22856 [in] setpts=PTS-STARTPTS [main];
22857 [main][over] overlay=16:16 [out]
22861 Read from a video4linux2 device, and overlay it on top of the input
22864 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
22865 [in] setpts=PTS-STARTPTS [main];
22866 [main][over] overlay=16:16 [out]
22870 Read the first video stream and the audio stream with id 0x81 from
22871 dvd.vob; the video is connected to the pad named "video" and the audio is
22872 connected to the pad named "audio":
22874 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
22878 @subsection Commands
22880 Both movie and amovie support the following commands:
22883 Perform seek using "av_seek_frame".
22884 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
22887 @var{stream_index}: If stream_index is -1, a default
22888 stream is selected, and @var{timestamp} is automatically converted
22889 from AV_TIME_BASE units to the stream specific time_base.
22891 @var{timestamp}: Timestamp in AVStream.time_base units
22892 or, if no stream is specified, in AV_TIME_BASE units.
22894 @var{flags}: Flags which select direction and seeking mode.
22898 Get movie duration in AV_TIME_BASE units.
22902 @c man end MULTIMEDIA SOURCES