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 @section Notes on filtergraph escaping
226 Filtergraph description composition entails several levels of
227 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
228 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
229 information about the employed escaping procedure.
231 A first level escaping affects the content of each filter option
232 value, which may contain the special character @code{:} used to
233 separate values, or one of the escaping characters @code{\'}.
235 A second level escaping affects the whole filter description, which
236 may contain the escaping characters @code{\'} or the special
237 characters @code{[],;} used by the filtergraph description.
239 Finally, when you specify a filtergraph on a shell commandline, you
240 need to perform a third level escaping for the shell special
241 characters contained within it.
243 For example, consider the following string to be embedded in
244 the @ref{drawtext} filter description @option{text} value:
246 this is a 'string': may contain one, or more, special characters
249 This string contains the @code{'} special escaping character, and the
250 @code{:} special character, so it needs to be escaped in this way:
252 text=this is a \'string\'\: may contain one, or more, special characters
255 A second level of escaping is required when embedding the filter
256 description in a filtergraph description, in order to escape all the
257 filtergraph special characters. Thus the example above becomes:
259 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
261 (note that in addition to the @code{\'} escaping special characters,
262 also @code{,} needs to be escaped).
264 Finally an additional level of escaping is needed when writing the
265 filtergraph description in a shell command, which depends on the
266 escaping rules of the adopted shell. For example, assuming that
267 @code{\} is special and needs to be escaped with another @code{\}, the
268 previous string will finally result in:
270 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
273 @chapter Timeline editing
275 Some filters support a generic @option{enable} option. For the filters
276 supporting timeline editing, this option can be set to an expression which is
277 evaluated before sending a frame to the filter. If the evaluation is non-zero,
278 the filter will be enabled, otherwise the frame will be sent unchanged to the
279 next filter in the filtergraph.
281 The expression accepts the following values:
284 timestamp expressed in seconds, NAN if the input timestamp is unknown
287 sequential number of the input frame, starting from 0
290 the position in the file of the input frame, NAN if unknown
294 width and height of the input frame if video
297 Additionally, these filters support an @option{enable} command that can be used
298 to re-define the expression.
300 Like any other filtering option, the @option{enable} option follows the same
303 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
304 minutes, and a @ref{curves} filter starting at 3 seconds:
306 smartblur = enable='between(t,10,3*60)',
307 curves = enable='gte(t,3)' : preset=cross_process
310 See @code{ffmpeg -filters} to view which filters have timeline support.
312 @c man end FILTERGRAPH DESCRIPTION
315 @chapter Options for filters with several inputs (framesync)
316 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
318 Some filters with several inputs support a common set of options.
319 These options can only be set by name, not with the short notation.
323 The action to take when EOF is encountered on the secondary input; it accepts
324 one of the following values:
328 Repeat the last frame (the default).
332 Pass the main input through.
336 If set to 1, force the output to terminate when the shortest input
337 terminates. Default value is 0.
340 If set to 1, force the filter to extend the last frame of secondary streams
341 until the end of the primary stream. A value of 0 disables this behavior.
345 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
347 @chapter Audio Filters
348 @c man begin AUDIO FILTERS
350 When you configure your FFmpeg build, you can disable any of the
351 existing filters using @code{--disable-filters}.
352 The configure output will show the audio filters included in your
355 Below is a description of the currently available audio filters.
359 A compressor is mainly used to reduce the dynamic range of a signal.
360 Especially modern music is mostly compressed at a high ratio to
361 improve the overall loudness. It's done to get the highest attention
362 of a listener, "fatten" the sound and bring more "power" to the track.
363 If a signal is compressed too much it may sound dull or "dead"
364 afterwards or it may start to "pump" (which could be a powerful effect
365 but can also destroy a track completely).
366 The right compression is the key to reach a professional sound and is
367 the high art of mixing and mastering. Because of its complex settings
368 it may take a long time to get the right feeling for this kind of effect.
370 Compression is done by detecting the volume above a chosen level
371 @code{threshold} and dividing it by the factor set with @code{ratio}.
372 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
373 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
374 the signal would cause distortion of the waveform the reduction can be
375 levelled over the time. This is done by setting "Attack" and "Release".
376 @code{attack} determines how long the signal has to rise above the threshold
377 before any reduction will occur and @code{release} sets the time the signal
378 has to fall below the threshold to reduce the reduction again. Shorter signals
379 than the chosen attack time will be left untouched.
380 The overall reduction of the signal can be made up afterwards with the
381 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
382 raising the makeup to this level results in a signal twice as loud than the
383 source. To gain a softer entry in the compression the @code{knee} flattens the
384 hard edge at the threshold in the range of the chosen decibels.
386 The filter accepts the following options:
390 Set input gain. Default is 1. Range is between 0.015625 and 64.
393 If a signal of stream rises above this level it will affect the gain
395 By default it is 0.125. Range is between 0.00097563 and 1.
398 Set a ratio by which the signal is reduced. 1:2 means that if the level
399 rose 4dB above the threshold, it will be only 2dB above after the reduction.
400 Default is 2. Range is between 1 and 20.
403 Amount of milliseconds the signal has to rise above the threshold before gain
404 reduction starts. Default is 20. Range is between 0.01 and 2000.
407 Amount of milliseconds the signal has to fall below the threshold before
408 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
411 Set the amount by how much signal will be amplified after processing.
412 Default is 1. Range is from 1 to 64.
415 Curve the sharp knee around the threshold to enter gain reduction more softly.
416 Default is 2.82843. Range is between 1 and 8.
419 Choose if the @code{average} level between all channels of input stream
420 or the louder(@code{maximum}) channel of input stream affects the
421 reduction. Default is @code{average}.
424 Should the exact signal be taken in case of @code{peak} or an RMS one in case
425 of @code{rms}. Default is @code{rms} which is mostly smoother.
428 How much to use compressed signal in output. Default is 1.
429 Range is between 0 and 1.
433 Simple audio dynamic range commpression/expansion filter.
435 The filter accepts the following options:
439 Set contrast. Default is 33. Allowed range is between 0 and 100.
444 Copy the input audio source unchanged to the output. This is mainly useful for
449 Apply cross fade from one input audio stream to another input audio stream.
450 The cross fade is applied for specified duration near the end of first stream.
452 The filter accepts the following options:
456 Specify the number of samples for which the cross fade effect has to last.
457 At the end of the cross fade effect the first input audio will be completely
458 silent. Default is 44100.
461 Specify the duration of the cross fade effect. See
462 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
463 for the accepted syntax.
464 By default the duration is determined by @var{nb_samples}.
465 If set this option is used instead of @var{nb_samples}.
468 Should first stream end overlap with second stream start. Default is enabled.
471 Set curve for cross fade transition for first stream.
474 Set curve for cross fade transition for second stream.
476 For description of available curve types see @ref{afade} filter description.
483 Cross fade from one input to another:
485 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
489 Cross fade from one input to another but without overlapping:
491 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
497 Reduce audio bit resolution.
499 This filter is bit crusher with enhanced functionality. A bit crusher
500 is used to audibly reduce number of bits an audio signal is sampled
501 with. This doesn't change the bit depth at all, it just produces the
502 effect. Material reduced in bit depth sounds more harsh and "digital".
503 This filter is able to even round to continuous values instead of discrete
505 Additionally it has a D/C offset which results in different crushing of
506 the lower and the upper half of the signal.
507 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
509 Another feature of this filter is the logarithmic mode.
510 This setting switches from linear distances between bits to logarithmic ones.
511 The result is a much more "natural" sounding crusher which doesn't gate low
512 signals for example. The human ear has a logarithmic perception, too
513 so this kind of crushing is much more pleasant.
514 Logarithmic crushing is also able to get anti-aliased.
516 The filter accepts the following options:
532 Can be linear: @code{lin} or logarithmic: @code{log}.
541 Set sample reduction.
544 Enable LFO. By default disabled.
555 Delay one or more audio channels.
557 Samples in delayed channel are filled with silence.
559 The filter accepts the following option:
563 Set list of delays in milliseconds for each channel separated by '|'.
564 Unused delays will be silently ignored. If number of given delays is
565 smaller than number of channels all remaining channels will not be delayed.
566 If you want to delay exact number of samples, append 'S' to number.
573 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
574 the second channel (and any other channels that may be present) unchanged.
580 Delay second channel by 500 samples, the third channel by 700 samples and leave
581 the first channel (and any other channels that may be present) unchanged.
589 Apply echoing to the input audio.
591 Echoes are reflected sound and can occur naturally amongst mountains
592 (and sometimes large buildings) when talking or shouting; digital echo
593 effects emulate this behaviour and are often used to help fill out the
594 sound of a single instrument or vocal. The time difference between the
595 original signal and the reflection is the @code{delay}, and the
596 loudness of the reflected signal is the @code{decay}.
597 Multiple echoes can have different delays and decays.
599 A description of the accepted parameters follows.
603 Set input gain of reflected signal. Default is @code{0.6}.
606 Set output gain of reflected signal. Default is @code{0.3}.
609 Set list of time intervals in milliseconds between original signal and reflections
610 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
611 Default is @code{1000}.
614 Set list of loudness of reflected signals separated by '|'.
615 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
616 Default is @code{0.5}.
623 Make it sound as if there are twice as many instruments as are actually playing:
625 aecho=0.8:0.88:60:0.4
629 If delay is very short, then it sound like a (metallic) robot playing music:
635 A longer delay will sound like an open air concert in the mountains:
637 aecho=0.8:0.9:1000:0.3
641 Same as above but with one more mountain:
643 aecho=0.8:0.9:1000|1800:0.3|0.25
648 Audio emphasis filter creates or restores material directly taken from LPs or
649 emphased CDs with different filter curves. E.g. to store music on vinyl the
650 signal has to be altered by a filter first to even out the disadvantages of
651 this recording medium.
652 Once the material is played back the inverse filter has to be applied to
653 restore the distortion of the frequency response.
655 The filter accepts the following options:
665 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
666 use @code{production} mode. Default is @code{reproduction} mode.
669 Set filter type. Selects medium. Can be one of the following:
681 select Compact Disc (CD).
687 select 50µs (FM-KF).
689 select 75µs (FM-KF).
695 Modify an audio signal according to the specified expressions.
697 This filter accepts one or more expressions (one for each channel),
698 which are evaluated and used to modify a corresponding audio signal.
700 It accepts the following parameters:
704 Set the '|'-separated expressions list for each separate channel. If
705 the number of input channels is greater than the number of
706 expressions, the last specified expression is used for the remaining
709 @item channel_layout, c
710 Set output channel layout. If not specified, the channel layout is
711 specified by the number of expressions. If set to @samp{same}, it will
712 use by default the same input channel layout.
715 Each expression in @var{exprs} can contain the following constants and functions:
719 channel number of the current expression
722 number of the evaluated sample, starting from 0
728 time of the evaluated sample expressed in seconds
731 @item nb_out_channels
732 input and output number of channels
735 the value of input channel with number @var{CH}
738 Note: this filter is slow. For faster processing you should use a
747 aeval=val(ch)/2:c=same
751 Invert phase of the second channel:
760 Apply fade-in/out effect to input audio.
762 A description of the accepted parameters follows.
766 Specify the effect type, can be either @code{in} for fade-in, or
767 @code{out} for a fade-out effect. Default is @code{in}.
769 @item start_sample, ss
770 Specify the number of the start sample for starting to apply the fade
771 effect. Default is 0.
774 Specify the number of samples for which the fade effect has to last. At
775 the end of the fade-in effect the output audio will have the same
776 volume as the input audio, at the end of the fade-out transition
777 the output audio will be silence. Default is 44100.
780 Specify the start time of the fade effect. Default is 0.
781 The value must be specified as a time duration; see
782 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
783 for the accepted syntax.
784 If set this option is used instead of @var{start_sample}.
787 Specify the duration of the fade effect. See
788 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
789 for the accepted syntax.
790 At the end of the fade-in effect the output audio will have the same
791 volume as the input audio, at the end of the fade-out transition
792 the output audio will be silence.
793 By default the duration is determined by @var{nb_samples}.
794 If set this option is used instead of @var{nb_samples}.
797 Set curve for fade transition.
799 It accepts the following values:
802 select triangular, linear slope (default)
804 select quarter of sine wave
806 select half of sine wave
808 select exponential sine wave
812 select inverted parabola
826 select inverted quarter of sine wave
828 select inverted half of sine wave
830 select double-exponential seat
832 select double-exponential sigmoid
840 Fade in first 15 seconds of audio:
846 Fade out last 25 seconds of a 900 seconds audio:
848 afade=t=out:st=875:d=25
853 Apply arbitrary expressions to samples in frequency domain.
857 Set frequency domain real expression for each separate channel separated
858 by '|'. Default is "1".
859 If the number of input channels is greater than the number of
860 expressions, the last specified expression is used for the remaining
864 Set frequency domain imaginary expression for each separate channel
865 separated by '|'. If not set, @var{real} option is used.
867 Each expression in @var{real} and @var{imag} can contain the following
875 current frequency bin number
878 number of available bins
881 channel number of the current expression
893 It accepts the following values:
909 Default is @code{w4096}
912 Set window function. Default is @code{hann}.
915 Set window overlap. If set to 1, the recommended overlap for selected
916 window function will be picked. Default is @code{0.75}.
923 Leave almost only low frequencies in audio:
925 afftfilt="1-clip((b/nb)*b,0,1)"
932 Apply an arbitrary Frequency Impulse Response filter.
934 This filter is designed for applying long FIR filters,
935 up to 30 seconds long.
937 It can be used as component for digital crossover filters,
938 room equalization, cross talk cancellation, wavefield synthesis,
939 auralization, ambiophonics and ambisonics.
941 This filter uses second stream as FIR coefficients.
942 If second stream holds single channel, it will be used
943 for all input channels in first stream, otherwise
944 number of channels in second stream must be same as
945 number of channels in first stream.
947 It accepts the following parameters:
951 Set dry gain. This sets input gain.
954 Set wet gain. This sets final output gain.
957 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
960 Enable applying gain measured from power of IR.
967 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
969 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
976 Set output format constraints for the input audio. The framework will
977 negotiate the most appropriate format to minimize conversions.
979 It accepts the following parameters:
983 A '|'-separated list of requested sample formats.
986 A '|'-separated list of requested sample rates.
988 @item channel_layouts
989 A '|'-separated list of requested channel layouts.
991 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
992 for the required syntax.
995 If a parameter is omitted, all values are allowed.
997 Force the output to either unsigned 8-bit or signed 16-bit stereo
999 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1004 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1005 processing reduces disturbing noise between useful signals.
1007 Gating is done by detecting the volume below a chosen level @var{threshold}
1008 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1009 floor is set via @var{range}. Because an exact manipulation of the signal
1010 would cause distortion of the waveform the reduction can be levelled over
1011 time. This is done by setting @var{attack} and @var{release}.
1013 @var{attack} determines how long the signal has to fall below the threshold
1014 before any reduction will occur and @var{release} sets the time the signal
1015 has to rise above the threshold to reduce the reduction again.
1016 Shorter signals than the chosen attack time will be left untouched.
1020 Set input level before filtering.
1021 Default is 1. Allowed range is from 0.015625 to 64.
1024 Set the level of gain reduction when the signal is below the threshold.
1025 Default is 0.06125. Allowed range is from 0 to 1.
1028 If a signal rises above this level the gain reduction is released.
1029 Default is 0.125. Allowed range is from 0 to 1.
1032 Set a ratio by which the signal is reduced.
1033 Default is 2. Allowed range is from 1 to 9000.
1036 Amount of milliseconds the signal has to rise above the threshold before gain
1038 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1041 Amount of milliseconds the signal has to fall below the threshold before the
1042 reduction is increased again. Default is 250 milliseconds.
1043 Allowed range is from 0.01 to 9000.
1046 Set amount of amplification of signal after processing.
1047 Default is 1. Allowed range is from 1 to 64.
1050 Curve the sharp knee around the threshold to enter gain reduction more softly.
1051 Default is 2.828427125. Allowed range is from 1 to 8.
1054 Choose if exact signal should be taken for detection or an RMS like one.
1055 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1058 Choose if the average level between all channels or the louder channel affects
1060 Default is @code{average}. Can be @code{average} or @code{maximum}.
1065 Apply an arbitrary Infinite Impulse Response filter.
1067 It accepts the following parameters:
1071 Set numerator/zeros coefficients.
1074 Set denominator/poles coefficients.
1086 Set coefficients format.
1092 Z-plane zeros/poles, cartesian (default)
1094 Z-plane zeros/poles, polar radians
1096 Z-plane zeros/poles, polar degrees
1100 Set kind of processing.
1101 Can be @code{d} - direct or @code{s} - serial cascading. Defauls is @code{s}.
1104 Set filtering precision.
1108 double-precision floating-point (default)
1110 single-precision floating-point
1119 Coefficients in @code{tf} format are separated by spaces and are in ascending
1122 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1123 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1126 Different coefficients and gains can be provided for every channel, in such case
1127 use '|' to separate coefficients or gains. Last provided coefficients will be
1128 used for all remaining channels.
1130 @subsection Examples
1134 Apply 2 pole elliptic notch at arround 5000Hz for 48000 Hz sample rate:
1136 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
1140 Same as above but in @code{zp} format:
1142 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
1148 The limiter prevents an input signal from rising over a desired threshold.
1149 This limiter uses lookahead technology to prevent your signal from distorting.
1150 It means that there is a small delay after the signal is processed. Keep in mind
1151 that the delay it produces is the attack time you set.
1153 The filter accepts the following options:
1157 Set input gain. Default is 1.
1160 Set output gain. Default is 1.
1163 Don't let signals above this level pass the limiter. Default is 1.
1166 The limiter will reach its attenuation level in this amount of time in
1167 milliseconds. Default is 5 milliseconds.
1170 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1171 Default is 50 milliseconds.
1174 When gain reduction is always needed ASC takes care of releasing to an
1175 average reduction level rather than reaching a reduction of 0 in the release
1179 Select how much the release time is affected by ASC, 0 means nearly no changes
1180 in release time while 1 produces higher release times.
1183 Auto level output signal. Default is enabled.
1184 This normalizes audio back to 0dB if enabled.
1187 Depending on picked setting it is recommended to upsample input 2x or 4x times
1188 with @ref{aresample} before applying this filter.
1192 Apply a two-pole all-pass filter with central frequency (in Hz)
1193 @var{frequency}, and filter-width @var{width}.
1194 An all-pass filter changes the audio's frequency to phase relationship
1195 without changing its frequency to amplitude relationship.
1197 The filter accepts the following options:
1201 Set frequency in Hz.
1204 Set method to specify band-width of filter.
1219 Specify the band-width of a filter in width_type units.
1222 Specify which channels to filter, by default all available are filtered.
1225 @subsection Commands
1227 This filter supports the following commands:
1230 Change allpass frequency.
1231 Syntax for the command is : "@var{frequency}"
1234 Change allpass width_type.
1235 Syntax for the command is : "@var{width_type}"
1238 Change allpass width.
1239 Syntax for the command is : "@var{width}"
1246 The filter accepts the following options:
1250 Set the number of loops. Setting this value to -1 will result in infinite loops.
1254 Set maximal number of samples. Default is 0.
1257 Set first sample of loop. Default is 0.
1263 Merge two or more audio streams into a single multi-channel stream.
1265 The filter accepts the following options:
1270 Set the number of inputs. Default is 2.
1274 If the channel layouts of the inputs are disjoint, and therefore compatible,
1275 the channel layout of the output will be set accordingly and the channels
1276 will be reordered as necessary. If the channel layouts of the inputs are not
1277 disjoint, the output will have all the channels of the first input then all
1278 the channels of the second input, in that order, and the channel layout of
1279 the output will be the default value corresponding to the total number of
1282 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1283 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1284 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1285 first input, b1 is the first channel of the second input).
1287 On the other hand, if both input are in stereo, the output channels will be
1288 in the default order: a1, a2, b1, b2, and the channel layout will be
1289 arbitrarily set to 4.0, which may or may not be the expected value.
1291 All inputs must have the same sample rate, and format.
1293 If inputs do not have the same duration, the output will stop with the
1296 @subsection Examples
1300 Merge two mono files into a stereo stream:
1302 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1306 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1308 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
1314 Mixes multiple audio inputs into a single output.
1316 Note that this filter only supports float samples (the @var{amerge}
1317 and @var{pan} audio filters support many formats). If the @var{amix}
1318 input has integer samples then @ref{aresample} will be automatically
1319 inserted to perform the conversion to float samples.
1323 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1325 will mix 3 input audio streams to a single output with the same duration as the
1326 first input and a dropout transition time of 3 seconds.
1328 It accepts the following parameters:
1332 The number of inputs. If unspecified, it defaults to 2.
1335 How to determine the end-of-stream.
1339 The duration of the longest input. (default)
1342 The duration of the shortest input.
1345 The duration of the first input.
1349 @item dropout_transition
1350 The transition time, in seconds, for volume renormalization when an input
1351 stream ends. The default value is 2 seconds.
1355 @section anequalizer
1357 High-order parametric multiband equalizer for each channel.
1359 It accepts the following parameters:
1363 This option string is in format:
1364 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1365 Each equalizer band is separated by '|'.
1369 Set channel number to which equalization will be applied.
1370 If input doesn't have that channel the entry is ignored.
1373 Set central frequency for band.
1374 If input doesn't have that frequency the entry is ignored.
1377 Set band width in hertz.
1380 Set band gain in dB.
1383 Set filter type for band, optional, can be:
1387 Butterworth, this is default.
1398 With this option activated frequency response of anequalizer is displayed
1402 Set video stream size. Only useful if curves option is activated.
1405 Set max gain that will be displayed. Only useful if curves option is activated.
1406 Setting this to a reasonable value makes it possible to display gain which is derived from
1407 neighbour bands which are too close to each other and thus produce higher gain
1408 when both are activated.
1411 Set frequency scale used to draw frequency response in video output.
1412 Can be linear or logarithmic. Default is logarithmic.
1415 Set color for each channel curve which is going to be displayed in video stream.
1416 This is list of color names separated by space or by '|'.
1417 Unrecognised or missing colors will be replaced by white color.
1420 @subsection Examples
1424 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1425 for first 2 channels using Chebyshev type 1 filter:
1427 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1431 @subsection Commands
1433 This filter supports the following commands:
1436 Alter existing filter parameters.
1437 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1439 @var{fN} is existing filter number, starting from 0, if no such filter is available
1441 @var{freq} set new frequency parameter.
1442 @var{width} set new width parameter in herz.
1443 @var{gain} set new gain parameter in dB.
1445 Full filter invocation with asendcmd may look like this:
1446 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1451 Pass the audio source unchanged to the output.
1455 Pad the end of an audio stream with silence.
1457 This can be used together with @command{ffmpeg} @option{-shortest} to
1458 extend audio streams to the same length as the video stream.
1460 A description of the accepted options follows.
1464 Set silence packet size. Default value is 4096.
1467 Set the number of samples of silence to add to the end. After the
1468 value is reached, the stream is terminated. This option is mutually
1469 exclusive with @option{whole_len}.
1472 Set the minimum total number of samples in the output audio stream. If
1473 the value is longer than the input audio length, silence is added to
1474 the end, until the value is reached. This option is mutually exclusive
1475 with @option{pad_len}.
1478 If neither the @option{pad_len} nor the @option{whole_len} option is
1479 set, the filter will add silence to the end of the input stream
1482 @subsection Examples
1486 Add 1024 samples of silence to the end of the input:
1492 Make sure the audio output will contain at least 10000 samples, pad
1493 the input with silence if required:
1495 apad=whole_len=10000
1499 Use @command{ffmpeg} to pad the audio input with silence, so that the
1500 video stream will always result the shortest and will be converted
1501 until the end in the output file when using the @option{shortest}
1504 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
1509 Add a phasing effect to the input audio.
1511 A phaser filter creates series of peaks and troughs in the frequency spectrum.
1512 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
1514 A description of the accepted parameters follows.
1518 Set input gain. Default is 0.4.
1521 Set output gain. Default is 0.74
1524 Set delay in milliseconds. Default is 3.0.
1527 Set decay. Default is 0.4.
1530 Set modulation speed in Hz. Default is 0.5.
1533 Set modulation type. Default is triangular.
1535 It accepts the following values:
1544 Audio pulsator is something between an autopanner and a tremolo.
1545 But it can produce funny stereo effects as well. Pulsator changes the volume
1546 of the left and right channel based on a LFO (low frequency oscillator) with
1547 different waveforms and shifted phases.
1548 This filter have the ability to define an offset between left and right
1549 channel. An offset of 0 means that both LFO shapes match each other.
1550 The left and right channel are altered equally - a conventional tremolo.
1551 An offset of 50% means that the shape of the right channel is exactly shifted
1552 in phase (or moved backwards about half of the frequency) - pulsator acts as
1553 an autopanner. At 1 both curves match again. Every setting in between moves the
1554 phase shift gapless between all stages and produces some "bypassing" sounds with
1555 sine and triangle waveforms. The more you set the offset near 1 (starting from
1556 the 0.5) the faster the signal passes from the left to the right speaker.
1558 The filter accepts the following options:
1562 Set input gain. By default it is 1. Range is [0.015625 - 64].
1565 Set output gain. By default it is 1. Range is [0.015625 - 64].
1568 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
1569 sawup or sawdown. Default is sine.
1572 Set modulation. Define how much of original signal is affected by the LFO.
1575 Set left channel offset. Default is 0. Allowed range is [0 - 1].
1578 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
1581 Set pulse width. Default is 1. Allowed range is [0 - 2].
1584 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
1587 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
1591 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
1595 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
1596 if timing is set to hz.
1602 Resample the input audio to the specified parameters, using the
1603 libswresample library. If none are specified then the filter will
1604 automatically convert between its input and output.
1606 This filter is also able to stretch/squeeze the audio data to make it match
1607 the timestamps or to inject silence / cut out audio to make it match the
1608 timestamps, do a combination of both or do neither.
1610 The filter accepts the syntax
1611 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
1612 expresses a sample rate and @var{resampler_options} is a list of
1613 @var{key}=@var{value} pairs, separated by ":". See the
1614 @ref{Resampler Options,,the "Resampler Options" section in the
1615 ffmpeg-resampler(1) manual,ffmpeg-resampler}
1616 for the complete list of supported options.
1618 @subsection Examples
1622 Resample the input audio to 44100Hz:
1628 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
1629 samples per second compensation:
1631 aresample=async=1000
1637 Reverse an audio clip.
1639 Warning: This filter requires memory to buffer the entire clip, so trimming
1642 @subsection Examples
1646 Take the first 5 seconds of a clip, and reverse it.
1648 atrim=end=5,areverse
1652 @section asetnsamples
1654 Set the number of samples per each output audio frame.
1656 The last output packet may contain a different number of samples, as
1657 the filter will flush all the remaining samples when the input audio
1660 The filter accepts the following options:
1664 @item nb_out_samples, n
1665 Set the number of frames per each output audio frame. The number is
1666 intended as the number of samples @emph{per each channel}.
1667 Default value is 1024.
1670 If set to 1, the filter will pad the last audio frame with zeroes, so
1671 that the last frame will contain the same number of samples as the
1672 previous ones. Default value is 1.
1675 For example, to set the number of per-frame samples to 1234 and
1676 disable padding for the last frame, use:
1678 asetnsamples=n=1234:p=0
1683 Set the sample rate without altering the PCM data.
1684 This will result in a change of speed and pitch.
1686 The filter accepts the following options:
1689 @item sample_rate, r
1690 Set the output sample rate. Default is 44100 Hz.
1695 Show a line containing various information for each input audio frame.
1696 The input audio is not modified.
1698 The shown line contains a sequence of key/value pairs of the form
1699 @var{key}:@var{value}.
1701 The following values are shown in the output:
1705 The (sequential) number of the input frame, starting from 0.
1708 The presentation timestamp of the input frame, in time base units; the time base
1709 depends on the filter input pad, and is usually 1/@var{sample_rate}.
1712 The presentation timestamp of the input frame in seconds.
1715 position of the frame in the input stream, -1 if this information in
1716 unavailable and/or meaningless (for example in case of synthetic audio)
1725 The sample rate for the audio frame.
1728 The number of samples (per channel) in the frame.
1731 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
1732 audio, the data is treated as if all the planes were concatenated.
1734 @item plane_checksums
1735 A list of Adler-32 checksums for each data plane.
1741 Display time domain statistical information about the audio channels.
1742 Statistics are calculated and displayed for each audio channel and,
1743 where applicable, an overall figure is also given.
1745 It accepts the following option:
1748 Short window length in seconds, used for peak and trough RMS measurement.
1749 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.1 - 10]}.
1753 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
1754 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
1757 Available keys for each channel are:
1791 For example full key look like this @code{lavfi.astats.1.DC_offset} or
1792 this @code{lavfi.astats.Overall.Peak_count}.
1794 For description what each key means read below.
1797 Set number of frame after which stats are going to be recalculated.
1798 Default is disabled.
1801 A description of each shown parameter follows:
1805 Mean amplitude displacement from zero.
1808 Minimal sample level.
1811 Maximal sample level.
1813 @item Min difference
1814 Minimal difference between two consecutive samples.
1816 @item Max difference
1817 Maximal difference between two consecutive samples.
1819 @item Mean difference
1820 Mean difference between two consecutive samples.
1821 The average of each difference between two consecutive samples.
1823 @item RMS difference
1824 Root Mean Square difference between two consecutive samples.
1828 Standard peak and RMS level measured in dBFS.
1832 Peak and trough values for RMS level measured over a short window.
1835 Standard ratio of peak to RMS level (note: not in dB).
1838 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
1839 (i.e. either @var{Min level} or @var{Max level}).
1842 Number of occasions (not the number of samples) that the signal attained either
1843 @var{Min level} or @var{Max level}.
1846 Overall bit depth of audio. Number of bits used for each sample.
1849 Measured dynamic range of audio in dB.
1856 The filter accepts exactly one parameter, the audio tempo. If not
1857 specified then the filter will assume nominal 1.0 tempo. Tempo must
1858 be in the [0.5, 2.0] range.
1860 @subsection Examples
1864 Slow down audio to 80% tempo:
1870 To speed up audio to 125% tempo:
1878 Trim the input so that the output contains one continuous subpart of the input.
1880 It accepts the following parameters:
1883 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
1884 sample with the timestamp @var{start} will be the first sample in the output.
1887 Specify time of the first audio sample that will be dropped, i.e. the
1888 audio sample immediately preceding the one with the timestamp @var{end} will be
1889 the last sample in the output.
1892 Same as @var{start}, except this option sets the start timestamp in samples
1896 Same as @var{end}, except this option sets the end timestamp in samples instead
1900 The maximum duration of the output in seconds.
1903 The number of the first sample that should be output.
1906 The number of the first sample that should be dropped.
1909 @option{start}, @option{end}, and @option{duration} are expressed as time
1910 duration specifications; see
1911 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
1913 Note that the first two sets of the start/end options and the @option{duration}
1914 option look at the frame timestamp, while the _sample options simply count the
1915 samples that pass through the filter. So start/end_pts and start/end_sample will
1916 give different results when the timestamps are wrong, inexact or do not start at
1917 zero. Also note that this filter does not modify the timestamps. If you wish
1918 to have the output timestamps start at zero, insert the asetpts filter after the
1921 If multiple start or end options are set, this filter tries to be greedy and
1922 keep all samples that match at least one of the specified constraints. To keep
1923 only the part that matches all the constraints at once, chain multiple atrim
1926 The defaults are such that all the input is kept. So it is possible to set e.g.
1927 just the end values to keep everything before the specified time.
1932 Drop everything except the second minute of input:
1934 ffmpeg -i INPUT -af atrim=60:120
1938 Keep only the first 1000 samples:
1940 ffmpeg -i INPUT -af atrim=end_sample=1000
1947 Apply a two-pole Butterworth band-pass filter with central
1948 frequency @var{frequency}, and (3dB-point) band-width width.
1949 The @var{csg} option selects a constant skirt gain (peak gain = Q)
1950 instead of the default: constant 0dB peak gain.
1951 The filter roll off at 6dB per octave (20dB per decade).
1953 The filter accepts the following options:
1957 Set the filter's central frequency. Default is @code{3000}.
1960 Constant skirt gain if set to 1. Defaults to 0.
1963 Set method to specify band-width of filter.
1978 Specify the band-width of a filter in width_type units.
1981 Specify which channels to filter, by default all available are filtered.
1984 @subsection Commands
1986 This filter supports the following commands:
1989 Change bandpass frequency.
1990 Syntax for the command is : "@var{frequency}"
1993 Change bandpass width_type.
1994 Syntax for the command is : "@var{width_type}"
1997 Change bandpass width.
1998 Syntax for the command is : "@var{width}"
2003 Apply a two-pole Butterworth band-reject filter with central
2004 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2005 The filter roll off at 6dB per octave (20dB per decade).
2007 The filter accepts the following options:
2011 Set the filter's central frequency. Default is @code{3000}.
2014 Set method to specify band-width of filter.
2029 Specify the band-width of a filter in width_type units.
2032 Specify which channels to filter, by default all available are filtered.
2035 @subsection Commands
2037 This filter supports the following commands:
2040 Change bandreject frequency.
2041 Syntax for the command is : "@var{frequency}"
2044 Change bandreject width_type.
2045 Syntax for the command is : "@var{width_type}"
2048 Change bandreject width.
2049 Syntax for the command is : "@var{width}"
2054 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2055 shelving filter with a response similar to that of a standard
2056 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2058 The filter accepts the following options:
2062 Give the gain at 0 Hz. Its useful range is about -20
2063 (for a large cut) to +20 (for a large boost).
2064 Beware of clipping when using a positive gain.
2067 Set the filter's central frequency and so can be used
2068 to extend or reduce the frequency range to be boosted or cut.
2069 The default value is @code{100} Hz.
2072 Set method to specify band-width of filter.
2087 Determine how steep is the filter's shelf transition.
2090 Specify which channels to filter, by default all available are filtered.
2093 @subsection Commands
2095 This filter supports the following commands:
2098 Change bass frequency.
2099 Syntax for the command is : "@var{frequency}"
2102 Change bass width_type.
2103 Syntax for the command is : "@var{width_type}"
2107 Syntax for the command is : "@var{width}"
2111 Syntax for the command is : "@var{gain}"
2116 Apply a biquad IIR filter with the given coefficients.
2117 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2118 are the numerator and denominator coefficients respectively.
2119 and @var{channels}, @var{c} specify which channels to filter, by default all
2120 available are filtered.
2122 @subsection Commands
2124 This filter supports the following commands:
2132 Change biquad parameter.
2133 Syntax for the command is : "@var{value}"
2137 Bauer stereo to binaural transformation, which improves headphone listening of
2138 stereo audio records.
2140 To enable compilation of this filter you need to configure FFmpeg with
2141 @code{--enable-libbs2b}.
2143 It accepts the following parameters:
2147 Pre-defined crossfeed level.
2151 Default level (fcut=700, feed=50).
2154 Chu Moy circuit (fcut=700, feed=60).
2157 Jan Meier circuit (fcut=650, feed=95).
2162 Cut frequency (in Hz).
2171 Remap input channels to new locations.
2173 It accepts the following parameters:
2176 Map channels from input to output. The argument is a '|'-separated list of
2177 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2178 @var{in_channel} form. @var{in_channel} can be either the name of the input
2179 channel (e.g. FL for front left) or its index in the input channel layout.
2180 @var{out_channel} is the name of the output channel or its index in the output
2181 channel layout. If @var{out_channel} is not given then it is implicitly an
2182 index, starting with zero and increasing by one for each mapping.
2184 @item channel_layout
2185 The channel layout of the output stream.
2188 If no mapping is present, the filter will implicitly map input channels to
2189 output channels, preserving indices.
2191 For example, assuming a 5.1+downmix input MOV file,
2193 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2195 will create an output WAV file tagged as stereo from the downmix channels of
2198 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2200 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2203 @section channelsplit
2205 Split each channel from an input audio stream into a separate output stream.
2207 It accepts the following parameters:
2209 @item channel_layout
2210 The channel layout of the input stream. The default is "stereo".
2213 For example, assuming a stereo input MP3 file,
2215 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2217 will create an output Matroska file with two audio streams, one containing only
2218 the left channel and the other the right channel.
2220 Split a 5.1 WAV file into per-channel files:
2222 ffmpeg -i in.wav -filter_complex
2223 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2224 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2225 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2230 Add a chorus effect to the audio.
2232 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2234 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2235 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2236 The modulation depth defines the range the modulated delay is played before or after
2237 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2238 sound tuned around the original one, like in a chorus where some vocals are slightly
2241 It accepts the following parameters:
2244 Set input gain. Default is 0.4.
2247 Set output gain. Default is 0.4.
2250 Set delays. A typical delay is around 40ms to 60ms.
2262 @subsection Examples
2268 chorus=0.7:0.9:55:0.4:0.25:2
2274 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2278 Fuller sounding chorus with three delays:
2280 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
2285 Compress or expand the audio's dynamic range.
2287 It accepts the following parameters:
2293 A list of times in seconds for each channel over which the instantaneous level
2294 of the input signal is averaged to determine its volume. @var{attacks} refers to
2295 increase of volume and @var{decays} refers to decrease of volume. For most
2296 situations, the attack time (response to the audio getting louder) should be
2297 shorter than the decay time, because the human ear is more sensitive to sudden
2298 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
2299 a typical value for decay is 0.8 seconds.
2300 If specified number of attacks & decays is lower than number of channels, the last
2301 set attack/decay will be used for all remaining channels.
2304 A list of points for the transfer function, specified in dB relative to the
2305 maximum possible signal amplitude. Each key points list must be defined using
2306 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
2307 @code{x0/y0 x1/y1 x2/y2 ....}
2309 The input values must be in strictly increasing order but the transfer function
2310 does not have to be monotonically rising. The point @code{0/0} is assumed but
2311 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
2312 function are @code{-70/-70|-60/-20|1/0}.
2315 Set the curve radius in dB for all joints. It defaults to 0.01.
2318 Set the additional gain in dB to be applied at all points on the transfer
2319 function. This allows for easy adjustment of the overall gain.
2323 Set an initial volume, in dB, to be assumed for each channel when filtering
2324 starts. This permits the user to supply a nominal level initially, so that, for
2325 example, a very large gain is not applied to initial signal levels before the
2326 companding has begun to operate. A typical value for audio which is initially
2327 quiet is -90 dB. It defaults to 0.
2330 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
2331 delayed before being fed to the volume adjuster. Specifying a delay
2332 approximately equal to the attack/decay times allows the filter to effectively
2333 operate in predictive rather than reactive mode. It defaults to 0.
2337 @subsection Examples
2341 Make music with both quiet and loud passages suitable for listening to in a
2344 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
2347 Another example for audio with whisper and explosion parts:
2349 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
2353 A noise gate for when the noise is at a lower level than the signal:
2355 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
2359 Here is another noise gate, this time for when the noise is at a higher level
2360 than the signal (making it, in some ways, similar to squelch):
2362 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
2366 2:1 compression starting at -6dB:
2368 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
2372 2:1 compression starting at -9dB:
2374 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
2378 2:1 compression starting at -12dB:
2380 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
2384 2:1 compression starting at -18dB:
2386 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
2390 3:1 compression starting at -15dB:
2392 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
2398 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
2404 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
2408 Hard limiter at -6dB:
2410 compand=attacks=0:points=-80/-80|-6/-6|20/-6
2414 Hard limiter at -12dB:
2416 compand=attacks=0:points=-80/-80|-12/-12|20/-12
2420 Hard noise gate at -35 dB:
2422 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
2428 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
2432 @section compensationdelay
2434 Compensation Delay Line is a metric based delay to compensate differing
2435 positions of microphones or speakers.
2437 For example, you have recorded guitar with two microphones placed in
2438 different location. Because the front of sound wave has fixed speed in
2439 normal conditions, the phasing of microphones can vary and depends on
2440 their location and interposition. The best sound mix can be achieved when
2441 these microphones are in phase (synchronized). Note that distance of
2442 ~30 cm between microphones makes one microphone to capture signal in
2443 antiphase to another microphone. That makes the final mix sounding moody.
2444 This filter helps to solve phasing problems by adding different delays
2445 to each microphone track and make them synchronized.
2447 The best result can be reached when you take one track as base and
2448 synchronize other tracks one by one with it.
2449 Remember that synchronization/delay tolerance depends on sample rate, too.
2450 Higher sample rates will give more tolerance.
2452 It accepts the following parameters:
2456 Set millimeters distance. This is compensation distance for fine tuning.
2460 Set cm distance. This is compensation distance for tightening distance setup.
2464 Set meters distance. This is compensation distance for hard distance setup.
2468 Set dry amount. Amount of unprocessed (dry) signal.
2472 Set wet amount. Amount of processed (wet) signal.
2476 Set temperature degree in Celsius. This is the temperature of the environment.
2481 Apply headphone crossfeed filter.
2483 Crossfeed is the process of blending the left and right channels of stereo
2485 It is mainly used to reduce extreme stereo separation of low frequencies.
2487 The intent is to produce more speaker like sound to the listener.
2489 The filter accepts the following options:
2493 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
2494 This sets gain of low shelf filter for side part of stereo image.
2495 Default is -6dB. Max allowed is -30db when strength is set to 1.
2498 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
2499 This sets cut off frequency of low shelf filter. Default is cut off near
2500 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
2503 Set input gain. Default is 0.9.
2506 Set output gain. Default is 1.
2509 @section crystalizer
2510 Simple algorithm to expand audio dynamic range.
2512 The filter accepts the following options:
2516 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
2517 (unchanged sound) to 10.0 (maximum effect).
2520 Enable clipping. By default is enabled.
2524 Apply a DC shift to the audio.
2526 This can be useful to remove a DC offset (caused perhaps by a hardware problem
2527 in the recording chain) from the audio. The effect of a DC offset is reduced
2528 headroom and hence volume. The @ref{astats} filter can be used to determine if
2529 a signal has a DC offset.
2533 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
2537 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
2538 used to prevent clipping.
2542 Dynamic Audio Normalizer.
2544 This filter applies a certain amount of gain to the input audio in order
2545 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
2546 contrast to more "simple" normalization algorithms, the Dynamic Audio
2547 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
2548 This allows for applying extra gain to the "quiet" sections of the audio
2549 while avoiding distortions or clipping the "loud" sections. In other words:
2550 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
2551 sections, in the sense that the volume of each section is brought to the
2552 same target level. Note, however, that the Dynamic Audio Normalizer achieves
2553 this goal *without* applying "dynamic range compressing". It will retain 100%
2554 of the dynamic range *within* each section of the audio file.
2558 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
2559 Default is 500 milliseconds.
2560 The Dynamic Audio Normalizer processes the input audio in small chunks,
2561 referred to as frames. This is required, because a peak magnitude has no
2562 meaning for just a single sample value. Instead, we need to determine the
2563 peak magnitude for a contiguous sequence of sample values. While a "standard"
2564 normalizer would simply use the peak magnitude of the complete file, the
2565 Dynamic Audio Normalizer determines the peak magnitude individually for each
2566 frame. The length of a frame is specified in milliseconds. By default, the
2567 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
2568 been found to give good results with most files.
2569 Note that the exact frame length, in number of samples, will be determined
2570 automatically, based on the sampling rate of the individual input audio file.
2573 Set the Gaussian filter window size. In range from 3 to 301, must be odd
2574 number. Default is 31.
2575 Probably the most important parameter of the Dynamic Audio Normalizer is the
2576 @code{window size} of the Gaussian smoothing filter. The filter's window size
2577 is specified in frames, centered around the current frame. For the sake of
2578 simplicity, this must be an odd number. Consequently, the default value of 31
2579 takes into account the current frame, as well as the 15 preceding frames and
2580 the 15 subsequent frames. Using a larger window results in a stronger
2581 smoothing effect and thus in less gain variation, i.e. slower gain
2582 adaptation. Conversely, using a smaller window results in a weaker smoothing
2583 effect and thus in more gain variation, i.e. faster gain adaptation.
2584 In other words, the more you increase this value, the more the Dynamic Audio
2585 Normalizer will behave like a "traditional" normalization filter. On the
2586 contrary, the more you decrease this value, the more the Dynamic Audio
2587 Normalizer will behave like a dynamic range compressor.
2590 Set the target peak value. This specifies the highest permissible magnitude
2591 level for the normalized audio input. This filter will try to approach the
2592 target peak magnitude as closely as possible, but at the same time it also
2593 makes sure that the normalized signal will never exceed the peak magnitude.
2594 A frame's maximum local gain factor is imposed directly by the target peak
2595 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
2596 It is not recommended to go above this value.
2599 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
2600 The Dynamic Audio Normalizer determines the maximum possible (local) gain
2601 factor for each input frame, i.e. the maximum gain factor that does not
2602 result in clipping or distortion. The maximum gain factor is determined by
2603 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
2604 additionally bounds the frame's maximum gain factor by a predetermined
2605 (global) maximum gain factor. This is done in order to avoid excessive gain
2606 factors in "silent" or almost silent frames. By default, the maximum gain
2607 factor is 10.0, For most inputs the default value should be sufficient and
2608 it usually is not recommended to increase this value. Though, for input
2609 with an extremely low overall volume level, it may be necessary to allow even
2610 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
2611 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
2612 Instead, a "sigmoid" threshold function will be applied. This way, the
2613 gain factors will smoothly approach the threshold value, but never exceed that
2617 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
2618 By default, the Dynamic Audio Normalizer performs "peak" normalization.
2619 This means that the maximum local gain factor for each frame is defined
2620 (only) by the frame's highest magnitude sample. This way, the samples can
2621 be amplified as much as possible without exceeding the maximum signal
2622 level, i.e. without clipping. Optionally, however, the Dynamic Audio
2623 Normalizer can also take into account the frame's root mean square,
2624 abbreviated RMS. In electrical engineering, the RMS is commonly used to
2625 determine the power of a time-varying signal. It is therefore considered
2626 that the RMS is a better approximation of the "perceived loudness" than
2627 just looking at the signal's peak magnitude. Consequently, by adjusting all
2628 frames to a constant RMS value, a uniform "perceived loudness" can be
2629 established. If a target RMS value has been specified, a frame's local gain
2630 factor is defined as the factor that would result in exactly that RMS value.
2631 Note, however, that the maximum local gain factor is still restricted by the
2632 frame's highest magnitude sample, in order to prevent clipping.
2635 Enable channels coupling. By default is enabled.
2636 By default, the Dynamic Audio Normalizer will amplify all channels by the same
2637 amount. This means the same gain factor will be applied to all channels, i.e.
2638 the maximum possible gain factor is determined by the "loudest" channel.
2639 However, in some recordings, it may happen that the volume of the different
2640 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
2641 In this case, this option can be used to disable the channel coupling. This way,
2642 the gain factor will be determined independently for each channel, depending
2643 only on the individual channel's highest magnitude sample. This allows for
2644 harmonizing the volume of the different channels.
2647 Enable DC bias correction. By default is disabled.
2648 An audio signal (in the time domain) is a sequence of sample values.
2649 In the Dynamic Audio Normalizer these sample values are represented in the
2650 -1.0 to 1.0 range, regardless of the original input format. Normally, the
2651 audio signal, or "waveform", should be centered around the zero point.
2652 That means if we calculate the mean value of all samples in a file, or in a
2653 single frame, then the result should be 0.0 or at least very close to that
2654 value. If, however, there is a significant deviation of the mean value from
2655 0.0, in either positive or negative direction, this is referred to as a
2656 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
2657 Audio Normalizer provides optional DC bias correction.
2658 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
2659 the mean value, or "DC correction" offset, of each input frame and subtract
2660 that value from all of the frame's sample values which ensures those samples
2661 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
2662 boundaries, the DC correction offset values will be interpolated smoothly
2663 between neighbouring frames.
2666 Enable alternative boundary mode. By default is disabled.
2667 The Dynamic Audio Normalizer takes into account a certain neighbourhood
2668 around each frame. This includes the preceding frames as well as the
2669 subsequent frames. However, for the "boundary" frames, located at the very
2670 beginning and at the very end of the audio file, not all neighbouring
2671 frames are available. In particular, for the first few frames in the audio
2672 file, the preceding frames are not known. And, similarly, for the last few
2673 frames in the audio file, the subsequent frames are not known. Thus, the
2674 question arises which gain factors should be assumed for the missing frames
2675 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
2676 to deal with this situation. The default boundary mode assumes a gain factor
2677 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
2678 "fade out" at the beginning and at the end of the input, respectively.
2681 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
2682 By default, the Dynamic Audio Normalizer does not apply "traditional"
2683 compression. This means that signal peaks will not be pruned and thus the
2684 full dynamic range will be retained within each local neighbourhood. However,
2685 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
2686 normalization algorithm with a more "traditional" compression.
2687 For this purpose, the Dynamic Audio Normalizer provides an optional compression
2688 (thresholding) function. If (and only if) the compression feature is enabled,
2689 all input frames will be processed by a soft knee thresholding function prior
2690 to the actual normalization process. Put simply, the thresholding function is
2691 going to prune all samples whose magnitude exceeds a certain threshold value.
2692 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
2693 value. Instead, the threshold value will be adjusted for each individual
2695 In general, smaller parameters result in stronger compression, and vice versa.
2696 Values below 3.0 are not recommended, because audible distortion may appear.
2701 Make audio easier to listen to on headphones.
2703 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
2704 so that when listened to on headphones the stereo image is moved from
2705 inside your head (standard for headphones) to outside and in front of
2706 the listener (standard for speakers).
2712 Apply a two-pole peaking equalisation (EQ) filter. With this
2713 filter, the signal-level at and around a selected frequency can
2714 be increased or decreased, whilst (unlike bandpass and bandreject
2715 filters) that at all other frequencies is unchanged.
2717 In order to produce complex equalisation curves, this filter can
2718 be given several times, each with a different central frequency.
2720 The filter accepts the following options:
2724 Set the filter's central frequency in Hz.
2727 Set method to specify band-width of filter.
2742 Specify the band-width of a filter in width_type units.
2745 Set the required gain or attenuation in dB.
2746 Beware of clipping when using a positive gain.
2749 Specify which channels to filter, by default all available are filtered.
2752 @subsection Examples
2755 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
2757 equalizer=f=1000:t=h:width=200:g=-10
2761 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
2763 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
2767 @subsection Commands
2769 This filter supports the following commands:
2772 Change equalizer frequency.
2773 Syntax for the command is : "@var{frequency}"
2776 Change equalizer width_type.
2777 Syntax for the command is : "@var{width_type}"
2780 Change equalizer width.
2781 Syntax for the command is : "@var{width}"
2784 Change equalizer gain.
2785 Syntax for the command is : "@var{gain}"
2788 @section extrastereo
2790 Linearly increases the difference between left and right channels which
2791 adds some sort of "live" effect to playback.
2793 The filter accepts the following options:
2797 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
2798 (average of both channels), with 1.0 sound will be unchanged, with
2799 -1.0 left and right channels will be swapped.
2802 Enable clipping. By default is enabled.
2805 @section firequalizer
2806 Apply FIR Equalization using arbitrary frequency response.
2808 The filter accepts the following option:
2812 Set gain curve equation (in dB). The expression can contain variables:
2815 the evaluated frequency
2819 channel number, set to 0 when multichannels evaluation is disabled
2821 channel id, see libavutil/channel_layout.h, set to the first channel id when
2822 multichannels evaluation is disabled
2826 channel_layout, see libavutil/channel_layout.h
2831 @item gain_interpolate(f)
2832 interpolate gain on frequency f based on gain_entry
2833 @item cubic_interpolate(f)
2834 same as gain_interpolate, but smoother
2836 This option is also available as command. Default is @code{gain_interpolate(f)}.
2839 Set gain entry for gain_interpolate function. The expression can
2843 store gain entry at frequency f with value g
2845 This option is also available as command.
2848 Set filter delay in seconds. Higher value means more accurate.
2849 Default is @code{0.01}.
2852 Set filter accuracy in Hz. Lower value means more accurate.
2853 Default is @code{5}.
2856 Set window function. Acceptable values are:
2859 rectangular window, useful when gain curve is already smooth
2861 hann window (default)
2867 3-terms continuous 1st derivative nuttall window
2869 minimum 3-terms discontinuous nuttall window
2871 4-terms continuous 1st derivative nuttall window
2873 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
2875 blackman-harris window
2881 If enabled, use fixed number of audio samples. This improves speed when
2882 filtering with large delay. Default is disabled.
2885 Enable multichannels evaluation on gain. Default is disabled.
2888 Enable zero phase mode by subtracting timestamp to compensate delay.
2889 Default is disabled.
2892 Set scale used by gain. Acceptable values are:
2895 linear frequency, linear gain
2897 linear frequency, logarithmic (in dB) gain (default)
2899 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
2901 logarithmic frequency, logarithmic gain
2905 Set file for dumping, suitable for gnuplot.
2908 Set scale for dumpfile. Acceptable values are same with scale option.
2912 Enable 2-channel convolution using complex FFT. This improves speed significantly.
2913 Default is disabled.
2916 Enable minimum phase impulse response. Default is disabled.
2919 @subsection Examples
2924 firequalizer=gain='if(lt(f,1000), 0, -INF)'
2927 lowpass at 1000 Hz with gain_entry:
2929 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
2932 custom equalization:
2934 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
2937 higher delay with zero phase to compensate delay:
2939 firequalizer=delay=0.1:fixed=on:zero_phase=on
2942 lowpass on left channel, highpass on right channel:
2944 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
2945 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
2950 Apply a flanging effect to the audio.
2952 The filter accepts the following options:
2956 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
2959 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
2962 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
2966 Set percentage of delayed signal mixed with original. Range from 0 to 100.
2967 Default value is 71.
2970 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
2973 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
2974 Default value is @var{sinusoidal}.
2977 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
2978 Default value is 25.
2981 Set delay-line interpolation, @var{linear} or @var{quadratic}.
2982 Default is @var{linear}.
2986 Apply Haas effect to audio.
2988 Note that this makes most sense to apply on mono signals.
2989 With this filter applied to mono signals it give some directionality and
2990 stretches its stereo image.
2992 The filter accepts the following options:
2996 Set input level. By default is @var{1}, or 0dB
2999 Set output level. By default is @var{1}, or 0dB.
3002 Set gain applied to side part of signal. By default is @var{1}.
3005 Set kind of middle source. Can be one of the following:
3015 Pick middle part signal of stereo image.
3018 Pick side part signal of stereo image.
3022 Change middle phase. By default is disabled.
3025 Set left channel delay. By default is @var{2.05} milliseconds.
3028 Set left channel balance. By default is @var{-1}.
3031 Set left channel gain. By default is @var{1}.
3034 Change left phase. By default is disabled.
3037 Set right channel delay. By defaults is @var{2.12} milliseconds.
3040 Set right channel balance. By default is @var{1}.
3043 Set right channel gain. By default is @var{1}.
3046 Change right phase. By default is enabled.
3051 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3052 embedded HDCD codes is expanded into a 20-bit PCM stream.
3054 The filter supports the Peak Extend and Low-level Gain Adjustment features
3055 of HDCD, and detects the Transient Filter flag.
3058 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3061 When using the filter with wav, note the default encoding for wav is 16-bit,
3062 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3063 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3065 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3066 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3069 The filter accepts the following options:
3072 @item disable_autoconvert
3073 Disable any automatic format conversion or resampling in the filter graph.
3075 @item process_stereo
3076 Process the stereo channels together. If target_gain does not match between
3077 channels, consider it invalid and use the last valid target_gain.
3080 Set the code detect timer period in ms.
3083 Always extend peaks above -3dBFS even if PE isn't signaled.
3086 Replace audio with a solid tone and adjust the amplitude to signal some
3087 specific aspect of the decoding process. The output file can be loaded in
3088 an audio editor alongside the original to aid analysis.
3090 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3097 Gain adjustment level at each sample
3099 Samples where peak extend occurs
3101 Samples where the code detect timer is active
3103 Samples where the target gain does not match between channels
3109 Apply head-related transfer functions (HRTFs) to create virtual
3110 loudspeakers around the user for binaural listening via headphones.
3111 The HRIRs are provided via additional streams, for each channel
3112 one stereo input stream is needed.
3114 The filter accepts the following options:
3118 Set mapping of input streams for convolution.
3119 The argument is a '|'-separated list of channel names in order as they
3120 are given as additional stream inputs for filter.
3121 This also specify number of input streams. Number of input streams
3122 must be not less than number of channels in first stream plus one.
3125 Set gain applied to audio. Value is in dB. Default is 0.
3128 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3129 processing audio in time domain which is slow.
3130 @var{freq} is processing audio in frequency domain which is fast.
3131 Default is @var{freq}.
3134 Set custom gain for LFE channels. Value is in dB. Default is 0.
3137 @subsection Examples
3141 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3142 each amovie filter use stereo file with IR coefficients as input.
3143 The files give coefficients for each position of virtual loudspeaker:
3145 ffmpeg -i input.wav -lavfi-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],[a:0][fl][fr][fc][lfe][bl][br][sl][sr]headphone=FL|FR|FC|LFE|BL|BR|SL|SR"
3152 Apply a high-pass filter with 3dB point frequency.
3153 The filter can be either single-pole, or double-pole (the default).
3154 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3156 The filter accepts the following options:
3160 Set frequency in Hz. Default is 3000.
3163 Set number of poles. Default is 2.
3166 Set method to specify band-width of filter.
3181 Specify the band-width of a filter in width_type units.
3182 Applies only to double-pole filter.
3183 The default is 0.707q and gives a Butterworth response.
3186 Specify which channels to filter, by default all available are filtered.
3189 @subsection Commands
3191 This filter supports the following commands:
3194 Change highpass frequency.
3195 Syntax for the command is : "@var{frequency}"
3198 Change highpass width_type.
3199 Syntax for the command is : "@var{width_type}"
3202 Change highpass width.
3203 Syntax for the command is : "@var{width}"
3208 Join multiple input streams into one multi-channel stream.
3210 It accepts the following parameters:
3214 The number of input streams. It defaults to 2.
3216 @item channel_layout
3217 The desired output channel layout. It defaults to stereo.
3220 Map channels from inputs to output. The argument is a '|'-separated list of
3221 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
3222 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
3223 can be either the name of the input channel (e.g. FL for front left) or its
3224 index in the specified input stream. @var{out_channel} is the name of the output
3228 The filter will attempt to guess the mappings when they are not specified
3229 explicitly. It does so by first trying to find an unused matching input channel
3230 and if that fails it picks the first unused input channel.
3232 Join 3 inputs (with properly set channel layouts):
3234 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
3237 Build a 5.1 output from 6 single-channel streams:
3239 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
3240 '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'
3246 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
3248 To enable compilation of this filter you need to configure FFmpeg with
3249 @code{--enable-ladspa}.
3253 Specifies the name of LADSPA plugin library to load. If the environment
3254 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
3255 each one of the directories specified by the colon separated list in
3256 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
3257 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
3258 @file{/usr/lib/ladspa/}.
3261 Specifies the plugin within the library. Some libraries contain only
3262 one plugin, but others contain many of them. If this is not set filter
3263 will list all available plugins within the specified library.
3266 Set the '|' separated list of controls which are zero or more floating point
3267 values that determine the behavior of the loaded plugin (for example delay,
3269 Controls need to be defined using the following syntax:
3270 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
3271 @var{valuei} is the value set on the @var{i}-th control.
3272 Alternatively they can be also defined using the following syntax:
3273 @var{value0}|@var{value1}|@var{value2}|..., where
3274 @var{valuei} is the value set on the @var{i}-th control.
3275 If @option{controls} is set to @code{help}, all available controls and
3276 their valid ranges are printed.
3278 @item sample_rate, s
3279 Specify the sample rate, default to 44100. Only used if plugin have
3283 Set the number of samples per channel per each output frame, default
3284 is 1024. Only used if plugin have zero inputs.
3287 Set the minimum duration of the sourced audio. See
3288 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
3289 for the accepted syntax.
3290 Note that the resulting duration may be greater than the specified duration,
3291 as the generated audio is always cut at the end of a complete frame.
3292 If not specified, or the expressed duration is negative, the audio is
3293 supposed to be generated forever.
3294 Only used if plugin have zero inputs.
3298 @subsection Examples
3302 List all available plugins within amp (LADSPA example plugin) library:
3308 List all available controls and their valid ranges for @code{vcf_notch}
3309 plugin from @code{VCF} library:
3311 ladspa=f=vcf:p=vcf_notch:c=help
3315 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
3318 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
3322 Add reverberation to the audio using TAP-plugins
3323 (Tom's Audio Processing plugins):
3325 ladspa=file=tap_reverb:tap_reverb
3329 Generate white noise, with 0.2 amplitude:
3331 ladspa=file=cmt:noise_source_white:c=c0=.2
3335 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
3336 @code{C* Audio Plugin Suite} (CAPS) library:
3338 ladspa=file=caps:Click:c=c1=20'
3342 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
3344 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
3348 Increase volume by 20dB using fast lookahead limiter from Steve Harris
3349 @code{SWH Plugins} collection:
3351 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
3355 Attenuate low frequencies using Multiband EQ from Steve Harris
3356 @code{SWH Plugins} collection:
3358 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
3362 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
3365 ladspa=caps:Narrower
3369 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
3371 ladspa=caps:White:.2
3375 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
3377 ladspa=caps:Fractal:c=c1=1
3381 Dynamic volume normalization using @code{VLevel} plugin:
3383 ladspa=vlevel-ladspa:vlevel_mono
3387 @subsection Commands
3389 This filter supports the following commands:
3392 Modify the @var{N}-th control value.
3394 If the specified value is not valid, it is ignored and prior one is kept.
3399 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
3400 Support for both single pass (livestreams, files) and double pass (files) modes.
3401 This algorithm can target IL, LRA, and maximum true peak. To accurately detect true peaks,
3402 the audio stream will be upsampled to 192 kHz unless the normalization mode is linear.
3403 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
3405 The filter accepts the following options:
3409 Set integrated loudness target.
3410 Range is -70.0 - -5.0. Default value is -24.0.
3413 Set loudness range target.
3414 Range is 1.0 - 20.0. Default value is 7.0.
3417 Set maximum true peak.
3418 Range is -9.0 - +0.0. Default value is -2.0.
3420 @item measured_I, measured_i
3421 Measured IL of input file.
3422 Range is -99.0 - +0.0.
3424 @item measured_LRA, measured_lra
3425 Measured LRA of input file.
3426 Range is 0.0 - 99.0.
3428 @item measured_TP, measured_tp
3429 Measured true peak of input file.
3430 Range is -99.0 - +99.0.
3432 @item measured_thresh
3433 Measured threshold of input file.
3434 Range is -99.0 - +0.0.
3437 Set offset gain. Gain is applied before the true-peak limiter.
3438 Range is -99.0 - +99.0. Default is +0.0.
3441 Normalize linearly if possible.
3442 measured_I, measured_LRA, measured_TP, and measured_thresh must also
3443 to be specified in order to use this mode.
3444 Options are true or false. Default is true.
3447 Treat mono input files as "dual-mono". If a mono file is intended for playback
3448 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
3449 If set to @code{true}, this option will compensate for this effect.
3450 Multi-channel input files are not affected by this option.
3451 Options are true or false. Default is false.
3454 Set print format for stats. Options are summary, json, or none.
3455 Default value is none.
3460 Apply a low-pass filter with 3dB point frequency.
3461 The filter can be either single-pole or double-pole (the default).
3462 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3464 The filter accepts the following options:
3468 Set frequency in Hz. Default is 500.
3471 Set number of poles. Default is 2.
3474 Set method to specify band-width of filter.
3489 Specify the band-width of a filter in width_type units.
3490 Applies only to double-pole filter.
3491 The default is 0.707q and gives a Butterworth response.
3494 Specify which channels to filter, by default all available are filtered.
3497 @subsection Examples
3500 Lowpass only LFE channel, it LFE is not present it does nothing:
3506 @subsection Commands
3508 This filter supports the following commands:
3511 Change lowpass frequency.
3512 Syntax for the command is : "@var{frequency}"
3515 Change lowpass width_type.
3516 Syntax for the command is : "@var{width_type}"
3519 Change lowpass width.
3520 Syntax for the command is : "@var{width}"
3525 Load a LV2 (LADSPA Version 2) plugin.
3527 To enable compilation of this filter you need to configure FFmpeg with
3528 @code{--enable-lv2}.
3532 Specifies the plugin URI. You may need to escape ':'.
3535 Set the '|' separated list of controls which are zero or more floating point
3536 values that determine the behavior of the loaded plugin (for example delay,
3538 If @option{controls} is set to @code{help}, all available controls and
3539 their valid ranges are printed.
3541 @item sample_rate, s
3542 Specify the sample rate, default to 44100. Only used if plugin have
3546 Set the number of samples per channel per each output frame, default
3547 is 1024. Only used if plugin have zero inputs.
3550 Set the minimum duration of the sourced audio. See
3551 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
3552 for the accepted syntax.
3553 Note that the resulting duration may be greater than the specified duration,
3554 as the generated audio is always cut at the end of a complete frame.
3555 If not specified, or the expressed duration is negative, the audio is
3556 supposed to be generated forever.
3557 Only used if plugin have zero inputs.
3560 @subsection Examples
3564 Apply bass enhancer plugin from Calf:
3566 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
3570 Apply bass vinyl plugin from Calf:
3572 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
3576 Apply bit crusher plugin from ArtyFX:
3578 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
3583 Multiband Compress or expand the audio's dynamic range.
3585 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
3586 This is akin to the crossover of a loudspeaker, and results in flat frequency
3587 response when absent compander action.
3589 It accepts the following parameters:
3593 This option syntax is:
3594 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
3595 For explanation of each item refer to compand filter documentation.
3601 Mix channels with specific gain levels. The filter accepts the output
3602 channel layout followed by a set of channels definitions.
3604 This filter is also designed to efficiently remap the channels of an audio
3607 The filter accepts parameters of the form:
3608 "@var{l}|@var{outdef}|@var{outdef}|..."
3612 output channel layout or number of channels
3615 output channel specification, of the form:
3616 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
3619 output channel to define, either a channel name (FL, FR, etc.) or a channel
3620 number (c0, c1, etc.)
3623 multiplicative coefficient for the channel, 1 leaving the volume unchanged
3626 input channel to use, see out_name for details; it is not possible to mix
3627 named and numbered input channels
3630 If the `=' in a channel specification is replaced by `<', then the gains for
3631 that specification will be renormalized so that the total is 1, thus
3632 avoiding clipping noise.
3634 @subsection Mixing examples
3636 For example, if you want to down-mix from stereo to mono, but with a bigger
3637 factor for the left channel:
3639 pan=1c|c0=0.9*c0+0.1*c1
3642 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
3643 7-channels surround:
3645 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
3648 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
3649 that should be preferred (see "-ac" option) unless you have very specific
3652 @subsection Remapping examples
3654 The channel remapping will be effective if, and only if:
3657 @item gain coefficients are zeroes or ones,
3658 @item only one input per channel output,
3661 If all these conditions are satisfied, the filter will notify the user ("Pure
3662 channel mapping detected"), and use an optimized and lossless method to do the
3665 For example, if you have a 5.1 source and want a stereo audio stream by
3666 dropping the extra channels:
3668 pan="stereo| c0=FL | c1=FR"
3671 Given the same source, you can also switch front left and front right channels
3672 and keep the input channel layout:
3674 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
3677 If the input is a stereo audio stream, you can mute the front left channel (and
3678 still keep the stereo channel layout) with:
3683 Still with a stereo audio stream input, you can copy the right channel in both
3684 front left and right:
3686 pan="stereo| c0=FR | c1=FR"
3691 ReplayGain scanner filter. This filter takes an audio stream as an input and
3692 outputs it unchanged.
3693 At end of filtering it displays @code{track_gain} and @code{track_peak}.
3697 Convert the audio sample format, sample rate and channel layout. It is
3698 not meant to be used directly.
3701 Apply time-stretching and pitch-shifting with librubberband.
3703 The filter accepts the following options:
3707 Set tempo scale factor.
3710 Set pitch scale factor.
3713 Set transients detector.
3714 Possible values are:
3723 Possible values are:
3732 Possible values are:
3739 Set processing window size.
3740 Possible values are:
3749 Possible values are:
3756 Enable formant preservation when shift pitching.
3757 Possible values are:
3765 Possible values are:
3774 Possible values are:
3781 @section sidechaincompress
3783 This filter acts like normal compressor but has the ability to compress
3784 detected signal using second input signal.
3785 It needs two input streams and returns one output stream.
3786 First input stream will be processed depending on second stream signal.
3787 The filtered signal then can be filtered with other filters in later stages of
3788 processing. See @ref{pan} and @ref{amerge} filter.
3790 The filter accepts the following options:
3794 Set input gain. Default is 1. Range is between 0.015625 and 64.
3797 If a signal of second stream raises above this level it will affect the gain
3798 reduction of first stream.
3799 By default is 0.125. Range is between 0.00097563 and 1.
3802 Set a ratio about which the signal is reduced. 1:2 means that if the level
3803 raised 4dB above the threshold, it will be only 2dB above after the reduction.
3804 Default is 2. Range is between 1 and 20.
3807 Amount of milliseconds the signal has to rise above the threshold before gain
3808 reduction starts. Default is 20. Range is between 0.01 and 2000.
3811 Amount of milliseconds the signal has to fall below the threshold before
3812 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
3815 Set the amount by how much signal will be amplified after processing.
3816 Default is 1. Range is from 1 to 64.
3819 Curve the sharp knee around the threshold to enter gain reduction more softly.
3820 Default is 2.82843. Range is between 1 and 8.
3823 Choose if the @code{average} level between all channels of side-chain stream
3824 or the louder(@code{maximum}) channel of side-chain stream affects the
3825 reduction. Default is @code{average}.
3828 Should the exact signal be taken in case of @code{peak} or an RMS one in case
3829 of @code{rms}. Default is @code{rms} which is mainly smoother.
3832 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
3835 How much to use compressed signal in output. Default is 1.
3836 Range is between 0 and 1.
3839 @subsection Examples
3843 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
3844 depending on the signal of 2nd input and later compressed signal to be
3845 merged with 2nd input:
3847 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
3851 @section sidechaingate
3853 A sidechain gate acts like a normal (wideband) gate but has the ability to
3854 filter the detected signal before sending it to the gain reduction stage.
3855 Normally a gate uses the full range signal to detect a level above the
3857 For example: If you cut all lower frequencies from your sidechain signal
3858 the gate will decrease the volume of your track only if not enough highs
3859 appear. With this technique you are able to reduce the resonation of a
3860 natural drum or remove "rumbling" of muted strokes from a heavily distorted
3862 It needs two input streams and returns one output stream.
3863 First input stream will be processed depending on second stream signal.
3865 The filter accepts the following options:
3869 Set input level before filtering.
3870 Default is 1. Allowed range is from 0.015625 to 64.
3873 Set the level of gain reduction when the signal is below the threshold.
3874 Default is 0.06125. Allowed range is from 0 to 1.
3877 If a signal rises above this level the gain reduction is released.
3878 Default is 0.125. Allowed range is from 0 to 1.
3881 Set a ratio about which the signal is reduced.
3882 Default is 2. Allowed range is from 1 to 9000.
3885 Amount of milliseconds the signal has to rise above the threshold before gain
3887 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
3890 Amount of milliseconds the signal has to fall below the threshold before the
3891 reduction is increased again. Default is 250 milliseconds.
3892 Allowed range is from 0.01 to 9000.
3895 Set amount of amplification of signal after processing.
3896 Default is 1. Allowed range is from 1 to 64.
3899 Curve the sharp knee around the threshold to enter gain reduction more softly.
3900 Default is 2.828427125. Allowed range is from 1 to 8.
3903 Choose if exact signal should be taken for detection or an RMS like one.
3904 Default is rms. Can be peak or rms.
3907 Choose if the average level between all channels or the louder channel affects
3909 Default is average. Can be average or maximum.
3912 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
3915 @section silencedetect
3917 Detect silence in an audio stream.
3919 This filter logs a message when it detects that the input audio volume is less
3920 or equal to a noise tolerance value for a duration greater or equal to the
3921 minimum detected noise duration.
3923 The printed times and duration are expressed in seconds.
3925 The filter accepts the following options:
3929 Set silence duration until notification (default is 2 seconds).
3932 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
3933 specified value) or amplitude ratio. Default is -60dB, or 0.001.
3936 @subsection Examples
3940 Detect 5 seconds of silence with -50dB noise tolerance:
3942 silencedetect=n=-50dB:d=5
3946 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
3947 tolerance in @file{silence.mp3}:
3949 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
3953 @section silenceremove
3955 Remove silence from the beginning, middle or end of the audio.
3957 The filter accepts the following options:
3961 This value is used to indicate if audio should be trimmed at beginning of
3962 the audio. A value of zero indicates no silence should be trimmed from the
3963 beginning. When specifying a non-zero value, it trims audio up until it
3964 finds non-silence. Normally, when trimming silence from beginning of audio
3965 the @var{start_periods} will be @code{1} but it can be increased to higher
3966 values to trim all audio up to specific count of non-silence periods.
3967 Default value is @code{0}.
3969 @item start_duration
3970 Specify the amount of time that non-silence must be detected before it stops
3971 trimming audio. By increasing the duration, bursts of noises can be treated
3972 as silence and trimmed off. Default value is @code{0}.
3974 @item start_threshold
3975 This indicates what sample value should be treated as silence. For digital
3976 audio, a value of @code{0} may be fine but for audio recorded from analog,
3977 you may wish to increase the value to account for background noise.
3978 Can be specified in dB (in case "dB" is appended to the specified value)
3979 or amplitude ratio. Default value is @code{0}.
3982 Set the count for trimming silence from the end of audio.
3983 To remove silence from the middle of a file, specify a @var{stop_periods}
3984 that is negative. This value is then treated as a positive value and is
3985 used to indicate the effect should restart processing as specified by
3986 @var{start_periods}, making it suitable for removing periods of silence
3987 in the middle of the audio.
3988 Default value is @code{0}.
3991 Specify a duration of silence that must exist before audio is not copied any
3992 more. By specifying a higher duration, silence that is wanted can be left in
3994 Default value is @code{0}.
3996 @item stop_threshold
3997 This is the same as @option{start_threshold} but for trimming silence from
3999 Can be specified in dB (in case "dB" is appended to the specified value)
4000 or amplitude ratio. Default value is @code{0}.
4003 This indicates that @var{stop_duration} length of audio should be left intact
4004 at the beginning of each period of silence.
4005 For example, if you want to remove long pauses between words but do not want
4006 to remove the pauses completely. Default value is @code{0}.
4009 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4010 and works better with digital silence which is exactly 0.
4011 Default value is @code{rms}.
4014 Set ratio used to calculate size of window for detecting silence.
4015 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4018 @subsection Examples
4022 The following example shows how this filter can be used to start a recording
4023 that does not contain the delay at the start which usually occurs between
4024 pressing the record button and the start of the performance:
4026 silenceremove=1:5:0.02
4030 Trim all silence encountered from beginning to end where there is more than 1
4031 second of silence in audio:
4033 silenceremove=0:0:0:-1:1:-90dB
4039 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4040 loudspeakers around the user for binaural listening via headphones (audio
4041 formats up to 9 channels supported).
4042 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4043 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4044 Austrian Academy of Sciences.
4046 To enable compilation of this filter you need to configure FFmpeg with
4047 @code{--enable-libmysofa}.
4049 The filter accepts the following options:
4053 Set the SOFA file used for rendering.
4056 Set gain applied to audio. Value is in dB. Default is 0.
4059 Set rotation of virtual loudspeakers in deg. Default is 0.
4062 Set elevation of virtual speakers in deg. Default is 0.
4065 Set distance in meters between loudspeakers and the listener with near-field
4066 HRTFs. Default is 1.
4069 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4070 processing audio in time domain which is slow.
4071 @var{freq} is processing audio in frequency domain which is fast.
4072 Default is @var{freq}.
4075 Set custom positions of virtual loudspeakers. Syntax for this option is:
4076 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4077 Each virtual loudspeaker is described with short channel name following with
4078 azimuth and elevation in degrees.
4079 Each virtual loudspeaker description is separated by '|'.
4080 For example to override front left and front right channel positions use:
4081 'speakers=FL 45 15|FR 345 15'.
4082 Descriptions with unrecognised channel names are ignored.
4085 Set custom gain for LFE channels. Value is in dB. Default is 0.
4088 @subsection Examples
4092 Using ClubFritz6 sofa file:
4094 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
4098 Using ClubFritz12 sofa file and bigger radius with small rotation:
4100 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
4104 Similar as above but with custom speaker positions for front left, front right, back left and back right
4105 and also with custom gain:
4107 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
4111 @section stereotools
4113 This filter has some handy utilities to manage stereo signals, for converting
4114 M/S stereo recordings to L/R signal while having control over the parameters
4115 or spreading the stereo image of master track.
4117 The filter accepts the following options:
4121 Set input level before filtering for both channels. Defaults is 1.
4122 Allowed range is from 0.015625 to 64.
4125 Set output level after filtering for both channels. Defaults is 1.
4126 Allowed range is from 0.015625 to 64.
4129 Set input balance between both channels. Default is 0.
4130 Allowed range is from -1 to 1.
4133 Set output balance between both channels. Default is 0.
4134 Allowed range is from -1 to 1.
4137 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
4138 clipping. Disabled by default.
4141 Mute the left channel. Disabled by default.
4144 Mute the right channel. Disabled by default.
4147 Change the phase of the left channel. Disabled by default.
4150 Change the phase of the right channel. Disabled by default.
4153 Set stereo mode. Available values are:
4157 Left/Right to Left/Right, this is default.
4160 Left/Right to Mid/Side.
4163 Mid/Side to Left/Right.
4166 Left/Right to Left/Left.
4169 Left/Right to Right/Right.
4172 Left/Right to Left + Right.
4175 Left/Right to Right/Left.
4178 Mid/Side to Left/Left.
4181 Mid/Side to Right/Right.
4185 Set level of side signal. Default is 1.
4186 Allowed range is from 0.015625 to 64.
4189 Set balance of side signal. Default is 0.
4190 Allowed range is from -1 to 1.
4193 Set level of the middle signal. Default is 1.
4194 Allowed range is from 0.015625 to 64.
4197 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
4200 Set stereo base between mono and inversed channels. Default is 0.
4201 Allowed range is from -1 to 1.
4204 Set delay in milliseconds how much to delay left from right channel and
4205 vice versa. Default is 0. Allowed range is from -20 to 20.
4208 Set S/C level. Default is 1. Allowed range is from 1 to 100.
4211 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
4213 @item bmode_in, bmode_out
4214 Set balance mode for balance_in/balance_out option.
4216 Can be one of the following:
4220 Classic balance mode. Attenuate one channel at time.
4221 Gain is raised up to 1.
4224 Similar as classic mode above but gain is raised up to 2.
4227 Equal power distribution, from -6dB to +6dB range.
4231 @subsection Examples
4235 Apply karaoke like effect:
4237 stereotools=mlev=0.015625
4241 Convert M/S signal to L/R:
4243 "stereotools=mode=ms>lr"
4247 @section stereowiden
4249 This filter enhance the stereo effect by suppressing signal common to both
4250 channels and by delaying the signal of left into right and vice versa,
4251 thereby widening the stereo effect.
4253 The filter accepts the following options:
4257 Time in milliseconds of the delay of left signal into right and vice versa.
4258 Default is 20 milliseconds.
4261 Amount of gain in delayed signal into right and vice versa. Gives a delay
4262 effect of left signal in right output and vice versa which gives widening
4263 effect. Default is 0.3.
4266 Cross feed of left into right with inverted phase. This helps in suppressing
4267 the mono. If the value is 1 it will cancel all the signal common to both
4268 channels. Default is 0.3.
4271 Set level of input signal of original channel. Default is 0.8.
4274 @section superequalizer
4275 Apply 18 band equalizer.
4277 The filter accepts the following options:
4284 Set 131Hz band gain.
4286 Set 185Hz band gain.
4288 Set 262Hz band gain.
4290 Set 370Hz band gain.
4292 Set 523Hz band gain.
4294 Set 740Hz band gain.
4296 Set 1047Hz band gain.
4298 Set 1480Hz band gain.
4300 Set 2093Hz band gain.
4302 Set 2960Hz band gain.
4304 Set 4186Hz band gain.
4306 Set 5920Hz band gain.
4308 Set 8372Hz band gain.
4310 Set 11840Hz band gain.
4312 Set 16744Hz band gain.
4314 Set 20000Hz band gain.
4318 Apply audio surround upmix filter.
4320 This filter allows to produce multichannel output from audio stream.
4322 The filter accepts the following options:
4326 Set output channel layout. By default, this is @var{5.1}.
4328 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4329 for the required syntax.
4332 Set input channel layout. By default, this is @var{stereo}.
4334 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4335 for the required syntax.
4338 Set input volume level. By default, this is @var{1}.
4341 Set output volume level. By default, this is @var{1}.
4344 Enable LFE channel output if output channel layout has it. By default, this is enabled.
4347 Set LFE low cut off frequency. By default, this is @var{128} Hz.
4350 Set LFE high cut off frequency. By default, this is @var{256} Hz.
4353 Set front center input volume. By default, this is @var{1}.
4356 Set front center output volume. By default, this is @var{1}.
4359 Set LFE input volume. By default, this is @var{1}.
4362 Set LFE output volume. By default, this is @var{1}.
4367 Boost or cut treble (upper) frequencies of the audio using a two-pole
4368 shelving filter with a response similar to that of a standard
4369 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
4371 The filter accepts the following options:
4375 Give the gain at whichever is the lower of ~22 kHz and the
4376 Nyquist frequency. Its useful range is about -20 (for a large cut)
4377 to +20 (for a large boost). Beware of clipping when using a positive gain.
4380 Set the filter's central frequency and so can be used
4381 to extend or reduce the frequency range to be boosted or cut.
4382 The default value is @code{3000} Hz.
4385 Set method to specify band-width of filter.
4400 Determine how steep is the filter's shelf transition.
4403 Specify which channels to filter, by default all available are filtered.
4406 @subsection Commands
4408 This filter supports the following commands:
4411 Change treble frequency.
4412 Syntax for the command is : "@var{frequency}"
4415 Change treble width_type.
4416 Syntax for the command is : "@var{width_type}"
4419 Change treble width.
4420 Syntax for the command is : "@var{width}"
4424 Syntax for the command is : "@var{gain}"
4429 Sinusoidal amplitude modulation.
4431 The filter accepts the following options:
4435 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
4436 (20 Hz or lower) will result in a tremolo effect.
4437 This filter may also be used as a ring modulator by specifying
4438 a modulation frequency higher than 20 Hz.
4439 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
4442 Depth of modulation as a percentage. Range is 0.0 - 1.0.
4443 Default value is 0.5.
4448 Sinusoidal phase modulation.
4450 The filter accepts the following options:
4454 Modulation frequency in Hertz.
4455 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
4458 Depth of modulation as a percentage. Range is 0.0 - 1.0.
4459 Default value is 0.5.
4464 Adjust the input audio volume.
4466 It accepts the following parameters:
4470 Set audio volume expression.
4472 Output values are clipped to the maximum value.
4474 The output audio volume is given by the relation:
4476 @var{output_volume} = @var{volume} * @var{input_volume}
4479 The default value for @var{volume} is "1.0".
4482 This parameter represents the mathematical precision.
4484 It determines which input sample formats will be allowed, which affects the
4485 precision of the volume scaling.
4489 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
4491 32-bit floating-point; this limits input sample format to FLT. (default)
4493 64-bit floating-point; this limits input sample format to DBL.
4497 Choose the behaviour on encountering ReplayGain side data in input frames.
4501 Remove ReplayGain side data, ignoring its contents (the default).
4504 Ignore ReplayGain side data, but leave it in the frame.
4507 Prefer the track gain, if present.
4510 Prefer the album gain, if present.
4513 @item replaygain_preamp
4514 Pre-amplification gain in dB to apply to the selected replaygain gain.
4516 Default value for @var{replaygain_preamp} is 0.0.
4519 Set when the volume expression is evaluated.
4521 It accepts the following values:
4524 only evaluate expression once during the filter initialization, or
4525 when the @samp{volume} command is sent
4528 evaluate expression for each incoming frame
4531 Default value is @samp{once}.
4534 The volume expression can contain the following parameters.
4538 frame number (starting at zero)
4541 @item nb_consumed_samples
4542 number of samples consumed by the filter
4544 number of samples in the current frame
4546 original frame position in the file
4552 PTS at start of stream
4554 time at start of stream
4560 last set volume value
4563 Note that when @option{eval} is set to @samp{once} only the
4564 @var{sample_rate} and @var{tb} variables are available, all other
4565 variables will evaluate to NAN.
4567 @subsection Commands
4569 This filter supports the following commands:
4572 Modify the volume expression.
4573 The command accepts the same syntax of the corresponding option.
4575 If the specified expression is not valid, it is kept at its current
4577 @item replaygain_noclip
4578 Prevent clipping by limiting the gain applied.
4580 Default value for @var{replaygain_noclip} is 1.
4584 @subsection Examples
4588 Halve the input audio volume:
4592 volume=volume=-6.0206dB
4595 In all the above example the named key for @option{volume} can be
4596 omitted, for example like in:
4602 Increase input audio power by 6 decibels using fixed-point precision:
4604 volume=volume=6dB:precision=fixed
4608 Fade volume after time 10 with an annihilation period of 5 seconds:
4610 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
4614 @section volumedetect
4616 Detect the volume of the input video.
4618 The filter has no parameters. The input is not modified. Statistics about
4619 the volume will be printed in the log when the input stream end is reached.
4621 In particular it will show the mean volume (root mean square), maximum
4622 volume (on a per-sample basis), and the beginning of a histogram of the
4623 registered volume values (from the maximum value to a cumulated 1/1000 of
4626 All volumes are in decibels relative to the maximum PCM value.
4628 @subsection Examples
4630 Here is an excerpt of the output:
4632 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
4633 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
4634 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
4635 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
4636 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
4637 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
4638 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
4639 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
4640 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
4646 The mean square energy is approximately -27 dB, or 10^-2.7.
4648 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
4650 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
4653 In other words, raising the volume by +4 dB does not cause any clipping,
4654 raising it by +5 dB causes clipping for 6 samples, etc.
4656 @c man end AUDIO FILTERS
4658 @chapter Audio Sources
4659 @c man begin AUDIO SOURCES
4661 Below is a description of the currently available audio sources.
4665 Buffer audio frames, and make them available to the filter chain.
4667 This source is mainly intended for a programmatic use, in particular
4668 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
4670 It accepts the following parameters:
4674 The timebase which will be used for timestamps of submitted frames. It must be
4675 either a floating-point number or in @var{numerator}/@var{denominator} form.
4678 The sample rate of the incoming audio buffers.
4681 The sample format of the incoming audio buffers.
4682 Either a sample format name or its corresponding integer representation from
4683 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
4685 @item channel_layout
4686 The channel layout of the incoming audio buffers.
4687 Either a channel layout name from channel_layout_map in
4688 @file{libavutil/channel_layout.c} or its corresponding integer representation
4689 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
4692 The number of channels of the incoming audio buffers.
4693 If both @var{channels} and @var{channel_layout} are specified, then they
4698 @subsection Examples
4701 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
4704 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
4705 Since the sample format with name "s16p" corresponds to the number
4706 6 and the "stereo" channel layout corresponds to the value 0x3, this is
4709 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
4714 Generate an audio signal specified by an expression.
4716 This source accepts in input one or more expressions (one for each
4717 channel), which are evaluated and used to generate a corresponding
4720 This source accepts the following options:
4724 Set the '|'-separated expressions list for each separate channel. In case the
4725 @option{channel_layout} option is not specified, the selected channel layout
4726 depends on the number of provided expressions. Otherwise the last
4727 specified expression is applied to the remaining output channels.
4729 @item channel_layout, c
4730 Set the channel layout. The number of channels in the specified layout
4731 must be equal to the number of specified expressions.
4734 Set the minimum duration of the sourced audio. See
4735 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4736 for the accepted syntax.
4737 Note that the resulting duration may be greater than the specified
4738 duration, as the generated audio is always cut at the end of a
4741 If not specified, or the expressed duration is negative, the audio is
4742 supposed to be generated forever.
4745 Set the number of samples per channel per each output frame,
4748 @item sample_rate, s
4749 Specify the sample rate, default to 44100.
4752 Each expression in @var{exprs} can contain the following constants:
4756 number of the evaluated sample, starting from 0
4759 time of the evaluated sample expressed in seconds, starting from 0
4766 @subsection Examples
4776 Generate a sin signal with frequency of 440 Hz, set sample rate to
4779 aevalsrc="sin(440*2*PI*t):s=8000"
4783 Generate a two channels signal, specify the channel layout (Front
4784 Center + Back Center) explicitly:
4786 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
4790 Generate white noise:
4792 aevalsrc="-2+random(0)"
4796 Generate an amplitude modulated signal:
4798 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
4802 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
4804 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
4811 The null audio source, return unprocessed audio frames. It is mainly useful
4812 as a template and to be employed in analysis / debugging tools, or as
4813 the source for filters which ignore the input data (for example the sox
4816 This source accepts the following options:
4820 @item channel_layout, cl
4822 Specifies the channel layout, and can be either an integer or a string
4823 representing a channel layout. The default value of @var{channel_layout}
4826 Check the channel_layout_map definition in
4827 @file{libavutil/channel_layout.c} for the mapping between strings and
4828 channel layout values.
4830 @item sample_rate, r
4831 Specifies the sample rate, and defaults to 44100.
4834 Set the number of samples per requested frames.
4838 @subsection Examples
4842 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
4844 anullsrc=r=48000:cl=4
4848 Do the same operation with a more obvious syntax:
4850 anullsrc=r=48000:cl=mono
4854 All the parameters need to be explicitly defined.
4858 Synthesize a voice utterance using the libflite library.
4860 To enable compilation of this filter you need to configure FFmpeg with
4861 @code{--enable-libflite}.
4863 Note that versions of the flite library prior to 2.0 are not thread-safe.
4865 The filter accepts the following options:
4870 If set to 1, list the names of the available voices and exit
4871 immediately. Default value is 0.
4874 Set the maximum number of samples per frame. Default value is 512.
4877 Set the filename containing the text to speak.
4880 Set the text to speak.
4883 Set the voice to use for the speech synthesis. Default value is
4884 @code{kal}. See also the @var{list_voices} option.
4887 @subsection Examples
4891 Read from file @file{speech.txt}, and synthesize the text using the
4892 standard flite voice:
4894 flite=textfile=speech.txt
4898 Read the specified text selecting the @code{slt} voice:
4900 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
4904 Input text to ffmpeg:
4906 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
4910 Make @file{ffplay} speak the specified text, using @code{flite} and
4911 the @code{lavfi} device:
4913 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
4917 For more information about libflite, check:
4918 @url{http://www.festvox.org/flite/}
4922 Generate a noise audio signal.
4924 The filter accepts the following options:
4927 @item sample_rate, r
4928 Specify the sample rate. Default value is 48000 Hz.
4931 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
4935 Specify the duration of the generated audio stream. Not specifying this option
4936 results in noise with an infinite length.
4938 @item color, colour, c
4939 Specify the color of noise. Available noise colors are white, pink, brown,
4940 blue and violet. Default color is white.
4943 Specify a value used to seed the PRNG.
4946 Set the number of samples per each output frame, default is 1024.
4949 @subsection Examples
4954 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
4956 anoisesrc=d=60:c=pink:r=44100:a=0.5
4962 Generate odd-tap Hilbert transform FIR coefficients.
4964 The resulting stream can be used with @ref{afir} filter for phase-shifting
4965 the signal by 90 degrees.
4967 This is used in many matrix coding schemes and for analytic signal generation.
4968 The process is often written as a multiplication by i (or j), the imaginary unit.
4970 The filter accepts the following options:
4974 @item sample_rate, s
4975 Set sample rate, default is 44100.
4978 Set length of FIR filter, default is 22051.
4981 Set number of samples per each frame.
4984 Set window function to be used when generating FIR coefficients.
4989 Generate an audio signal made of a sine wave with amplitude 1/8.
4991 The audio signal is bit-exact.
4993 The filter accepts the following options:
4998 Set the carrier frequency. Default is 440 Hz.
5000 @item beep_factor, b
5001 Enable a periodic beep every second with frequency @var{beep_factor} times
5002 the carrier frequency. Default is 0, meaning the beep is disabled.
5004 @item sample_rate, r
5005 Specify the sample rate, default is 44100.
5008 Specify the duration of the generated audio stream.
5010 @item samples_per_frame
5011 Set the number of samples per output frame.
5013 The expression can contain the following constants:
5017 The (sequential) number of the output audio frame, starting from 0.
5020 The PTS (Presentation TimeStamp) of the output audio frame,
5021 expressed in @var{TB} units.
5024 The PTS of the output audio frame, expressed in seconds.
5027 The timebase of the output audio frames.
5030 Default is @code{1024}.
5033 @subsection Examples
5038 Generate a simple 440 Hz sine wave:
5044 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
5048 sine=frequency=220:beep_factor=4:duration=5
5052 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
5055 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
5059 @c man end AUDIO SOURCES
5061 @chapter Audio Sinks
5062 @c man begin AUDIO SINKS
5064 Below is a description of the currently available audio sinks.
5066 @section abuffersink
5068 Buffer audio frames, and make them available to the end of filter chain.
5070 This sink is mainly intended for programmatic use, in particular
5071 through the interface defined in @file{libavfilter/buffersink.h}
5072 or the options system.
5074 It accepts a pointer to an AVABufferSinkContext structure, which
5075 defines the incoming buffers' formats, to be passed as the opaque
5076 parameter to @code{avfilter_init_filter} for initialization.
5079 Null audio sink; do absolutely nothing with the input audio. It is
5080 mainly useful as a template and for use in analysis / debugging
5083 @c man end AUDIO SINKS
5085 @chapter Video Filters
5086 @c man begin VIDEO FILTERS
5088 When you configure your FFmpeg build, you can disable any of the
5089 existing filters using @code{--disable-filters}.
5090 The configure output will show the video filters included in your
5093 Below is a description of the currently available video filters.
5095 @section alphaextract
5097 Extract the alpha component from the input as a grayscale video. This
5098 is especially useful with the @var{alphamerge} filter.
5102 Add or replace the alpha component of the primary input with the
5103 grayscale value of a second input. This is intended for use with
5104 @var{alphaextract} to allow the transmission or storage of frame
5105 sequences that have alpha in a format that doesn't support an alpha
5108 For example, to reconstruct full frames from a normal YUV-encoded video
5109 and a separate video created with @var{alphaextract}, you might use:
5111 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
5114 Since this filter is designed for reconstruction, it operates on frame
5115 sequences without considering timestamps, and terminates when either
5116 input reaches end of stream. This will cause problems if your encoding
5117 pipeline drops frames. If you're trying to apply an image as an
5118 overlay to a video stream, consider the @var{overlay} filter instead.
5122 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
5123 and libavformat to work. On the other hand, it is limited to ASS (Advanced
5124 Substation Alpha) subtitles files.
5126 This filter accepts the following option in addition to the common options from
5127 the @ref{subtitles} filter:
5131 Set the shaping engine
5133 Available values are:
5136 The default libass shaping engine, which is the best available.
5138 Fast, font-agnostic shaper that can do only substitutions
5140 Slower shaper using OpenType for substitutions and positioning
5143 The default is @code{auto}.
5147 Apply an Adaptive Temporal Averaging Denoiser to the video input.
5149 The filter accepts the following options:
5153 Set threshold A for 1st plane. Default is 0.02.
5154 Valid range is 0 to 0.3.
5157 Set threshold B for 1st plane. Default is 0.04.
5158 Valid range is 0 to 5.
5161 Set threshold A for 2nd plane. Default is 0.02.
5162 Valid range is 0 to 0.3.
5165 Set threshold B for 2nd plane. Default is 0.04.
5166 Valid range is 0 to 5.
5169 Set threshold A for 3rd plane. Default is 0.02.
5170 Valid range is 0 to 0.3.
5173 Set threshold B for 3rd plane. Default is 0.04.
5174 Valid range is 0 to 5.
5176 Threshold A is designed to react on abrupt changes in the input signal and
5177 threshold B is designed to react on continuous changes in the input signal.
5180 Set number of frames filter will use for averaging. Default is 33. Must be odd
5181 number in range [5, 129].
5184 Set what planes of frame filter will use for averaging. Default is all.
5189 Apply average blur filter.
5191 The filter accepts the following options:
5195 Set horizontal kernel size.
5198 Set which planes to filter. By default all planes are filtered.
5201 Set vertical kernel size, if zero it will be same as @code{sizeX}.
5202 Default is @code{0}.
5207 Compute the bounding box for the non-black pixels in the input frame
5210 This filter computes the bounding box containing all the pixels with a
5211 luminance value greater than the minimum allowed value.
5212 The parameters describing the bounding box are printed on the filter
5215 The filter accepts the following option:
5219 Set the minimal luminance value. Default is @code{16}.
5222 @section bitplanenoise
5224 Show and measure bit plane noise.
5226 The filter accepts the following options:
5230 Set which plane to analyze. Default is @code{1}.
5233 Filter out noisy pixels from @code{bitplane} set above.
5234 Default is disabled.
5237 @section blackdetect
5239 Detect video intervals that are (almost) completely black. Can be
5240 useful to detect chapter transitions, commercials, or invalid
5241 recordings. Output lines contains the time for the start, end and
5242 duration of the detected black interval expressed in seconds.
5244 In order to display the output lines, you need to set the loglevel at
5245 least to the AV_LOG_INFO value.
5247 The filter accepts the following options:
5250 @item black_min_duration, d
5251 Set the minimum detected black duration expressed in seconds. It must
5252 be a non-negative floating point number.
5254 Default value is 2.0.
5256 @item picture_black_ratio_th, pic_th
5257 Set the threshold for considering a picture "black".
5258 Express the minimum value for the ratio:
5260 @var{nb_black_pixels} / @var{nb_pixels}
5263 for which a picture is considered black.
5264 Default value is 0.98.
5266 @item pixel_black_th, pix_th
5267 Set the threshold for considering a pixel "black".
5269 The threshold expresses the maximum pixel luminance value for which a
5270 pixel is considered "black". The provided value is scaled according to
5271 the following equation:
5273 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
5276 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
5277 the input video format, the range is [0-255] for YUV full-range
5278 formats and [16-235] for YUV non full-range formats.
5280 Default value is 0.10.
5283 The following example sets the maximum pixel threshold to the minimum
5284 value, and detects only black intervals of 2 or more seconds:
5286 blackdetect=d=2:pix_th=0.00
5291 Detect frames that are (almost) completely black. Can be useful to
5292 detect chapter transitions or commercials. Output lines consist of
5293 the frame number of the detected frame, the percentage of blackness,
5294 the position in the file if known or -1 and the timestamp in seconds.
5296 In order to display the output lines, you need to set the loglevel at
5297 least to the AV_LOG_INFO value.
5299 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
5300 The value represents the percentage of pixels in the picture that
5301 are below the threshold value.
5303 It accepts the following parameters:
5308 The percentage of the pixels that have to be below the threshold; it defaults to
5311 @item threshold, thresh
5312 The threshold below which a pixel value is considered black; it defaults to
5317 @section blend, tblend
5319 Blend two video frames into each other.
5321 The @code{blend} filter takes two input streams and outputs one
5322 stream, the first input is the "top" layer and second input is
5323 "bottom" layer. By default, the output terminates when the longest input terminates.
5325 The @code{tblend} (time blend) filter takes two consecutive frames
5326 from one single stream, and outputs the result obtained by blending
5327 the new frame on top of the old frame.
5329 A description of the accepted options follows.
5337 Set blend mode for specific pixel component or all pixel components in case
5338 of @var{all_mode}. Default value is @code{normal}.
5340 Available values for component modes are:
5382 Set blend opacity for specific pixel component or all pixel components in case
5383 of @var{all_opacity}. Only used in combination with pixel component blend modes.
5390 Set blend expression for specific pixel component or all pixel components in case
5391 of @var{all_expr}. Note that related mode options will be ignored if those are set.
5393 The expressions can use the following variables:
5397 The sequential number of the filtered frame, starting from @code{0}.
5401 the coordinates of the current sample
5405 the width and height of currently filtered plane
5409 Width and height scale depending on the currently filtered plane. It is the
5410 ratio between the corresponding luma plane number of pixels and the current
5411 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
5412 @code{0.5,0.5} for chroma planes.
5415 Time of the current frame, expressed in seconds.
5418 Value of pixel component at current location for first video frame (top layer).
5421 Value of pixel component at current location for second video frame (bottom layer).
5425 The @code{blend} filter also supports the @ref{framesync} options.
5427 @subsection Examples
5431 Apply transition from bottom layer to top layer in first 10 seconds:
5433 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
5437 Apply linear horizontal transition from top layer to bottom layer:
5439 blend=all_expr='A*(X/W)+B*(1-X/W)'
5443 Apply 1x1 checkerboard effect:
5445 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
5449 Apply uncover left effect:
5451 blend=all_expr='if(gte(N*SW+X,W),A,B)'
5455 Apply uncover down effect:
5457 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
5461 Apply uncover up-left effect:
5463 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
5467 Split diagonally video and shows top and bottom layer on each side:
5469 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
5473 Display differences between the current and the previous frame:
5475 tblend=all_mode=grainextract
5481 Apply a boxblur algorithm to the input video.
5483 It accepts the following parameters:
5487 @item luma_radius, lr
5488 @item luma_power, lp
5489 @item chroma_radius, cr
5490 @item chroma_power, cp
5491 @item alpha_radius, ar
5492 @item alpha_power, ap
5496 A description of the accepted options follows.
5499 @item luma_radius, lr
5500 @item chroma_radius, cr
5501 @item alpha_radius, ar
5502 Set an expression for the box radius in pixels used for blurring the
5503 corresponding input plane.
5505 The radius value must be a non-negative number, and must not be
5506 greater than the value of the expression @code{min(w,h)/2} for the
5507 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
5510 Default value for @option{luma_radius} is "2". If not specified,
5511 @option{chroma_radius} and @option{alpha_radius} default to the
5512 corresponding value set for @option{luma_radius}.
5514 The expressions can contain the following constants:
5518 The input width and height in pixels.
5522 The input chroma image width and height in pixels.
5526 The horizontal and vertical chroma subsample values. For example, for the
5527 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
5530 @item luma_power, lp
5531 @item chroma_power, cp
5532 @item alpha_power, ap
5533 Specify how many times the boxblur filter is applied to the
5534 corresponding plane.
5536 Default value for @option{luma_power} is 2. If not specified,
5537 @option{chroma_power} and @option{alpha_power} default to the
5538 corresponding value set for @option{luma_power}.
5540 A value of 0 will disable the effect.
5543 @subsection Examples
5547 Apply a boxblur filter with the luma, chroma, and alpha radii
5550 boxblur=luma_radius=2:luma_power=1
5555 Set the luma radius to 2, and alpha and chroma radius to 0:
5557 boxblur=2:1:cr=0:ar=0
5561 Set the luma and chroma radii to a fraction of the video dimension:
5563 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
5569 Deinterlace the input video ("bwdif" stands for "Bob Weaver
5570 Deinterlacing Filter").
5572 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
5573 interpolation algorithms.
5574 It accepts the following parameters:
5578 The interlacing mode to adopt. It accepts one of the following values:
5582 Output one frame for each frame.
5584 Output one frame for each field.
5587 The default value is @code{send_field}.
5590 The picture field parity assumed for the input interlaced video. It accepts one
5591 of the following values:
5595 Assume the top field is first.
5597 Assume the bottom field is first.
5599 Enable automatic detection of field parity.
5602 The default value is @code{auto}.
5603 If the interlacing is unknown or the decoder does not export this information,
5604 top field first will be assumed.
5607 Specify which frames to deinterlace. Accept one of the following
5612 Deinterlace all frames.
5614 Only deinterlace frames marked as interlaced.
5617 The default value is @code{all}.
5621 YUV colorspace color/chroma keying.
5623 The filter accepts the following options:
5627 The color which will be replaced with transparency.
5630 Similarity percentage with the key color.
5632 0.01 matches only the exact key color, while 1.0 matches everything.
5637 0.0 makes pixels either fully transparent, or not transparent at all.
5639 Higher values result in semi-transparent pixels, with a higher transparency
5640 the more similar the pixels color is to the key color.
5643 Signals that the color passed is already in YUV instead of RGB.
5645 Literal colors like "green" or "red" don't make sense with this enabled anymore.
5646 This can be used to pass exact YUV values as hexadecimal numbers.
5649 @subsection Examples
5653 Make every green pixel in the input image transparent:
5655 ffmpeg -i input.png -vf chromakey=green out.png
5659 Overlay a greenscreen-video on top of a static black background.
5661 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
5667 Display CIE color diagram with pixels overlaid onto it.
5669 The filter accepts the following options:
5684 @item uhdtv, rec2020
5697 Set what gamuts to draw.
5699 See @code{system} option for available values.
5702 Set ciescope size, by default set to 512.
5705 Set intensity used to map input pixel values to CIE diagram.
5708 Set contrast used to draw tongue colors that are out of active color system gamut.
5711 Correct gamma displayed on scope, by default enabled.
5714 Show white point on CIE diagram, by default disabled.
5717 Set input gamma. Used only with XYZ input color space.
5722 Visualize information exported by some codecs.
5724 Some codecs can export information through frames using side-data or other
5725 means. For example, some MPEG based codecs export motion vectors through the
5726 @var{export_mvs} flag in the codec @option{flags2} option.
5728 The filter accepts the following option:
5732 Set motion vectors to visualize.
5734 Available flags for @var{mv} are:
5738 forward predicted MVs of P-frames
5740 forward predicted MVs of B-frames
5742 backward predicted MVs of B-frames
5746 Display quantization parameters using the chroma planes.
5749 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
5751 Available flags for @var{mv_type} are:
5755 forward predicted MVs
5757 backward predicted MVs
5760 @item frame_type, ft
5761 Set frame type to visualize motion vectors of.
5763 Available flags for @var{frame_type} are:
5767 intra-coded frames (I-frames)
5769 predicted frames (P-frames)
5771 bi-directionally predicted frames (B-frames)
5775 @subsection Examples
5779 Visualize forward predicted MVs of all frames using @command{ffplay}:
5781 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
5785 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
5787 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
5791 @section colorbalance
5792 Modify intensity of primary colors (red, green and blue) of input frames.
5794 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
5795 regions for the red-cyan, green-magenta or blue-yellow balance.
5797 A positive adjustment value shifts the balance towards the primary color, a negative
5798 value towards the complementary color.
5800 The filter accepts the following options:
5806 Adjust red, green and blue shadows (darkest pixels).
5811 Adjust red, green and blue midtones (medium pixels).
5816 Adjust red, green and blue highlights (brightest pixels).
5818 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
5821 @subsection Examples
5825 Add red color cast to shadows:
5832 RGB colorspace color keying.
5834 The filter accepts the following options:
5838 The color which will be replaced with transparency.
5841 Similarity percentage with the key color.
5843 0.01 matches only the exact key color, while 1.0 matches everything.
5848 0.0 makes pixels either fully transparent, or not transparent at all.
5850 Higher values result in semi-transparent pixels, with a higher transparency
5851 the more similar the pixels color is to the key color.
5854 @subsection Examples
5858 Make every green pixel in the input image transparent:
5860 ffmpeg -i input.png -vf colorkey=green out.png
5864 Overlay a greenscreen-video on top of a static background image.
5866 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
5870 @section colorlevels
5872 Adjust video input frames using levels.
5874 The filter accepts the following options:
5881 Adjust red, green, blue and alpha input black point.
5882 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
5888 Adjust red, green, blue and alpha input white point.
5889 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
5891 Input levels are used to lighten highlights (bright tones), darken shadows
5892 (dark tones), change the balance of bright and dark tones.
5898 Adjust red, green, blue and alpha output black point.
5899 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
5905 Adjust red, green, blue and alpha output white point.
5906 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
5908 Output levels allows manual selection of a constrained output level range.
5911 @subsection Examples
5915 Make video output darker:
5917 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
5923 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
5927 Make video output lighter:
5929 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
5933 Increase brightness:
5935 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
5939 @section colorchannelmixer
5941 Adjust video input frames by re-mixing color channels.
5943 This filter modifies a color channel by adding the values associated to
5944 the other channels of the same pixels. For example if the value to
5945 modify is red, the output value will be:
5947 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
5950 The filter accepts the following options:
5957 Adjust contribution of input red, green, blue and alpha channels for output red channel.
5958 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
5964 Adjust contribution of input red, green, blue and alpha channels for output green channel.
5965 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
5971 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
5972 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
5978 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
5979 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
5981 Allowed ranges for options are @code{[-2.0, 2.0]}.
5984 @subsection Examples
5988 Convert source to grayscale:
5990 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
5993 Simulate sepia tones:
5995 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
5999 @section colormatrix
6001 Convert color matrix.
6003 The filter accepts the following options:
6008 Specify the source and destination color matrix. Both values must be
6011 The accepted values are:
6039 For example to convert from BT.601 to SMPTE-240M, use the command:
6041 colormatrix=bt601:smpte240m
6046 Convert colorspace, transfer characteristics or color primaries.
6047 Input video needs to have an even size.
6049 The filter accepts the following options:
6054 Specify all color properties at once.
6056 The accepted values are:
6086 Specify output colorspace.
6088 The accepted values are:
6097 BT.470BG or BT.601-6 625
6100 SMPTE-170M or BT.601-6 525
6109 BT.2020 with non-constant luminance
6115 Specify output transfer characteristics.
6117 The accepted values are:
6129 Constant gamma of 2.2
6132 Constant gamma of 2.8
6135 SMPTE-170M, BT.601-6 625 or BT.601-6 525
6153 BT.2020 for 10-bits content
6156 BT.2020 for 12-bits content
6162 Specify output color primaries.
6164 The accepted values are:
6173 BT.470BG or BT.601-6 625
6176 SMPTE-170M or BT.601-6 525
6200 Specify output color range.
6202 The accepted values are:
6205 TV (restricted) range
6208 MPEG (restricted) range
6219 Specify output color format.
6221 The accepted values are:
6224 YUV 4:2:0 planar 8-bits
6227 YUV 4:2:0 planar 10-bits
6230 YUV 4:2:0 planar 12-bits
6233 YUV 4:2:2 planar 8-bits
6236 YUV 4:2:2 planar 10-bits
6239 YUV 4:2:2 planar 12-bits
6242 YUV 4:4:4 planar 8-bits
6245 YUV 4:4:4 planar 10-bits
6248 YUV 4:4:4 planar 12-bits
6253 Do a fast conversion, which skips gamma/primary correction. This will take
6254 significantly less CPU, but will be mathematically incorrect. To get output
6255 compatible with that produced by the colormatrix filter, use fast=1.
6258 Specify dithering mode.
6260 The accepted values are:
6266 Floyd-Steinberg dithering
6270 Whitepoint adaptation mode.
6272 The accepted values are:
6275 Bradford whitepoint adaptation
6278 von Kries whitepoint adaptation
6281 identity whitepoint adaptation (i.e. no whitepoint adaptation)
6285 Override all input properties at once. Same accepted values as @ref{all}.
6288 Override input colorspace. Same accepted values as @ref{space}.
6291 Override input color primaries. Same accepted values as @ref{primaries}.
6294 Override input transfer characteristics. Same accepted values as @ref{trc}.
6297 Override input color range. Same accepted values as @ref{range}.
6301 The filter converts the transfer characteristics, color space and color
6302 primaries to the specified user values. The output value, if not specified,
6303 is set to a default value based on the "all" property. If that property is
6304 also not specified, the filter will log an error. The output color range and
6305 format default to the same value as the input color range and format. The
6306 input transfer characteristics, color space, color primaries and color range
6307 should be set on the input data. If any of these are missing, the filter will
6308 log an error and no conversion will take place.
6310 For example to convert the input to SMPTE-240M, use the command:
6312 colorspace=smpte240m
6315 @section convolution
6317 Apply convolution 3x3, 5x5 or 7x7 filter.
6319 The filter accepts the following options:
6326 Set matrix for each plane.
6327 Matrix is sequence of 9, 25 or 49 signed integers.
6333 Set multiplier for calculated value for each plane.
6339 Set bias for each plane. This value is added to the result of the multiplication.
6340 Useful for making the overall image brighter or darker. Default is 0.0.
6343 @subsection Examples
6349 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"
6355 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"
6361 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"
6367 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"
6371 Apply laplacian edge detector which includes diagonals:
6373 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"
6379 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"
6385 Apply 2D convolution of video stream in frequency domain using second stream
6388 The filter accepts the following options:
6392 Set which planes to process.
6395 Set which impulse video frames will be processed, can be @var{first}
6396 or @var{all}. Default is @var{all}.
6399 The @code{convolve} filter also supports the @ref{framesync} options.
6403 Copy the input video source unchanged to the output. This is mainly useful for
6408 Video filtering on GPU using Apple's CoreImage API on OSX.
6410 Hardware acceleration is based on an OpenGL context. Usually, this means it is
6411 processed by video hardware. However, software-based OpenGL implementations
6412 exist which means there is no guarantee for hardware processing. It depends on
6415 There are many filters and image generators provided by Apple that come with a
6416 large variety of options. The filter has to be referenced by its name along
6419 The coreimage filter accepts the following options:
6422 List all available filters and generators along with all their respective
6423 options as well as possible minimum and maximum values along with the default
6430 Specify all filters by their respective name and options.
6431 Use @var{list_filters} to determine all valid filter names and options.
6432 Numerical options are specified by a float value and are automatically clamped
6433 to their respective value range. Vector and color options have to be specified
6434 by a list of space separated float values. Character escaping has to be done.
6435 A special option name @code{default} is available to use default options for a
6438 It is required to specify either @code{default} or at least one of the filter options.
6439 All omitted options are used with their default values.
6440 The syntax of the filter string is as follows:
6442 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
6446 Specify a rectangle where the output of the filter chain is copied into the
6447 input image. It is given by a list of space separated float values:
6449 output_rect=x\ y\ width\ height
6451 If not given, the output rectangle equals the dimensions of the input image.
6452 The output rectangle is automatically cropped at the borders of the input
6453 image. Negative values are valid for each component.
6455 output_rect=25\ 25\ 100\ 100
6459 Several filters can be chained for successive processing without GPU-HOST
6460 transfers allowing for fast processing of complex filter chains.
6461 Currently, only filters with zero (generators) or exactly one (filters) input
6462 image and one output image are supported. Also, transition filters are not yet
6465 Some filters generate output images with additional padding depending on the
6466 respective filter kernel. The padding is automatically removed to ensure the
6467 filter output has the same size as the input image.
6469 For image generators, the size of the output image is determined by the
6470 previous output image of the filter chain or the input image of the whole
6471 filterchain, respectively. The generators do not use the pixel information of
6472 this image to generate their output. However, the generated output is
6473 blended onto this image, resulting in partial or complete coverage of the
6476 The @ref{coreimagesrc} video source can be used for generating input images
6477 which are directly fed into the filter chain. By using it, providing input
6478 images by another video source or an input video is not required.
6480 @subsection Examples
6485 List all filters available:
6487 coreimage=list_filters=true
6491 Use the CIBoxBlur filter with default options to blur an image:
6493 coreimage=filter=CIBoxBlur@@default
6497 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
6498 its center at 100x100 and a radius of 50 pixels:
6500 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
6504 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
6505 given as complete and escaped command-line for Apple's standard bash shell:
6507 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
6513 Crop the input video to given dimensions.
6515 It accepts the following parameters:
6519 The width of the output video. It defaults to @code{iw}.
6520 This expression is evaluated only once during the filter
6521 configuration, or when the @samp{w} or @samp{out_w} command is sent.
6524 The height of the output video. It defaults to @code{ih}.
6525 This expression is evaluated only once during the filter
6526 configuration, or when the @samp{h} or @samp{out_h} command is sent.
6529 The horizontal position, in the input video, of the left edge of the output
6530 video. It defaults to @code{(in_w-out_w)/2}.
6531 This expression is evaluated per-frame.
6534 The vertical position, in the input video, of the top edge of the output video.
6535 It defaults to @code{(in_h-out_h)/2}.
6536 This expression is evaluated per-frame.
6539 If set to 1 will force the output display aspect ratio
6540 to be the same of the input, by changing the output sample aspect
6541 ratio. It defaults to 0.
6544 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
6545 width/height/x/y as specified and will not be rounded to nearest smaller value.
6549 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
6550 expressions containing the following constants:
6555 The computed values for @var{x} and @var{y}. They are evaluated for
6560 The input width and height.
6564 These are the same as @var{in_w} and @var{in_h}.
6568 The output (cropped) width and height.
6572 These are the same as @var{out_w} and @var{out_h}.
6575 same as @var{iw} / @var{ih}
6578 input sample aspect ratio
6581 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
6585 horizontal and vertical chroma subsample values. For example for the
6586 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
6589 The number of the input frame, starting from 0.
6592 the position in the file of the input frame, NAN if unknown
6595 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
6599 The expression for @var{out_w} may depend on the value of @var{out_h},
6600 and the expression for @var{out_h} may depend on @var{out_w}, but they
6601 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
6602 evaluated after @var{out_w} and @var{out_h}.
6604 The @var{x} and @var{y} parameters specify the expressions for the
6605 position of the top-left corner of the output (non-cropped) area. They
6606 are evaluated for each frame. If the evaluated value is not valid, it
6607 is approximated to the nearest valid value.
6609 The expression for @var{x} may depend on @var{y}, and the expression
6610 for @var{y} may depend on @var{x}.
6612 @subsection Examples
6616 Crop area with size 100x100 at position (12,34).
6621 Using named options, the example above becomes:
6623 crop=w=100:h=100:x=12:y=34
6627 Crop the central input area with size 100x100:
6633 Crop the central input area with size 2/3 of the input video:
6635 crop=2/3*in_w:2/3*in_h
6639 Crop the input video central square:
6646 Delimit the rectangle with the top-left corner placed at position
6647 100:100 and the right-bottom corner corresponding to the right-bottom
6648 corner of the input image.
6650 crop=in_w-100:in_h-100:100:100
6654 Crop 10 pixels from the left and right borders, and 20 pixels from
6655 the top and bottom borders
6657 crop=in_w-2*10:in_h-2*20
6661 Keep only the bottom right quarter of the input image:
6663 crop=in_w/2:in_h/2:in_w/2:in_h/2
6667 Crop height for getting Greek harmony:
6669 crop=in_w:1/PHI*in_w
6673 Apply trembling effect:
6675 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)
6679 Apply erratic camera effect depending on timestamp:
6681 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)"
6685 Set x depending on the value of y:
6687 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
6691 @subsection Commands
6693 This filter supports the following commands:
6699 Set width/height of the output video and the horizontal/vertical position
6701 The command accepts the same syntax of the corresponding option.
6703 If the specified expression is not valid, it is kept at its current
6709 Auto-detect the crop size.
6711 It calculates the necessary cropping parameters and prints the
6712 recommended parameters via the logging system. The detected dimensions
6713 correspond to the non-black area of the input video.
6715 It accepts the following parameters:
6720 Set higher black value threshold, which can be optionally specified
6721 from nothing (0) to everything (255 for 8-bit based formats). An intensity
6722 value greater to the set value is considered non-black. It defaults to 24.
6723 You can also specify a value between 0.0 and 1.0 which will be scaled depending
6724 on the bitdepth of the pixel format.
6727 The value which the width/height should be divisible by. It defaults to
6728 16. The offset is automatically adjusted to center the video. Use 2 to
6729 get only even dimensions (needed for 4:2:2 video). 16 is best when
6730 encoding to most video codecs.
6732 @item reset_count, reset
6733 Set the counter that determines after how many frames cropdetect will
6734 reset the previously detected largest video area and start over to
6735 detect the current optimal crop area. Default value is 0.
6737 This can be useful when channel logos distort the video area. 0
6738 indicates 'never reset', and returns the largest area encountered during
6745 Apply color adjustments using curves.
6747 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
6748 component (red, green and blue) has its values defined by @var{N} key points
6749 tied from each other using a smooth curve. The x-axis represents the pixel
6750 values from the input frame, and the y-axis the new pixel values to be set for
6753 By default, a component curve is defined by the two points @var{(0;0)} and
6754 @var{(1;1)}. This creates a straight line where each original pixel value is
6755 "adjusted" to its own value, which means no change to the image.
6757 The filter allows you to redefine these two points and add some more. A new
6758 curve (using a natural cubic spline interpolation) will be define to pass
6759 smoothly through all these new coordinates. The new defined points needs to be
6760 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
6761 be in the @var{[0;1]} interval. If the computed curves happened to go outside
6762 the vector spaces, the values will be clipped accordingly.
6764 The filter accepts the following options:
6768 Select one of the available color presets. This option can be used in addition
6769 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
6770 options takes priority on the preset values.
6771 Available presets are:
6774 @item color_negative
6777 @item increase_contrast
6779 @item linear_contrast
6780 @item medium_contrast
6782 @item strong_contrast
6785 Default is @code{none}.
6787 Set the master key points. These points will define a second pass mapping. It
6788 is sometimes called a "luminance" or "value" mapping. It can be used with
6789 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
6790 post-processing LUT.
6792 Set the key points for the red component.
6794 Set the key points for the green component.
6796 Set the key points for the blue component.
6798 Set the key points for all components (not including master).
6799 Can be used in addition to the other key points component
6800 options. In this case, the unset component(s) will fallback on this
6801 @option{all} setting.
6803 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
6805 Save Gnuplot script of the curves in specified file.
6808 To avoid some filtergraph syntax conflicts, each key points list need to be
6809 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
6811 @subsection Examples
6815 Increase slightly the middle level of blue:
6817 curves=blue='0/0 0.5/0.58 1/1'
6823 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'
6825 Here we obtain the following coordinates for each components:
6828 @code{(0;0.11) (0.42;0.51) (1;0.95)}
6830 @code{(0;0) (0.50;0.48) (1;1)}
6832 @code{(0;0.22) (0.49;0.44) (1;0.80)}
6836 The previous example can also be achieved with the associated built-in preset:
6838 curves=preset=vintage
6848 Use a Photoshop preset and redefine the points of the green component:
6850 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
6854 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
6855 and @command{gnuplot}:
6857 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
6858 gnuplot -p /tmp/curves.plt
6864 Video data analysis filter.
6866 This filter shows hexadecimal pixel values of part of video.
6868 The filter accepts the following options:
6872 Set output video size.
6875 Set x offset from where to pick pixels.
6878 Set y offset from where to pick pixels.
6881 Set scope mode, can be one of the following:
6884 Draw hexadecimal pixel values with white color on black background.
6887 Draw hexadecimal pixel values with input video pixel color on black
6891 Draw hexadecimal pixel values on color background picked from input video,
6892 the text color is picked in such way so its always visible.
6896 Draw rows and columns numbers on left and top of video.
6899 Set background opacity.
6904 Denoise frames using 2D DCT (frequency domain filtering).
6906 This filter is not designed for real time.
6908 The filter accepts the following options:
6912 Set the noise sigma constant.
6914 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
6915 coefficient (absolute value) below this threshold with be dropped.
6917 If you need a more advanced filtering, see @option{expr}.
6919 Default is @code{0}.
6922 Set number overlapping pixels for each block. Since the filter can be slow, you
6923 may want to reduce this value, at the cost of a less effective filter and the
6924 risk of various artefacts.
6926 If the overlapping value doesn't permit processing the whole input width or
6927 height, a warning will be displayed and according borders won't be denoised.
6929 Default value is @var{blocksize}-1, which is the best possible setting.
6932 Set the coefficient factor expression.
6934 For each coefficient of a DCT block, this expression will be evaluated as a
6935 multiplier value for the coefficient.
6937 If this is option is set, the @option{sigma} option will be ignored.
6939 The absolute value of the coefficient can be accessed through the @var{c}
6943 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
6944 @var{blocksize}, which is the width and height of the processed blocks.
6946 The default value is @var{3} (8x8) and can be raised to @var{4} for a
6947 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
6948 on the speed processing. Also, a larger block size does not necessarily means a
6952 @subsection Examples
6954 Apply a denoise with a @option{sigma} of @code{4.5}:
6959 The same operation can be achieved using the expression system:
6961 dctdnoiz=e='gte(c, 4.5*3)'
6964 Violent denoise using a block size of @code{16x16}:
6971 Remove banding artifacts from input video.
6972 It works by replacing banded pixels with average value of referenced pixels.
6974 The filter accepts the following options:
6981 Set banding detection threshold for each plane. Default is 0.02.
6982 Valid range is 0.00003 to 0.5.
6983 If difference between current pixel and reference pixel is less than threshold,
6984 it will be considered as banded.
6987 Banding detection range in pixels. Default is 16. If positive, random number
6988 in range 0 to set value will be used. If negative, exact absolute value
6990 The range defines square of four pixels around current pixel.
6993 Set direction in radians from which four pixel will be compared. If positive,
6994 random direction from 0 to set direction will be picked. If negative, exact of
6995 absolute value will be picked. For example direction 0, -PI or -2*PI radians
6996 will pick only pixels on same row and -PI/2 will pick only pixels on same
7000 If enabled, current pixel is compared with average value of all four
7001 surrounding pixels. The default is enabled. If disabled current pixel is
7002 compared with all four surrounding pixels. The pixel is considered banded
7003 if only all four differences with surrounding pixels are less than threshold.
7006 If enabled, current pixel is changed if and only if all pixel components are banded,
7007 e.g. banding detection threshold is triggered for all color components.
7008 The default is disabled.
7014 Drop duplicated frames at regular intervals.
7016 The filter accepts the following options:
7020 Set the number of frames from which one will be dropped. Setting this to
7021 @var{N} means one frame in every batch of @var{N} frames will be dropped.
7022 Default is @code{5}.
7025 Set the threshold for duplicate detection. If the difference metric for a frame
7026 is less than or equal to this value, then it is declared as duplicate. Default
7030 Set scene change threshold. Default is @code{15}.
7034 Set the size of the x and y-axis blocks used during metric calculations.
7035 Larger blocks give better noise suppression, but also give worse detection of
7036 small movements. Must be a power of two. Default is @code{32}.
7039 Mark main input as a pre-processed input and activate clean source input
7040 stream. This allows the input to be pre-processed with various filters to help
7041 the metrics calculation while keeping the frame selection lossless. When set to
7042 @code{1}, the first stream is for the pre-processed input, and the second
7043 stream is the clean source from where the kept frames are chosen. Default is
7047 Set whether or not chroma is considered in the metric calculations. Default is
7053 Apply 2D deconvolution of video stream in frequency domain using second stream
7056 The filter accepts the following options:
7060 Set which planes to process.
7063 Set which impulse video frames will be processed, can be @var{first}
7064 or @var{all}. Default is @var{all}.
7067 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
7068 and height are not same and not power of 2 or if stream prior to convolving
7072 The @code{deconvolve} filter also supports the @ref{framesync} options.
7076 Apply deflate effect to the video.
7078 This filter replaces the pixel by the local(3x3) average by taking into account
7079 only values lower than the pixel.
7081 It accepts the following options:
7088 Limit the maximum change for each plane, default is 65535.
7089 If 0, plane will remain unchanged.
7094 Remove temporal frame luminance variations.
7096 It accepts the following options:
7100 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
7103 Set averaging mode to smooth temporal luminance variations.
7105 Available values are:
7130 Do not actually modify frame. Useful when one only wants metadata.
7135 Remove judder produced by partially interlaced telecined content.
7137 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
7138 source was partially telecined content then the output of @code{pullup,dejudder}
7139 will have a variable frame rate. May change the recorded frame rate of the
7140 container. Aside from that change, this filter will not affect constant frame
7143 The option available in this filter is:
7147 Specify the length of the window over which the judder repeats.
7149 Accepts any integer greater than 1. Useful values are:
7153 If the original was telecined from 24 to 30 fps (Film to NTSC).
7156 If the original was telecined from 25 to 30 fps (PAL to NTSC).
7159 If a mixture of the two.
7162 The default is @samp{4}.
7167 Suppress a TV station logo by a simple interpolation of the surrounding
7168 pixels. Just set a rectangle covering the logo and watch it disappear
7169 (and sometimes something even uglier appear - your mileage may vary).
7171 It accepts the following parameters:
7176 Specify the top left corner coordinates of the logo. They must be
7181 Specify the width and height of the logo to clear. They must be
7185 Specify the thickness of the fuzzy edge of the rectangle (added to
7186 @var{w} and @var{h}). The default value is 1. This option is
7187 deprecated, setting higher values should no longer be necessary and
7191 When set to 1, a green rectangle is drawn on the screen to simplify
7192 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
7193 The default value is 0.
7195 The rectangle is drawn on the outermost pixels which will be (partly)
7196 replaced with interpolated values. The values of the next pixels
7197 immediately outside this rectangle in each direction will be used to
7198 compute the interpolated pixel values inside the rectangle.
7202 @subsection Examples
7206 Set a rectangle covering the area with top left corner coordinates 0,0
7207 and size 100x77, and a band of size 10:
7209 delogo=x=0:y=0:w=100:h=77:band=10
7216 Attempt to fix small changes in horizontal and/or vertical shift. This
7217 filter helps remove camera shake from hand-holding a camera, bumping a
7218 tripod, moving on a vehicle, etc.
7220 The filter accepts the following options:
7228 Specify a rectangular area where to limit the search for motion
7230 If desired the search for motion vectors can be limited to a
7231 rectangular area of the frame defined by its top left corner, width
7232 and height. These parameters have the same meaning as the drawbox
7233 filter which can be used to visualise the position of the bounding
7236 This is useful when simultaneous movement of subjects within the frame
7237 might be confused for camera motion by the motion vector search.
7239 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
7240 then the full frame is used. This allows later options to be set
7241 without specifying the bounding box for the motion vector search.
7243 Default - search the whole frame.
7247 Specify the maximum extent of movement in x and y directions in the
7248 range 0-64 pixels. Default 16.
7251 Specify how to generate pixels to fill blanks at the edge of the
7252 frame. Available values are:
7255 Fill zeroes at blank locations
7257 Original image at blank locations
7259 Extruded edge value at blank locations
7261 Mirrored edge at blank locations
7263 Default value is @samp{mirror}.
7266 Specify the blocksize to use for motion search. Range 4-128 pixels,
7270 Specify the contrast threshold for blocks. Only blocks with more than
7271 the specified contrast (difference between darkest and lightest
7272 pixels) will be considered. Range 1-255, default 125.
7275 Specify the search strategy. Available values are:
7278 Set exhaustive search
7280 Set less exhaustive search.
7282 Default value is @samp{exhaustive}.
7285 If set then a detailed log of the motion search is written to the
7292 Remove unwanted contamination of foreground colors, caused by reflected color of
7293 greenscreen or bluescreen.
7295 This filter accepts the following options:
7299 Set what type of despill to use.
7302 Set how spillmap will be generated.
7305 Set how much to get rid of still remaining spill.
7308 Controls amount of red in spill area.
7311 Controls amount of green in spill area.
7312 Should be -1 for greenscreen.
7315 Controls amount of blue in spill area.
7316 Should be -1 for bluescreen.
7319 Controls brightness of spill area, preserving colors.
7322 Modify alpha from generated spillmap.
7327 Apply an exact inverse of the telecine operation. It requires a predefined
7328 pattern specified using the pattern option which must be the same as that passed
7329 to the telecine filter.
7331 This filter accepts the following options:
7340 The default value is @code{top}.
7344 A string of numbers representing the pulldown pattern you wish to apply.
7345 The default value is @code{23}.
7348 A number representing position of the first frame with respect to the telecine
7349 pattern. This is to be used if the stream is cut. The default value is @code{0}.
7354 Apply dilation effect to the video.
7356 This filter replaces the pixel by the local(3x3) maximum.
7358 It accepts the following options:
7365 Limit the maximum change for each plane, default is 65535.
7366 If 0, plane will remain unchanged.
7369 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
7372 Flags to local 3x3 coordinates maps like this:
7381 Displace pixels as indicated by second and third input stream.
7383 It takes three input streams and outputs one stream, the first input is the
7384 source, and second and third input are displacement maps.
7386 The second input specifies how much to displace pixels along the
7387 x-axis, while the third input specifies how much to displace pixels
7389 If one of displacement map streams terminates, last frame from that
7390 displacement map will be used.
7392 Note that once generated, displacements maps can be reused over and over again.
7394 A description of the accepted options follows.
7398 Set displace behavior for pixels that are out of range.
7400 Available values are:
7403 Missing pixels are replaced by black pixels.
7406 Adjacent pixels will spread out to replace missing pixels.
7409 Out of range pixels are wrapped so they point to pixels of other side.
7412 Out of range pixels will be replaced with mirrored pixels.
7414 Default is @samp{smear}.
7418 @subsection Examples
7422 Add ripple effect to rgb input of video size hd720:
7424 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
7428 Add wave effect to rgb input of video size hd720:
7430 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
7436 Draw a colored box on the input image.
7438 It accepts the following parameters:
7443 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
7447 The expressions which specify the width and height of the box; if 0 they are interpreted as
7448 the input width and height. It defaults to 0.
7451 Specify the color of the box to write. For the general syntax of this option,
7452 check the "Color" section in the ffmpeg-utils manual. If the special
7453 value @code{invert} is used, the box edge color is the same as the
7454 video with inverted luma.
7457 The expression which sets the thickness of the box edge.
7458 A value of @code{fill} will create a filled box. Default value is @code{3}.
7460 See below for the list of accepted constants.
7463 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
7464 will overwrite the video's color and alpha pixels.
7465 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
7468 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
7469 following constants:
7473 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
7477 horizontal and vertical chroma subsample values. For example for the
7478 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
7482 The input width and height.
7485 The input sample aspect ratio.
7489 The x and y offset coordinates where the box is drawn.
7493 The width and height of the drawn box.
7496 The thickness of the drawn box.
7498 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
7499 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
7503 @subsection Examples
7507 Draw a black box around the edge of the input image:
7513 Draw a box with color red and an opacity of 50%:
7515 drawbox=10:20:200:60:red@@0.5
7518 The previous example can be specified as:
7520 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
7524 Fill the box with pink color:
7526 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
7530 Draw a 2-pixel red 2.40:1 mask:
7532 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
7538 Draw a grid on the input image.
7540 It accepts the following parameters:
7545 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
7549 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
7550 input width and height, respectively, minus @code{thickness}, so image gets
7551 framed. Default to 0.
7554 Specify the color of the grid. For the general syntax of this option,
7555 check the "Color" section in the ffmpeg-utils manual. If the special
7556 value @code{invert} is used, the grid color is the same as the
7557 video with inverted luma.
7560 The expression which sets the thickness of the grid line. Default value is @code{1}.
7562 See below for the list of accepted constants.
7565 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
7566 will overwrite the video's color and alpha pixels.
7567 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
7570 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
7571 following constants:
7575 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
7579 horizontal and vertical chroma subsample values. For example for the
7580 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
7584 The input grid cell width and height.
7587 The input sample aspect ratio.
7591 The x and y coordinates of some point of grid intersection (meant to configure offset).
7595 The width and height of the drawn cell.
7598 The thickness of the drawn cell.
7600 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
7601 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
7605 @subsection Examples
7609 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
7611 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
7615 Draw a white 3x3 grid with an opacity of 50%:
7617 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
7624 Draw a text string or text from a specified file on top of a video, using the
7625 libfreetype library.
7627 To enable compilation of this filter, you need to configure FFmpeg with
7628 @code{--enable-libfreetype}.
7629 To enable default font fallback and the @var{font} option you need to
7630 configure FFmpeg with @code{--enable-libfontconfig}.
7631 To enable the @var{text_shaping} option, you need to configure FFmpeg with
7632 @code{--enable-libfribidi}.
7636 It accepts the following parameters:
7641 Used to draw a box around text using the background color.
7642 The value must be either 1 (enable) or 0 (disable).
7643 The default value of @var{box} is 0.
7646 Set the width of the border to be drawn around the box using @var{boxcolor}.
7647 The default value of @var{boxborderw} is 0.
7650 The color to be used for drawing box around text. For the syntax of this
7651 option, check the "Color" section in the ffmpeg-utils manual.
7653 The default value of @var{boxcolor} is "white".
7656 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
7657 The default value of @var{line_spacing} is 0.
7660 Set the width of the border to be drawn around the text using @var{bordercolor}.
7661 The default value of @var{borderw} is 0.
7664 Set the color to be used for drawing border around text. For the syntax of this
7665 option, check the "Color" section in the ffmpeg-utils manual.
7667 The default value of @var{bordercolor} is "black".
7670 Select how the @var{text} is expanded. Can be either @code{none},
7671 @code{strftime} (deprecated) or
7672 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
7676 Set a start time for the count. Value is in microseconds. Only applied
7677 in the deprecated strftime expansion mode. To emulate in normal expansion
7678 mode use the @code{pts} function, supplying the start time (in seconds)
7679 as the second argument.
7682 If true, check and fix text coords to avoid clipping.
7685 The color to be used for drawing fonts. For the syntax of this option, check
7686 the "Color" section in the ffmpeg-utils manual.
7688 The default value of @var{fontcolor} is "black".
7690 @item fontcolor_expr
7691 String which is expanded the same way as @var{text} to obtain dynamic
7692 @var{fontcolor} value. By default this option has empty value and is not
7693 processed. When this option is set, it overrides @var{fontcolor} option.
7696 The font family to be used for drawing text. By default Sans.
7699 The font file to be used for drawing text. The path must be included.
7700 This parameter is mandatory if the fontconfig support is disabled.
7703 Draw the text applying alpha blending. The value can
7704 be a number between 0.0 and 1.0.
7705 The expression accepts the same variables @var{x, y} as well.
7706 The default value is 1.
7707 Please see @var{fontcolor_expr}.
7710 The font size to be used for drawing text.
7711 The default value of @var{fontsize} is 16.
7714 If set to 1, attempt to shape the text (for example, reverse the order of
7715 right-to-left text and join Arabic characters) before drawing it.
7716 Otherwise, just draw the text exactly as given.
7717 By default 1 (if supported).
7720 The flags to be used for loading the fonts.
7722 The flags map the corresponding flags supported by libfreetype, and are
7723 a combination of the following values:
7730 @item vertical_layout
7731 @item force_autohint
7734 @item ignore_global_advance_width
7736 @item ignore_transform
7742 Default value is "default".
7744 For more information consult the documentation for the FT_LOAD_*
7748 The color to be used for drawing a shadow behind the drawn text. For the
7749 syntax of this option, check the "Color" section in the ffmpeg-utils manual.
7751 The default value of @var{shadowcolor} is "black".
7755 The x and y offsets for the text shadow position with respect to the
7756 position of the text. They can be either positive or negative
7757 values. The default value for both is "0".
7760 The starting frame number for the n/frame_num variable. The default value
7764 The size in number of spaces to use for rendering the tab.
7768 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
7769 format. It can be used with or without text parameter. @var{timecode_rate}
7770 option must be specified.
7772 @item timecode_rate, rate, r
7773 Set the timecode frame rate (timecode only). Value will be rounded to nearest
7774 integer. Minimum value is "1".
7775 Drop-frame timecode is supported for frame rates 30 & 60.
7778 If set to 1, the output of the timecode option will wrap around at 24 hours.
7779 Default is 0 (disabled).
7782 The text string to be drawn. The text must be a sequence of UTF-8
7784 This parameter is mandatory if no file is specified with the parameter
7788 A text file containing text to be drawn. The text must be a sequence
7789 of UTF-8 encoded characters.
7791 This parameter is mandatory if no text string is specified with the
7792 parameter @var{text}.
7794 If both @var{text} and @var{textfile} are specified, an error is thrown.
7797 If set to 1, the @var{textfile} will be reloaded before each frame.
7798 Be sure to update it atomically, or it may be read partially, or even fail.
7802 The expressions which specify the offsets where text will be drawn
7803 within the video frame. They are relative to the top/left border of the
7806 The default value of @var{x} and @var{y} is "0".
7808 See below for the list of accepted constants and functions.
7811 The parameters for @var{x} and @var{y} are expressions containing the
7812 following constants and functions:
7816 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
7820 horizontal and vertical chroma subsample values. For example for the
7821 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
7824 the height of each text line
7832 @item max_glyph_a, ascent
7833 the maximum distance from the baseline to the highest/upper grid
7834 coordinate used to place a glyph outline point, for all the rendered
7836 It is a positive value, due to the grid's orientation with the Y axis
7839 @item max_glyph_d, descent
7840 the maximum distance from the baseline to the lowest grid coordinate
7841 used to place a glyph outline point, for all the rendered glyphs.
7842 This is a negative value, due to the grid's orientation, with the Y axis
7846 maximum glyph height, that is the maximum height for all the glyphs
7847 contained in the rendered text, it is equivalent to @var{ascent} -
7851 maximum glyph width, that is the maximum width for all the glyphs
7852 contained in the rendered text
7855 the number of input frame, starting from 0
7857 @item rand(min, max)
7858 return a random number included between @var{min} and @var{max}
7861 The input sample aspect ratio.
7864 timestamp expressed in seconds, NAN if the input timestamp is unknown
7867 the height of the rendered text
7870 the width of the rendered text
7874 the x and y offset coordinates where the text is drawn.
7876 These parameters allow the @var{x} and @var{y} expressions to refer
7877 each other, so you can for example specify @code{y=x/dar}.
7880 @anchor{drawtext_expansion}
7881 @subsection Text expansion
7883 If @option{expansion} is set to @code{strftime},
7884 the filter recognizes strftime() sequences in the provided text and
7885 expands them accordingly. Check the documentation of strftime(). This
7886 feature is deprecated.
7888 If @option{expansion} is set to @code{none}, the text is printed verbatim.
7890 If @option{expansion} is set to @code{normal} (which is the default),
7891 the following expansion mechanism is used.
7893 The backslash character @samp{\}, followed by any character, always expands to
7894 the second character.
7896 Sequences of the form @code{%@{...@}} are expanded. The text between the
7897 braces is a function name, possibly followed by arguments separated by ':'.
7898 If the arguments contain special characters or delimiters (':' or '@}'),
7899 they should be escaped.
7901 Note that they probably must also be escaped as the value for the
7902 @option{text} option in the filter argument string and as the filter
7903 argument in the filtergraph description, and possibly also for the shell,
7904 that makes up to four levels of escaping; using a text file avoids these
7907 The following functions are available:
7912 The expression evaluation result.
7914 It must take one argument specifying the expression to be evaluated,
7915 which accepts the same constants and functions as the @var{x} and
7916 @var{y} values. Note that not all constants should be used, for
7917 example the text size is not known when evaluating the expression, so
7918 the constants @var{text_w} and @var{text_h} will have an undefined
7921 @item expr_int_format, eif
7922 Evaluate the expression's value and output as formatted integer.
7924 The first argument is the expression to be evaluated, just as for the @var{expr} function.
7925 The second argument specifies the output format. Allowed values are @samp{x},
7926 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
7927 @code{printf} function.
7928 The third parameter is optional and sets the number of positions taken by the output.
7929 It can be used to add padding with zeros from the left.
7932 The time at which the filter is running, expressed in UTC.
7933 It can accept an argument: a strftime() format string.
7936 The time at which the filter is running, expressed in the local time zone.
7937 It can accept an argument: a strftime() format string.
7940 Frame metadata. Takes one or two arguments.
7942 The first argument is mandatory and specifies the metadata key.
7944 The second argument is optional and specifies a default value, used when the
7945 metadata key is not found or empty.
7948 The frame number, starting from 0.
7951 A 1 character description of the current picture type.
7954 The timestamp of the current frame.
7955 It can take up to three arguments.
7957 The first argument is the format of the timestamp; it defaults to @code{flt}
7958 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
7959 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
7960 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
7961 @code{localtime} stands for the timestamp of the frame formatted as
7962 local time zone time.
7964 The second argument is an offset added to the timestamp.
7966 If the format is set to @code{localtime} or @code{gmtime},
7967 a third argument may be supplied: a strftime() format string.
7968 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
7971 @subsection Examples
7975 Draw "Test Text" with font FreeSerif, using the default values for the
7976 optional parameters.
7979 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
7983 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
7984 and y=50 (counting from the top-left corner of the screen), text is
7985 yellow with a red box around it. Both the text and the box have an
7989 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
7990 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
7993 Note that the double quotes are not necessary if spaces are not used
7994 within the parameter list.
7997 Show the text at the center of the video frame:
7999 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
8003 Show the text at a random position, switching to a new position every 30 seconds:
8005 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)"
8009 Show a text line sliding from right to left in the last row of the video
8010 frame. The file @file{LONG_LINE} is assumed to contain a single line
8013 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
8017 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
8019 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
8023 Draw a single green letter "g", at the center of the input video.
8024 The glyph baseline is placed at half screen height.
8026 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
8030 Show text for 1 second every 3 seconds:
8032 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
8036 Use fontconfig to set the font. Note that the colons need to be escaped.
8038 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
8042 Print the date of a real-time encoding (see strftime(3)):
8044 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
8048 Show text fading in and out (appearing/disappearing):
8051 DS=1.0 # display start
8052 DE=10.0 # display end
8053 FID=1.5 # fade in duration
8054 FOD=5 # fade out duration
8055 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 @}"
8059 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
8060 and the @option{fontsize} value are included in the @option{y} offset.
8062 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
8063 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
8068 For more information about libfreetype, check:
8069 @url{http://www.freetype.org/}.
8071 For more information about fontconfig, check:
8072 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
8074 For more information about libfribidi, check:
8075 @url{http://fribidi.org/}.
8079 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
8081 The filter accepts the following options:
8086 Set low and high threshold values used by the Canny thresholding
8089 The high threshold selects the "strong" edge pixels, which are then
8090 connected through 8-connectivity with the "weak" edge pixels selected
8091 by the low threshold.
8093 @var{low} and @var{high} threshold values must be chosen in the range
8094 [0,1], and @var{low} should be lesser or equal to @var{high}.
8096 Default value for @var{low} is @code{20/255}, and default value for @var{high}
8100 Define the drawing mode.
8104 Draw white/gray wires on black background.
8107 Mix the colors to create a paint/cartoon effect.
8110 Default value is @var{wires}.
8113 @subsection Examples
8117 Standard edge detection with custom values for the hysteresis thresholding:
8119 edgedetect=low=0.1:high=0.4
8123 Painting effect without thresholding:
8125 edgedetect=mode=colormix:high=0
8130 Set brightness, contrast, saturation and approximate gamma adjustment.
8132 The filter accepts the following options:
8136 Set the contrast expression. The value must be a float value in range
8137 @code{-2.0} to @code{2.0}. The default value is "1".
8140 Set the brightness expression. The value must be a float value in
8141 range @code{-1.0} to @code{1.0}. The default value is "0".
8144 Set the saturation expression. The value must be a float in
8145 range @code{0.0} to @code{3.0}. The default value is "1".
8148 Set the gamma expression. The value must be a float in range
8149 @code{0.1} to @code{10.0}. The default value is "1".
8152 Set the gamma expression for red. The value must be a float in
8153 range @code{0.1} to @code{10.0}. The default value is "1".
8156 Set the gamma expression for green. The value must be a float in range
8157 @code{0.1} to @code{10.0}. The default value is "1".
8160 Set the gamma expression for blue. The value must be a float in range
8161 @code{0.1} to @code{10.0}. The default value is "1".
8164 Set the gamma weight expression. It can be used to reduce the effect
8165 of a high gamma value on bright image areas, e.g. keep them from
8166 getting overamplified and just plain white. The value must be a float
8167 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
8168 gamma correction all the way down while @code{1.0} leaves it at its
8169 full strength. Default is "1".
8172 Set when the expressions for brightness, contrast, saturation and
8173 gamma expressions are evaluated.
8175 It accepts the following values:
8178 only evaluate expressions once during the filter initialization or
8179 when a command is processed
8182 evaluate expressions for each incoming frame
8185 Default value is @samp{init}.
8188 The expressions accept the following parameters:
8191 frame count of the input frame starting from 0
8194 byte position of the corresponding packet in the input file, NAN if
8198 frame rate of the input video, NAN if the input frame rate is unknown
8201 timestamp expressed in seconds, NAN if the input timestamp is unknown
8204 @subsection Commands
8205 The filter supports the following commands:
8209 Set the contrast expression.
8212 Set the brightness expression.
8215 Set the saturation expression.
8218 Set the gamma expression.
8221 Set the gamma_r expression.
8224 Set gamma_g expression.
8227 Set gamma_b expression.
8230 Set gamma_weight expression.
8232 The command accepts the same syntax of the corresponding option.
8234 If the specified expression is not valid, it is kept at its current
8241 Apply erosion effect to the video.
8243 This filter replaces the pixel by the local(3x3) minimum.
8245 It accepts the following options:
8252 Limit the maximum change for each plane, default is 65535.
8253 If 0, plane will remain unchanged.
8256 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
8259 Flags to local 3x3 coordinates maps like this:
8266 @section extractplanes
8268 Extract color channel components from input video stream into
8269 separate grayscale video streams.
8271 The filter accepts the following option:
8275 Set plane(s) to extract.
8277 Available values for planes are:
8288 Choosing planes not available in the input will result in an error.
8289 That means you cannot select @code{r}, @code{g}, @code{b} planes
8290 with @code{y}, @code{u}, @code{v} planes at same time.
8293 @subsection Examples
8297 Extract luma, u and v color channel component from input video frame
8298 into 3 grayscale outputs:
8300 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
8306 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
8308 For each input image, the filter will compute the optimal mapping from
8309 the input to the output given the codebook length, that is the number
8310 of distinct output colors.
8312 This filter accepts the following options.
8315 @item codebook_length, l
8316 Set codebook length. The value must be a positive integer, and
8317 represents the number of distinct output colors. Default value is 256.
8320 Set the maximum number of iterations to apply for computing the optimal
8321 mapping. The higher the value the better the result and the higher the
8322 computation time. Default value is 1.
8325 Set a random seed, must be an integer included between 0 and
8326 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
8327 will try to use a good random seed on a best effort basis.
8330 Set pal8 output pixel format. This option does not work with codebook
8331 length greater than 256.
8336 Measure graylevel entropy in histogram of color channels of video frames.
8338 It accepts the following parameters:
8342 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
8344 @var{diff} mode measures entropy of histogram delta values, absolute differences
8345 between neighbour histogram values.
8350 Apply a fade-in/out effect to the input video.
8352 It accepts the following parameters:
8356 The effect type can be either "in" for a fade-in, or "out" for a fade-out
8358 Default is @code{in}.
8360 @item start_frame, s
8361 Specify the number of the frame to start applying the fade
8362 effect at. Default is 0.
8365 The number of frames that the fade effect lasts. At the end of the
8366 fade-in effect, the output video will have the same intensity as the input video.
8367 At the end of the fade-out transition, the output video will be filled with the
8368 selected @option{color}.
8372 If set to 1, fade only alpha channel, if one exists on the input.
8375 @item start_time, st
8376 Specify the timestamp (in seconds) of the frame to start to apply the fade
8377 effect. If both start_frame and start_time are specified, the fade will start at
8378 whichever comes last. Default is 0.
8381 The number of seconds for which the fade effect has to last. At the end of the
8382 fade-in effect the output video will have the same intensity as the input video,
8383 at the end of the fade-out transition the output video will be filled with the
8384 selected @option{color}.
8385 If both duration and nb_frames are specified, duration is used. Default is 0
8386 (nb_frames is used by default).
8389 Specify the color of the fade. Default is "black".
8392 @subsection Examples
8396 Fade in the first 30 frames of video:
8401 The command above is equivalent to:
8407 Fade out the last 45 frames of a 200-frame video:
8410 fade=type=out:start_frame=155:nb_frames=45
8414 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
8416 fade=in:0:25, fade=out:975:25
8420 Make the first 5 frames yellow, then fade in from frame 5-24:
8422 fade=in:5:20:color=yellow
8426 Fade in alpha over first 25 frames of video:
8428 fade=in:0:25:alpha=1
8432 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
8434 fade=t=in:st=5.5:d=0.5
8440 Apply arbitrary expressions to samples in frequency domain
8444 Adjust the dc value (gain) of the luma plane of the image. The filter
8445 accepts an integer value in range @code{0} to @code{1000}. The default
8446 value is set to @code{0}.
8449 Adjust the dc value (gain) of the 1st chroma plane of the image. The
8450 filter accepts an integer value in range @code{0} to @code{1000}. The
8451 default value is set to @code{0}.
8454 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
8455 filter accepts an integer value in range @code{0} to @code{1000}. The
8456 default value is set to @code{0}.
8459 Set the frequency domain weight expression for the luma plane.
8462 Set the frequency domain weight expression for the 1st chroma plane.
8465 Set the frequency domain weight expression for the 2nd chroma plane.
8468 Set when the expressions are evaluated.
8470 It accepts the following values:
8473 Only evaluate expressions once during the filter initialization.
8476 Evaluate expressions for each incoming frame.
8479 Default value is @samp{init}.
8481 The filter accepts the following variables:
8484 The coordinates of the current sample.
8488 The width and height of the image.
8491 The number of input frame, starting from 0.
8494 @subsection Examples
8500 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
8506 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
8512 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
8518 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
8525 Extract a single field from an interlaced image using stride
8526 arithmetic to avoid wasting CPU time. The output frames are marked as
8529 The filter accepts the following options:
8533 Specify whether to extract the top (if the value is @code{0} or
8534 @code{top}) or the bottom field (if the value is @code{1} or
8540 Create new frames by copying the top and bottom fields from surrounding frames
8541 supplied as numbers by the hint file.
8545 Set file containing hints: absolute/relative frame numbers.
8547 There must be one line for each frame in a clip. Each line must contain two
8548 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
8549 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
8550 is current frame number for @code{absolute} mode or out of [-1, 1] range
8551 for @code{relative} mode. First number tells from which frame to pick up top
8552 field and second number tells from which frame to pick up bottom field.
8554 If optionally followed by @code{+} output frame will be marked as interlaced,
8555 else if followed by @code{-} output frame will be marked as progressive, else
8556 it will be marked same as input frame.
8557 If line starts with @code{#} or @code{;} that line is skipped.
8560 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
8563 Example of first several lines of @code{hint} file for @code{relative} mode:
8566 1,0 - # second frame, use third's frame top field and second's frame bottom field
8567 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
8584 Field matching filter for inverse telecine. It is meant to reconstruct the
8585 progressive frames from a telecined stream. The filter does not drop duplicated
8586 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
8587 followed by a decimation filter such as @ref{decimate} in the filtergraph.
8589 The separation of the field matching and the decimation is notably motivated by
8590 the possibility of inserting a de-interlacing filter fallback between the two.
8591 If the source has mixed telecined and real interlaced content,
8592 @code{fieldmatch} will not be able to match fields for the interlaced parts.
8593 But these remaining combed frames will be marked as interlaced, and thus can be
8594 de-interlaced by a later filter such as @ref{yadif} before decimation.
8596 In addition to the various configuration options, @code{fieldmatch} can take an
8597 optional second stream, activated through the @option{ppsrc} option. If
8598 enabled, the frames reconstruction will be based on the fields and frames from
8599 this second stream. This allows the first input to be pre-processed in order to
8600 help the various algorithms of the filter, while keeping the output lossless
8601 (assuming the fields are matched properly). Typically, a field-aware denoiser,
8602 or brightness/contrast adjustments can help.
8604 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
8605 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
8606 which @code{fieldmatch} is based on. While the semantic and usage are very
8607 close, some behaviour and options names can differ.
8609 The @ref{decimate} filter currently only works for constant frame rate input.
8610 If your input has mixed telecined (30fps) and progressive content with a lower
8611 framerate like 24fps use the following filterchain to produce the necessary cfr
8612 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
8614 The filter accepts the following options:
8618 Specify the assumed field order of the input stream. Available values are:
8622 Auto detect parity (use FFmpeg's internal parity value).
8624 Assume bottom field first.
8626 Assume top field first.
8629 Note that it is sometimes recommended not to trust the parity announced by the
8632 Default value is @var{auto}.
8635 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
8636 sense that it won't risk creating jerkiness due to duplicate frames when
8637 possible, but if there are bad edits or blended fields it will end up
8638 outputting combed frames when a good match might actually exist. On the other
8639 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
8640 but will almost always find a good frame if there is one. The other values are
8641 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
8642 jerkiness and creating duplicate frames versus finding good matches in sections
8643 with bad edits, orphaned fields, blended fields, etc.
8645 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
8647 Available values are:
8651 2-way matching (p/c)
8653 2-way matching, and trying 3rd match if still combed (p/c + n)
8655 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
8657 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
8658 still combed (p/c + n + u/b)
8660 3-way matching (p/c/n)
8662 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
8663 detected as combed (p/c/n + u/b)
8666 The parenthesis at the end indicate the matches that would be used for that
8667 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
8670 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
8673 Default value is @var{pc_n}.
8676 Mark the main input stream as a pre-processed input, and enable the secondary
8677 input stream as the clean source to pick the fields from. See the filter
8678 introduction for more details. It is similar to the @option{clip2} feature from
8681 Default value is @code{0} (disabled).
8684 Set the field to match from. It is recommended to set this to the same value as
8685 @option{order} unless you experience matching failures with that setting. In
8686 certain circumstances changing the field that is used to match from can have a
8687 large impact on matching performance. Available values are:
8691 Automatic (same value as @option{order}).
8693 Match from the bottom field.
8695 Match from the top field.
8698 Default value is @var{auto}.
8701 Set whether or not chroma is included during the match comparisons. In most
8702 cases it is recommended to leave this enabled. You should set this to @code{0}
8703 only if your clip has bad chroma problems such as heavy rainbowing or other
8704 artifacts. Setting this to @code{0} could also be used to speed things up at
8705 the cost of some accuracy.
8707 Default value is @code{1}.
8711 These define an exclusion band which excludes the lines between @option{y0} and
8712 @option{y1} from being included in the field matching decision. An exclusion
8713 band can be used to ignore subtitles, a logo, or other things that may
8714 interfere with the matching. @option{y0} sets the starting scan line and
8715 @option{y1} sets the ending line; all lines in between @option{y0} and
8716 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
8717 @option{y0} and @option{y1} to the same value will disable the feature.
8718 @option{y0} and @option{y1} defaults to @code{0}.
8721 Set the scene change detection threshold as a percentage of maximum change on
8722 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
8723 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
8724 @option{scthresh} is @code{[0.0, 100.0]}.
8726 Default value is @code{12.0}.
8729 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
8730 account the combed scores of matches when deciding what match to use as the
8731 final match. Available values are:
8735 No final matching based on combed scores.
8737 Combed scores are only used when a scene change is detected.
8739 Use combed scores all the time.
8742 Default is @var{sc}.
8745 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
8746 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
8747 Available values are:
8751 No forced calculation.
8753 Force p/c/n calculations.
8755 Force p/c/n/u/b calculations.
8758 Default value is @var{none}.
8761 This is the area combing threshold used for combed frame detection. This
8762 essentially controls how "strong" or "visible" combing must be to be detected.
8763 Larger values mean combing must be more visible and smaller values mean combing
8764 can be less visible or strong and still be detected. Valid settings are from
8765 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
8766 be detected as combed). This is basically a pixel difference value. A good
8767 range is @code{[8, 12]}.
8769 Default value is @code{9}.
8772 Sets whether or not chroma is considered in the combed frame decision. Only
8773 disable this if your source has chroma problems (rainbowing, etc.) that are
8774 causing problems for the combed frame detection with chroma enabled. Actually,
8775 using @option{chroma}=@var{0} is usually more reliable, except for the case
8776 where there is chroma only combing in the source.
8778 Default value is @code{0}.
8782 Respectively set the x-axis and y-axis size of the window used during combed
8783 frame detection. This has to do with the size of the area in which
8784 @option{combpel} pixels are required to be detected as combed for a frame to be
8785 declared combed. See the @option{combpel} parameter description for more info.
8786 Possible values are any number that is a power of 2 starting at 4 and going up
8789 Default value is @code{16}.
8792 The number of combed pixels inside any of the @option{blocky} by
8793 @option{blockx} size blocks on the frame for the frame to be detected as
8794 combed. While @option{cthresh} controls how "visible" the combing must be, this
8795 setting controls "how much" combing there must be in any localized area (a
8796 window defined by the @option{blockx} and @option{blocky} settings) on the
8797 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
8798 which point no frames will ever be detected as combed). This setting is known
8799 as @option{MI} in TFM/VFM vocabulary.
8801 Default value is @code{80}.
8804 @anchor{p/c/n/u/b meaning}
8805 @subsection p/c/n/u/b meaning
8807 @subsubsection p/c/n
8809 We assume the following telecined stream:
8812 Top fields: 1 2 2 3 4
8813 Bottom fields: 1 2 3 4 4
8816 The numbers correspond to the progressive frame the fields relate to. Here, the
8817 first two frames are progressive, the 3rd and 4th are combed, and so on.
8819 When @code{fieldmatch} is configured to run a matching from bottom
8820 (@option{field}=@var{bottom}) this is how this input stream get transformed:
8825 B 1 2 3 4 4 <-- matching reference
8834 As a result of the field matching, we can see that some frames get duplicated.
8835 To perform a complete inverse telecine, you need to rely on a decimation filter
8836 after this operation. See for instance the @ref{decimate} filter.
8838 The same operation now matching from top fields (@option{field}=@var{top})
8843 T 1 2 2 3 4 <-- matching reference
8853 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
8854 basically, they refer to the frame and field of the opposite parity:
8857 @item @var{p} matches the field of the opposite parity in the previous frame
8858 @item @var{c} matches the field of the opposite parity in the current frame
8859 @item @var{n} matches the field of the opposite parity in the next frame
8864 The @var{u} and @var{b} matching are a bit special in the sense that they match
8865 from the opposite parity flag. In the following examples, we assume that we are
8866 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
8867 'x' is placed above and below each matched fields.
8869 With bottom matching (@option{field}=@var{bottom}):
8874 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
8875 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
8883 With top matching (@option{field}=@var{top}):
8888 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
8889 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
8897 @subsection Examples
8899 Simple IVTC of a top field first telecined stream:
8901 fieldmatch=order=tff:combmatch=none, decimate
8904 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
8906 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
8911 Transform the field order of the input video.
8913 It accepts the following parameters:
8918 The output field order. Valid values are @var{tff} for top field first or @var{bff}
8919 for bottom field first.
8922 The default value is @samp{tff}.
8924 The transformation is done by shifting the picture content up or down
8925 by one line, and filling the remaining line with appropriate picture content.
8926 This method is consistent with most broadcast field order converters.
8928 If the input video is not flagged as being interlaced, or it is already
8929 flagged as being of the required output field order, then this filter does
8930 not alter the incoming video.
8932 It is very useful when converting to or from PAL DV material,
8933 which is bottom field first.
8937 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
8940 @section fifo, afifo
8942 Buffer input images and send them when they are requested.
8944 It is mainly useful when auto-inserted by the libavfilter
8947 It does not take parameters.
8949 @section fillborders
8951 Fill borders of the input video, without changing video stream dimensions.
8952 Sometimes video can have garbage at the four edges and you may not want to
8953 crop video input to keep size multiple of some number.
8955 This filter accepts the following options:
8959 Number of pixels to fill from left border.
8962 Number of pixels to fill from right border.
8965 Number of pixels to fill from top border.
8968 Number of pixels to fill from bottom border.
8973 It accepts the following values:
8976 fill pixels using outermost pixels
8979 fill pixels using mirroring
8982 fill pixels with constant value
8985 Default is @var{smear}.
8988 Set color for pixels in fixed mode. Default is @var{black}.
8993 Find a rectangular object
8995 It accepts the following options:
8999 Filepath of the object image, needs to be in gray8.
9002 Detection threshold, default is 0.5.
9005 Number of mipmaps, default is 3.
9007 @item xmin, ymin, xmax, ymax
9008 Specifies the rectangle in which to search.
9011 @subsection Examples
9015 Generate a representative palette of a given video using @command{ffmpeg}:
9017 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
9023 Cover a rectangular object
9025 It accepts the following options:
9029 Filepath of the optional cover image, needs to be in yuv420.
9034 It accepts the following values:
9037 cover it by the supplied image
9039 cover it by interpolating the surrounding pixels
9042 Default value is @var{blur}.
9045 @subsection Examples
9049 Generate a representative palette of a given video using @command{ffmpeg}:
9051 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
9057 Flood area with values of same pixel components with another values.
9059 It accepts the following options:
9062 Set pixel x coordinate.
9065 Set pixel y coordinate.
9068 Set source #0 component value.
9071 Set source #1 component value.
9074 Set source #2 component value.
9077 Set source #3 component value.
9080 Set destination #0 component value.
9083 Set destination #1 component value.
9086 Set destination #2 component value.
9089 Set destination #3 component value.
9095 Convert the input video to one of the specified pixel formats.
9096 Libavfilter will try to pick one that is suitable as input to
9099 It accepts the following parameters:
9103 A '|'-separated list of pixel format names, such as
9104 "pix_fmts=yuv420p|monow|rgb24".
9108 @subsection Examples
9112 Convert the input video to the @var{yuv420p} format
9114 format=pix_fmts=yuv420p
9117 Convert the input video to any of the formats in the list
9119 format=pix_fmts=yuv420p|yuv444p|yuv410p
9126 Convert the video to specified constant frame rate by duplicating or dropping
9127 frames as necessary.
9129 It accepts the following parameters:
9133 The desired output frame rate. The default is @code{25}.
9136 Assume the first PTS should be the given value, in seconds. This allows for
9137 padding/trimming at the start of stream. By default, no assumption is made
9138 about the first frame's expected PTS, so no padding or trimming is done.
9139 For example, this could be set to 0 to pad the beginning with duplicates of
9140 the first frame if a video stream starts after the audio stream or to trim any
9141 frames with a negative PTS.
9144 Timestamp (PTS) rounding method.
9146 Possible values are:
9153 round towards -infinity
9155 round towards +infinity
9159 The default is @code{near}.
9162 Action performed when reading the last frame.
9164 Possible values are:
9167 Use same timestamp rounding method as used for other frames.
9169 Pass through last frame if input duration has not been reached yet.
9171 The default is @code{round}.
9175 Alternatively, the options can be specified as a flat string:
9176 @var{fps}[:@var{start_time}[:@var{round}]].
9178 See also the @ref{setpts} filter.
9180 @subsection Examples
9184 A typical usage in order to set the fps to 25:
9190 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
9192 fps=fps=film:round=near
9198 Pack two different video streams into a stereoscopic video, setting proper
9199 metadata on supported codecs. The two views should have the same size and
9200 framerate and processing will stop when the shorter video ends. Please note
9201 that you may conveniently adjust view properties with the @ref{scale} and
9204 It accepts the following parameters:
9208 The desired packing format. Supported values are:
9213 The views are next to each other (default).
9216 The views are on top of each other.
9219 The views are packed by line.
9222 The views are packed by column.
9225 The views are temporally interleaved.
9234 # Convert left and right views into a frame-sequential video
9235 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
9237 # Convert views into a side-by-side video with the same output resolution as the input
9238 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
9243 Change the frame rate by interpolating new video output frames from the source
9246 This filter is not designed to function correctly with interlaced media. If
9247 you wish to change the frame rate of interlaced media then you are required
9248 to deinterlace before this filter and re-interlace after this filter.
9250 A description of the accepted options follows.
9254 Specify the output frames per second. This option can also be specified
9255 as a value alone. The default is @code{50}.
9258 Specify the start of a range where the output frame will be created as a
9259 linear interpolation of two frames. The range is [@code{0}-@code{255}],
9260 the default is @code{15}.
9263 Specify the end of a range where the output frame will be created as a
9264 linear interpolation of two frames. The range is [@code{0}-@code{255}],
9265 the default is @code{240}.
9268 Specify the level at which a scene change is detected as a value between
9269 0 and 100 to indicate a new scene; a low value reflects a low
9270 probability for the current frame to introduce a new scene, while a higher
9271 value means the current frame is more likely to be one.
9272 The default is @code{8.2}.
9275 Specify flags influencing the filter process.
9277 Available value for @var{flags} is:
9280 @item scene_change_detect, scd
9281 Enable scene change detection using the value of the option @var{scene}.
9282 This flag is enabled by default.
9288 Select one frame every N-th frame.
9290 This filter accepts the following option:
9293 Select frame after every @code{step} frames.
9294 Allowed values are positive integers higher than 0. Default value is @code{1}.
9300 Apply a frei0r effect to the input video.
9302 To enable the compilation of this filter, you need to install the frei0r
9303 header and configure FFmpeg with @code{--enable-frei0r}.
9305 It accepts the following parameters:
9310 The name of the frei0r effect to load. If the environment variable
9311 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
9312 directories specified by the colon-separated list in @env{FREI0R_PATH}.
9313 Otherwise, the standard frei0r paths are searched, in this order:
9314 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
9315 @file{/usr/lib/frei0r-1/}.
9318 A '|'-separated list of parameters to pass to the frei0r effect.
9322 A frei0r effect parameter can be a boolean (its value is either
9323 "y" or "n"), a double, a color (specified as
9324 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
9325 numbers between 0.0 and 1.0, inclusive) or by a color description specified in the "Color"
9326 section in the ffmpeg-utils manual), a position (specified as @var{X}/@var{Y}, where
9327 @var{X} and @var{Y} are floating point numbers) and/or a string.
9329 The number and types of parameters depend on the loaded effect. If an
9330 effect parameter is not specified, the default value is set.
9332 @subsection Examples
9336 Apply the distort0r effect, setting the first two double parameters:
9338 frei0r=filter_name=distort0r:filter_params=0.5|0.01
9342 Apply the colordistance effect, taking a color as the first parameter:
9344 frei0r=colordistance:0.2/0.3/0.4
9345 frei0r=colordistance:violet
9346 frei0r=colordistance:0x112233
9350 Apply the perspective effect, specifying the top left and top right image
9353 frei0r=perspective:0.2/0.2|0.8/0.2
9357 For more information, see
9358 @url{http://frei0r.dyne.org}
9362 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
9364 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
9365 processing filter, one of them is performed once per block, not per pixel.
9366 This allows for much higher speed.
9368 The filter accepts the following options:
9372 Set quality. This option defines the number of levels for averaging. It accepts
9373 an integer in the range 4-5. Default value is @code{4}.
9376 Force a constant quantization parameter. It accepts an integer in range 0-63.
9377 If not set, the filter will use the QP from the video stream (if available).
9380 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
9381 more details but also more artifacts, while higher values make the image smoother
9382 but also blurrier. Default value is @code{0} − PSNR optimal.
9385 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
9386 option may cause flicker since the B-Frames have often larger QP. Default is
9387 @code{0} (not enabled).
9393 Apply Gaussian blur filter.
9395 The filter accepts the following options:
9399 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
9402 Set number of steps for Gaussian approximation. Defauls is @code{1}.
9405 Set which planes to filter. By default all planes are filtered.
9408 Set vertical sigma, if negative it will be same as @code{sigma}.
9409 Default is @code{-1}.
9414 The filter accepts the following options:
9418 Set the luminance expression.
9420 Set the chrominance blue expression.
9422 Set the chrominance red expression.
9424 Set the alpha expression.
9426 Set the red expression.
9428 Set the green expression.
9430 Set the blue expression.
9433 The colorspace is selected according to the specified options. If one
9434 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
9435 options is specified, the filter will automatically select a YCbCr
9436 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
9437 @option{blue_expr} options is specified, it will select an RGB
9440 If one of the chrominance expression is not defined, it falls back on the other
9441 one. If no alpha expression is specified it will evaluate to opaque value.
9442 If none of chrominance expressions are specified, they will evaluate
9443 to the luminance expression.
9445 The expressions can use the following variables and functions:
9449 The sequential number of the filtered frame, starting from @code{0}.
9453 The coordinates of the current sample.
9457 The width and height of the image.
9461 Width and height scale depending on the currently filtered plane. It is the
9462 ratio between the corresponding luma plane number of pixels and the current
9463 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
9464 @code{0.5,0.5} for chroma planes.
9467 Time of the current frame, expressed in seconds.
9470 Return the value of the pixel at location (@var{x},@var{y}) of the current
9474 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
9478 Return the value of the pixel at location (@var{x},@var{y}) of the
9479 blue-difference chroma plane. Return 0 if there is no such plane.
9482 Return the value of the pixel at location (@var{x},@var{y}) of the
9483 red-difference chroma plane. Return 0 if there is no such plane.
9488 Return the value of the pixel at location (@var{x},@var{y}) of the
9489 red/green/blue component. Return 0 if there is no such component.
9492 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
9493 plane. Return 0 if there is no such plane.
9496 For functions, if @var{x} and @var{y} are outside the area, the value will be
9497 automatically clipped to the closer edge.
9499 @subsection Examples
9503 Flip the image horizontally:
9509 Generate a bidimensional sine wave, with angle @code{PI/3} and a
9510 wavelength of 100 pixels:
9512 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
9516 Generate a fancy enigmatic moving light:
9518 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
9522 Generate a quick emboss effect:
9524 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
9528 Modify RGB components depending on pixel position:
9530 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
9534 Create a radial gradient that is the same size as the input (also see
9535 the @ref{vignette} filter):
9537 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
9543 Fix the banding artifacts that are sometimes introduced into nearly flat
9544 regions by truncation to 8-bit color depth.
9545 Interpolate the gradients that should go where the bands are, and
9548 It is designed for playback only. Do not use it prior to
9549 lossy compression, because compression tends to lose the dither and
9550 bring back the bands.
9552 It accepts the following parameters:
9557 The maximum amount by which the filter will change any one pixel. This is also
9558 the threshold for detecting nearly flat regions. Acceptable values range from
9559 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
9563 The neighborhood to fit the gradient to. A larger radius makes for smoother
9564 gradients, but also prevents the filter from modifying the pixels near detailed
9565 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
9566 values will be clipped to the valid range.
9570 Alternatively, the options can be specified as a flat string:
9571 @var{strength}[:@var{radius}]
9573 @subsection Examples
9577 Apply the filter with a @code{3.5} strength and radius of @code{8}:
9583 Specify radius, omitting the strength (which will fall-back to the default
9594 Apply a Hald CLUT to a video stream.
9596 First input is the video stream to process, and second one is the Hald CLUT.
9597 The Hald CLUT input can be a simple picture or a complete video stream.
9599 The filter accepts the following options:
9603 Force termination when the shortest input terminates. Default is @code{0}.
9605 Continue applying the last CLUT after the end of the stream. A value of
9606 @code{0} disable the filter after the last frame of the CLUT is reached.
9607 Default is @code{1}.
9610 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
9611 filters share the same internals).
9613 More information about the Hald CLUT can be found on Eskil Steenberg's website
9614 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
9616 @subsection Workflow examples
9618 @subsubsection Hald CLUT video stream
9620 Generate an identity Hald CLUT stream altered with various effects:
9622 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
9625 Note: make sure you use a lossless codec.
9627 Then use it with @code{haldclut} to apply it on some random stream:
9629 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
9632 The Hald CLUT will be applied to the 10 first seconds (duration of
9633 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
9634 to the remaining frames of the @code{mandelbrot} stream.
9636 @subsubsection Hald CLUT with preview
9638 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
9639 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
9640 biggest possible square starting at the top left of the picture. The remaining
9641 padding pixels (bottom or right) will be ignored. This area can be used to add
9642 a preview of the Hald CLUT.
9644 Typically, the following generated Hald CLUT will be supported by the
9645 @code{haldclut} filter:
9648 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
9649 pad=iw+320 [padded_clut];
9650 smptebars=s=320x256, split [a][b];
9651 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
9652 [main][b] overlay=W-320" -frames:v 1 clut.png
9655 It contains the original and a preview of the effect of the CLUT: SMPTE color
9656 bars are displayed on the right-top, and below the same color bars processed by
9659 Then, the effect of this Hald CLUT can be visualized with:
9661 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
9666 Flip the input video horizontally.
9668 For example, to horizontally flip the input video with @command{ffmpeg}:
9670 ffmpeg -i in.avi -vf "hflip" out.avi
9674 This filter applies a global color histogram equalization on a
9677 It can be used to correct video that has a compressed range of pixel
9678 intensities. The filter redistributes the pixel intensities to
9679 equalize their distribution across the intensity range. It may be
9680 viewed as an "automatically adjusting contrast filter". This filter is
9681 useful only for correcting degraded or poorly captured source
9684 The filter accepts the following options:
9688 Determine the amount of equalization to be applied. As the strength
9689 is reduced, the distribution of pixel intensities more-and-more
9690 approaches that of the input frame. The value must be a float number
9691 in the range [0,1] and defaults to 0.200.
9694 Set the maximum intensity that can generated and scale the output
9695 values appropriately. The strength should be set as desired and then
9696 the intensity can be limited if needed to avoid washing-out. The value
9697 must be a float number in the range [0,1] and defaults to 0.210.
9700 Set the antibanding level. If enabled the filter will randomly vary
9701 the luminance of output pixels by a small amount to avoid banding of
9702 the histogram. Possible values are @code{none}, @code{weak} or
9703 @code{strong}. It defaults to @code{none}.
9708 Compute and draw a color distribution histogram for the input video.
9710 The computed histogram is a representation of the color component
9711 distribution in an image.
9713 Standard histogram displays the color components distribution in an image.
9714 Displays color graph for each color component. Shows distribution of
9715 the Y, U, V, A or R, G, B components, depending on input format, in the
9716 current frame. Below each graph a color component scale meter is shown.
9718 The filter accepts the following options:
9722 Set height of level. Default value is @code{200}.
9723 Allowed range is [50, 2048].
9726 Set height of color scale. Default value is @code{12}.
9727 Allowed range is [0, 40].
9731 It accepts the following values:
9734 Per color component graphs are placed below each other.
9737 Per color component graphs are placed side by side.
9740 Presents information identical to that in the @code{parade}, except
9741 that the graphs representing color components are superimposed directly
9744 Default is @code{stack}.
9747 Set mode. Can be either @code{linear}, or @code{logarithmic}.
9748 Default is @code{linear}.
9751 Set what color components to display.
9752 Default is @code{7}.
9755 Set foreground opacity. Default is @code{0.7}.
9758 Set background opacity. Default is @code{0.5}.
9761 @subsection Examples
9766 Calculate and draw histogram:
9768 ffplay -i input -vf histogram
9776 This is a high precision/quality 3d denoise filter. It aims to reduce
9777 image noise, producing smooth images and making still images really
9778 still. It should enhance compressibility.
9780 It accepts the following optional parameters:
9784 A non-negative floating point number which specifies spatial luma strength.
9787 @item chroma_spatial
9788 A non-negative floating point number which specifies spatial chroma strength.
9789 It defaults to 3.0*@var{luma_spatial}/4.0.
9792 A floating point number which specifies luma temporal strength. It defaults to
9793 6.0*@var{luma_spatial}/4.0.
9796 A floating point number which specifies chroma temporal strength. It defaults to
9797 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
9802 Download hardware frames to system memory.
9804 The input must be in hardware frames, and the output a non-hardware format.
9805 Not all formats will be supported on the output - it may be necessary to insert
9806 an additional @option{format} filter immediately following in the graph to get
9807 the output in a supported format.
9811 Map hardware frames to system memory or to another device.
9813 This filter has several different modes of operation; which one is used depends
9814 on the input and output formats:
9817 Hardware frame input, normal frame output
9819 Map the input frames to system memory and pass them to the output. If the
9820 original hardware frame is later required (for example, after overlaying
9821 something else on part of it), the @option{hwmap} filter can be used again
9822 in the next mode to retrieve it.
9824 Normal frame input, hardware frame output
9826 If the input is actually a software-mapped hardware frame, then unmap it -
9827 that is, return the original hardware frame.
9829 Otherwise, a device must be provided. Create new hardware surfaces on that
9830 device for the output, then map them back to the software format at the input
9831 and give those frames to the preceding filter. This will then act like the
9832 @option{hwupload} filter, but may be able to avoid an additional copy when
9833 the input is already in a compatible format.
9835 Hardware frame input and output
9837 A device must be supplied for the output, either directly or with the
9838 @option{derive_device} option. The input and output devices must be of
9839 different types and compatible - the exact meaning of this is
9840 system-dependent, but typically it means that they must refer to the same
9841 underlying hardware context (for example, refer to the same graphics card).
9843 If the input frames were originally created on the output device, then unmap
9844 to retrieve the original frames.
9846 Otherwise, map the frames to the output device - create new hardware frames
9847 on the output corresponding to the frames on the input.
9850 The following additional parameters are accepted:
9854 Set the frame mapping mode. Some combination of:
9857 The mapped frame should be readable.
9859 The mapped frame should be writeable.
9861 The mapping will always overwrite the entire frame.
9863 This may improve performance in some cases, as the original contents of the
9864 frame need not be loaded.
9866 The mapping must not involve any copying.
9868 Indirect mappings to copies of frames are created in some cases where either
9869 direct mapping is not possible or it would have unexpected properties.
9870 Setting this flag ensures that the mapping is direct and will fail if that is
9873 Defaults to @var{read+write} if not specified.
9875 @item derive_device @var{type}
9876 Rather than using the device supplied at initialisation, instead derive a new
9877 device of type @var{type} from the device the input frames exist on.
9880 In a hardware to hardware mapping, map in reverse - create frames in the sink
9881 and map them back to the source. This may be necessary in some cases where
9882 a mapping in one direction is required but only the opposite direction is
9883 supported by the devices being used.
9885 This option is dangerous - it may break the preceding filter in undefined
9886 ways if there are any additional constraints on that filter's output.
9887 Do not use it without fully understanding the implications of its use.
9892 Upload system memory frames to hardware surfaces.
9894 The device to upload to must be supplied when the filter is initialised. If
9895 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
9898 @anchor{hwupload_cuda}
9899 @section hwupload_cuda
9901 Upload system memory frames to a CUDA device.
9903 It accepts the following optional parameters:
9907 The number of the CUDA device to use
9912 Apply a high-quality magnification filter designed for pixel art. This filter
9913 was originally created by Maxim Stepin.
9915 It accepts the following option:
9919 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
9920 @code{hq3x} and @code{4} for @code{hq4x}.
9921 Default is @code{3}.
9925 Stack input videos horizontally.
9927 All streams must be of same pixel format and of same height.
9929 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
9930 to create same output.
9932 The filter accept the following option:
9936 Set number of input streams. Default is 2.
9939 If set to 1, force the output to terminate when the shortest input
9940 terminates. Default value is 0.
9945 Modify the hue and/or the saturation of the input.
9947 It accepts the following parameters:
9951 Specify the hue angle as a number of degrees. It accepts an expression,
9952 and defaults to "0".
9955 Specify the saturation in the [-10,10] range. It accepts an expression and
9959 Specify the hue angle as a number of radians. It accepts an
9960 expression, and defaults to "0".
9963 Specify the brightness in the [-10,10] range. It accepts an expression and
9967 @option{h} and @option{H} are mutually exclusive, and can't be
9968 specified at the same time.
9970 The @option{b}, @option{h}, @option{H} and @option{s} option values are
9971 expressions containing the following constants:
9975 frame count of the input frame starting from 0
9978 presentation timestamp of the input frame expressed in time base units
9981 frame rate of the input video, NAN if the input frame rate is unknown
9984 timestamp expressed in seconds, NAN if the input timestamp is unknown
9987 time base of the input video
9990 @subsection Examples
9994 Set the hue to 90 degrees and the saturation to 1.0:
10000 Same command but expressing the hue in radians:
10006 Rotate hue and make the saturation swing between 0
10007 and 2 over a period of 1 second:
10009 hue="H=2*PI*t: s=sin(2*PI*t)+1"
10013 Apply a 3 seconds saturation fade-in effect starting at 0:
10015 hue="s=min(t/3\,1)"
10018 The general fade-in expression can be written as:
10020 hue="s=min(0\, max((t-START)/DURATION\, 1))"
10024 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
10026 hue="s=max(0\, min(1\, (8-t)/3))"
10029 The general fade-out expression can be written as:
10031 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
10036 @subsection Commands
10038 This filter supports the following commands:
10044 Modify the hue and/or the saturation and/or brightness of the input video.
10045 The command accepts the same syntax of the corresponding option.
10047 If the specified expression is not valid, it is kept at its current
10051 @section hysteresis
10053 Grow first stream into second stream by connecting components.
10054 This makes it possible to build more robust edge masks.
10056 This filter accepts the following options:
10060 Set which planes will be processed as bitmap, unprocessed planes will be
10061 copied from first stream.
10062 By default value 0xf, all planes will be processed.
10065 Set threshold which is used in filtering. If pixel component value is higher than
10066 this value filter algorithm for connecting components is activated.
10067 By default value is 0.
10072 Detect video interlacing type.
10074 This filter tries to detect if the input frames are interlaced, progressive,
10075 top or bottom field first. It will also try to detect fields that are
10076 repeated between adjacent frames (a sign of telecine).
10078 Single frame detection considers only immediately adjacent frames when classifying each frame.
10079 Multiple frame detection incorporates the classification history of previous frames.
10081 The filter will log these metadata values:
10084 @item single.current_frame
10085 Detected type of current frame using single-frame detection. One of:
10086 ``tff'' (top field first), ``bff'' (bottom field first),
10087 ``progressive'', or ``undetermined''
10090 Cumulative number of frames detected as top field first using single-frame detection.
10093 Cumulative number of frames detected as top field first using multiple-frame detection.
10096 Cumulative number of frames detected as bottom field first using single-frame detection.
10098 @item multiple.current_frame
10099 Detected type of current frame using multiple-frame detection. One of:
10100 ``tff'' (top field first), ``bff'' (bottom field first),
10101 ``progressive'', or ``undetermined''
10104 Cumulative number of frames detected as bottom field first using multiple-frame detection.
10106 @item single.progressive
10107 Cumulative number of frames detected as progressive using single-frame detection.
10109 @item multiple.progressive
10110 Cumulative number of frames detected as progressive using multiple-frame detection.
10112 @item single.undetermined
10113 Cumulative number of frames that could not be classified using single-frame detection.
10115 @item multiple.undetermined
10116 Cumulative number of frames that could not be classified using multiple-frame detection.
10118 @item repeated.current_frame
10119 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
10121 @item repeated.neither
10122 Cumulative number of frames with no repeated field.
10125 Cumulative number of frames with the top field repeated from the previous frame's top field.
10127 @item repeated.bottom
10128 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
10131 The filter accepts the following options:
10135 Set interlacing threshold.
10137 Set progressive threshold.
10139 Threshold for repeated field detection.
10141 Number of frames after which a given frame's contribution to the
10142 statistics is halved (i.e., it contributes only 0.5 to its
10143 classification). The default of 0 means that all frames seen are given
10144 full weight of 1.0 forever.
10145 @item analyze_interlaced_flag
10146 When this is not 0 then idet will use the specified number of frames to determine
10147 if the interlaced flag is accurate, it will not count undetermined frames.
10148 If the flag is found to be accurate it will be used without any further
10149 computations, if it is found to be inaccurate it will be cleared without any
10150 further computations. This allows inserting the idet filter as a low computational
10151 method to clean up the interlaced flag
10156 Deinterleave or interleave fields.
10158 This filter allows one to process interlaced images fields without
10159 deinterlacing them. Deinterleaving splits the input frame into 2
10160 fields (so called half pictures). Odd lines are moved to the top
10161 half of the output image, even lines to the bottom half.
10162 You can process (filter) them independently and then re-interleave them.
10164 The filter accepts the following options:
10168 @item chroma_mode, c
10169 @item alpha_mode, a
10170 Available values for @var{luma_mode}, @var{chroma_mode} and
10171 @var{alpha_mode} are:
10177 @item deinterleave, d
10178 Deinterleave fields, placing one above the other.
10180 @item interleave, i
10181 Interleave fields. Reverse the effect of deinterleaving.
10183 Default value is @code{none}.
10185 @item luma_swap, ls
10186 @item chroma_swap, cs
10187 @item alpha_swap, as
10188 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
10193 Apply inflate effect to the video.
10195 This filter replaces the pixel by the local(3x3) average by taking into account
10196 only values higher than the pixel.
10198 It accepts the following options:
10205 Limit the maximum change for each plane, default is 65535.
10206 If 0, plane will remain unchanged.
10211 Simple interlacing filter from progressive contents. This interleaves upper (or
10212 lower) lines from odd frames with lower (or upper) lines from even frames,
10213 halving the frame rate and preserving image height.
10216 Original Original New Frame
10217 Frame 'j' Frame 'j+1' (tff)
10218 ========== =========== ==================
10219 Line 0 --------------------> Frame 'j' Line 0
10220 Line 1 Line 1 ----> Frame 'j+1' Line 1
10221 Line 2 ---------------------> Frame 'j' Line 2
10222 Line 3 Line 3 ----> Frame 'j+1' Line 3
10224 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
10227 It accepts the following optional parameters:
10231 This determines whether the interlaced frame is taken from the even
10232 (tff - default) or odd (bff) lines of the progressive frame.
10235 Vertical lowpass filter to avoid twitter interlacing and
10236 reduce moire patterns.
10240 Disable vertical lowpass filter
10243 Enable linear filter (default)
10246 Enable complex filter. This will slightly less reduce twitter and moire
10247 but better retain detail and subjective sharpness impression.
10254 Deinterlace input video by applying Donald Graft's adaptive kernel
10255 deinterling. Work on interlaced parts of a video to produce
10256 progressive frames.
10258 The description of the accepted parameters follows.
10262 Set the threshold which affects the filter's tolerance when
10263 determining if a pixel line must be processed. It must be an integer
10264 in the range [0,255] and defaults to 10. A value of 0 will result in
10265 applying the process on every pixels.
10268 Paint pixels exceeding the threshold value to white if set to 1.
10272 Set the fields order. Swap fields if set to 1, leave fields alone if
10276 Enable additional sharpening if set to 1. Default is 0.
10279 Enable twoway sharpening if set to 1. Default is 0.
10282 @subsection Examples
10286 Apply default values:
10288 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
10292 Enable additional sharpening:
10298 Paint processed pixels in white:
10304 @section lenscorrection
10306 Correct radial lens distortion
10308 This filter can be used to correct for radial distortion as can result from the use
10309 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
10310 one can use tools available for example as part of opencv or simply trial-and-error.
10311 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
10312 and extract the k1 and k2 coefficients from the resulting matrix.
10314 Note that effectively the same filter is available in the open-source tools Krita and
10315 Digikam from the KDE project.
10317 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
10318 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
10319 brightness distribution, so you may want to use both filters together in certain
10320 cases, though you will have to take care of ordering, i.e. whether vignetting should
10321 be applied before or after lens correction.
10323 @subsection Options
10325 The filter accepts the following options:
10329 Relative x-coordinate of the focal point of the image, and thereby the center of the
10330 distortion. This value has a range [0,1] and is expressed as fractions of the image
10333 Relative y-coordinate of the focal point of the image, and thereby the center of the
10334 distortion. This value has a range [0,1] and is expressed as fractions of the image
10337 Coefficient of the quadratic correction term. 0.5 means no correction.
10339 Coefficient of the double quadratic correction term. 0.5 means no correction.
10342 The formula that generates the correction is:
10344 @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)
10346 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
10347 distances from the focal point in the source and target images, respectively.
10351 Obtain the VMAF (Video Multi-Method Assessment Fusion)
10352 score between two input videos.
10354 The obtained VMAF score is printed through the logging system.
10356 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
10357 After installing the library it can be enabled using:
10358 @code{./configure --enable-libvmaf}.
10359 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
10361 The filter has following options:
10365 Set the model path which is to be used for SVM.
10366 Default value: @code{"vmaf_v0.6.1.pkl"}
10369 Set the file path to be used to store logs.
10372 Set the format of the log file (xml or json).
10374 @item enable_transform
10375 Enables transform for computing vmaf.
10378 Invokes the phone model which will generate VMAF scores higher than in the
10379 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
10382 Enables computing psnr along with vmaf.
10385 Enables computing ssim along with vmaf.
10388 Enables computing ms_ssim along with vmaf.
10391 Set the pool method (mean, min or harmonic mean) to be used for computing vmaf.
10394 This filter also supports the @ref{framesync} options.
10396 On the below examples the input file @file{main.mpg} being processed is
10397 compared with the reference file @file{ref.mpg}.
10400 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
10403 Example with options:
10405 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:enable-transform=1" -f null -
10410 Limits the pixel components values to the specified range [min, max].
10412 The filter accepts the following options:
10416 Lower bound. Defaults to the lowest allowed value for the input.
10419 Upper bound. Defaults to the highest allowed value for the input.
10422 Specify which planes will be processed. Defaults to all available.
10429 The filter accepts the following options:
10433 Set the number of loops. Setting this value to -1 will result in infinite loops.
10437 Set maximal size in number of frames. Default is 0.
10440 Set first frame of loop. Default is 0.
10446 Apply a 3D LUT to an input video.
10448 The filter accepts the following options:
10452 Set the 3D LUT file name.
10454 Currently supported formats:
10466 Select interpolation mode.
10468 Available values are:
10472 Use values from the nearest defined point.
10474 Interpolate values using the 8 points defining a cube.
10476 Interpolate values using a tetrahedron.
10480 This filter also supports the @ref{framesync} options.
10484 Turn certain luma values into transparency.
10486 The filter accepts the following options:
10490 Set the luma which will be used as base for transparency.
10491 Default value is @code{0}.
10494 Set the range of luma values to be keyed out.
10495 Default value is @code{0}.
10498 Set the range of softness. Default value is @code{0}.
10499 Use this to control gradual transition from zero to full transparency.
10502 @section lut, lutrgb, lutyuv
10504 Compute a look-up table for binding each pixel component input value
10505 to an output value, and apply it to the input video.
10507 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
10508 to an RGB input video.
10510 These filters accept the following parameters:
10513 set first pixel component expression
10515 set second pixel component expression
10517 set third pixel component expression
10519 set fourth pixel component expression, corresponds to the alpha component
10522 set red component expression
10524 set green component expression
10526 set blue component expression
10528 alpha component expression
10531 set Y/luminance component expression
10533 set U/Cb component expression
10535 set V/Cr component expression
10538 Each of them specifies the expression to use for computing the lookup table for
10539 the corresponding pixel component values.
10541 The exact component associated to each of the @var{c*} options depends on the
10544 The @var{lut} filter requires either YUV or RGB pixel formats in input,
10545 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
10547 The expressions can contain the following constants and functions:
10552 The input width and height.
10555 The input value for the pixel component.
10558 The input value, clipped to the @var{minval}-@var{maxval} range.
10561 The maximum value for the pixel component.
10564 The minimum value for the pixel component.
10567 The negated value for the pixel component value, clipped to the
10568 @var{minval}-@var{maxval} range; it corresponds to the expression
10569 "maxval-clipval+minval".
10572 The computed value in @var{val}, clipped to the
10573 @var{minval}-@var{maxval} range.
10575 @item gammaval(gamma)
10576 The computed gamma correction value of the pixel component value,
10577 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
10579 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
10583 All expressions default to "val".
10585 @subsection Examples
10589 Negate input video:
10591 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
10592 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
10595 The above is the same as:
10597 lutrgb="r=negval:g=negval:b=negval"
10598 lutyuv="y=negval:u=negval:v=negval"
10608 Remove chroma components, turning the video into a graytone image:
10610 lutyuv="u=128:v=128"
10614 Apply a luma burning effect:
10620 Remove green and blue components:
10626 Set a constant alpha channel value on input:
10628 format=rgba,lutrgb=a="maxval-minval/2"
10632 Correct luminance gamma by a factor of 0.5:
10634 lutyuv=y=gammaval(0.5)
10638 Discard least significant bits of luma:
10640 lutyuv=y='bitand(val, 128+64+32)'
10644 Technicolor like effect:
10646 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
10650 @section lut2, tlut2
10652 The @code{lut2} filter takes two input streams and outputs one
10655 The @code{tlut2} (time lut2) filter takes two consecutive frames
10656 from one single stream.
10658 This filter accepts the following parameters:
10661 set first pixel component expression
10663 set second pixel component expression
10665 set third pixel component expression
10667 set fourth pixel component expression, corresponds to the alpha component
10670 Each of them specifies the expression to use for computing the lookup table for
10671 the corresponding pixel component values.
10673 The exact component associated to each of the @var{c*} options depends on the
10676 The expressions can contain the following constants:
10681 The input width and height.
10684 The first input value for the pixel component.
10687 The second input value for the pixel component.
10690 The first input video bit depth.
10693 The second input video bit depth.
10696 All expressions default to "x".
10698 @subsection Examples
10702 Highlight differences between two RGB video streams:
10704 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)'
10708 Highlight differences between two YUV video streams:
10710 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)'
10714 Show max difference between two video streams:
10716 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)))'
10720 @section maskedclamp
10722 Clamp the first input stream with the second input and third input stream.
10724 Returns the value of first stream to be between second input
10725 stream - @code{undershoot} and third input stream + @code{overshoot}.
10727 This filter accepts the following options:
10730 Default value is @code{0}.
10733 Default value is @code{0}.
10736 Set which planes will be processed as bitmap, unprocessed planes will be
10737 copied from first stream.
10738 By default value 0xf, all planes will be processed.
10741 @section maskedmerge
10743 Merge the first input stream with the second input stream using per pixel
10744 weights in the third input stream.
10746 A value of 0 in the third stream pixel component means that pixel component
10747 from first stream is returned unchanged, while maximum value (eg. 255 for
10748 8-bit videos) means that pixel component from second stream is returned
10749 unchanged. Intermediate values define the amount of merging between both
10750 input stream's pixel components.
10752 This filter accepts the following options:
10755 Set which planes will be processed as bitmap, unprocessed planes will be
10756 copied from first stream.
10757 By default value 0xf, all planes will be processed.
10762 Apply motion-compensation deinterlacing.
10764 It needs one field per frame as input and must thus be used together
10765 with yadif=1/3 or equivalent.
10767 This filter accepts the following options:
10770 Set the deinterlacing mode.
10772 It accepts one of the following values:
10777 use iterative motion estimation
10779 like @samp{slow}, but use multiple reference frames.
10781 Default value is @samp{fast}.
10784 Set the picture field parity assumed for the input video. It must be
10785 one of the following values:
10789 assume top field first
10791 assume bottom field first
10794 Default value is @samp{bff}.
10797 Set per-block quantization parameter (QP) used by the internal
10800 Higher values should result in a smoother motion vector field but less
10801 optimal individual vectors. Default value is 1.
10804 @section mergeplanes
10806 Merge color channel components from several video streams.
10808 The filter accepts up to 4 input streams, and merge selected input
10809 planes to the output video.
10811 This filter accepts the following options:
10814 Set input to output plane mapping. Default is @code{0}.
10816 The mappings is specified as a bitmap. It should be specified as a
10817 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
10818 mapping for the first plane of the output stream. 'A' sets the number of
10819 the input stream to use (from 0 to 3), and 'a' the plane number of the
10820 corresponding input to use (from 0 to 3). The rest of the mappings is
10821 similar, 'Bb' describes the mapping for the output stream second
10822 plane, 'Cc' describes the mapping for the output stream third plane and
10823 'Dd' describes the mapping for the output stream fourth plane.
10826 Set output pixel format. Default is @code{yuva444p}.
10829 @subsection Examples
10833 Merge three gray video streams of same width and height into single video stream:
10835 [a0][a1][a2]mergeplanes=0x001020:yuv444p
10839 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
10841 [a0][a1]mergeplanes=0x00010210:yuva444p
10845 Swap Y and A plane in yuva444p stream:
10847 format=yuva444p,mergeplanes=0x03010200:yuva444p
10851 Swap U and V plane in yuv420p stream:
10853 format=yuv420p,mergeplanes=0x000201:yuv420p
10857 Cast a rgb24 clip to yuv444p:
10859 format=rgb24,mergeplanes=0x000102:yuv444p
10865 Estimate and export motion vectors using block matching algorithms.
10866 Motion vectors are stored in frame side data to be used by other filters.
10868 This filter accepts the following options:
10871 Specify the motion estimation method. Accepts one of the following values:
10875 Exhaustive search algorithm.
10877 Three step search algorithm.
10879 Two dimensional logarithmic search algorithm.
10881 New three step search algorithm.
10883 Four step search algorithm.
10885 Diamond search algorithm.
10887 Hexagon-based search algorithm.
10889 Enhanced predictive zonal search algorithm.
10891 Uneven multi-hexagon search algorithm.
10893 Default value is @samp{esa}.
10896 Macroblock size. Default @code{16}.
10899 Search parameter. Default @code{7}.
10902 @section midequalizer
10904 Apply Midway Image Equalization effect using two video streams.
10906 Midway Image Equalization adjusts a pair of images to have the same
10907 histogram, while maintaining their dynamics as much as possible. It's
10908 useful for e.g. matching exposures from a pair of stereo cameras.
10910 This filter has two inputs and one output, which must be of same pixel format, but
10911 may be of different sizes. The output of filter is first input adjusted with
10912 midway histogram of both inputs.
10914 This filter accepts the following option:
10918 Set which planes to process. Default is @code{15}, which is all available planes.
10921 @section minterpolate
10923 Convert the video to specified frame rate using motion interpolation.
10925 This filter accepts the following options:
10928 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}.
10931 Motion interpolation mode. Following values are accepted:
10934 Duplicate previous or next frame for interpolating new ones.
10936 Blend source frames. Interpolated frame is mean of previous and next frames.
10938 Motion compensated interpolation. Following options are effective when this mode is selected:
10942 Motion compensation mode. Following values are accepted:
10945 Overlapped block motion compensation.
10947 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
10949 Default mode is @samp{obmc}.
10952 Motion estimation mode. Following values are accepted:
10955 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
10957 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
10959 Default mode is @samp{bilat}.
10962 The algorithm to be used for motion estimation. Following values are accepted:
10965 Exhaustive search algorithm.
10967 Three step search algorithm.
10969 Two dimensional logarithmic search algorithm.
10971 New three step search algorithm.
10973 Four step search algorithm.
10975 Diamond search algorithm.
10977 Hexagon-based search algorithm.
10979 Enhanced predictive zonal search algorithm.
10981 Uneven multi-hexagon search algorithm.
10983 Default algorithm is @samp{epzs}.
10986 Macroblock size. Default @code{16}.
10989 Motion estimation search parameter. Default @code{32}.
10992 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).
10997 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:
11000 Disable scene change detection.
11002 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
11004 Default method is @samp{fdiff}.
11006 @item scd_threshold
11007 Scene change detection threshold. Default is @code{5.0}.
11012 Mix several video input streams into one video stream.
11014 A description of the accepted options follows.
11018 The number of inputs. If unspecified, it defaults to 2.
11021 Specify weight of each input video stream as sequence.
11022 Each weight is separated by space.
11025 Specify how end of stream is determined.
11028 The duration of the longest input. (default)
11031 The duration of the shortest input.
11034 The duration of the first input.
11038 @section mpdecimate
11040 Drop frames that do not differ greatly from the previous frame in
11041 order to reduce frame rate.
11043 The main use of this filter is for very-low-bitrate encoding
11044 (e.g. streaming over dialup modem), but it could in theory be used for
11045 fixing movies that were inverse-telecined incorrectly.
11047 A description of the accepted options follows.
11051 Set the maximum number of consecutive frames which can be dropped (if
11052 positive), or the minimum interval between dropped frames (if
11053 negative). If the value is 0, the frame is dropped disregarding the
11054 number of previous sequentially dropped frames.
11056 Default value is 0.
11061 Set the dropping threshold values.
11063 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
11064 represent actual pixel value differences, so a threshold of 64
11065 corresponds to 1 unit of difference for each pixel, or the same spread
11066 out differently over the block.
11068 A frame is a candidate for dropping if no 8x8 blocks differ by more
11069 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
11070 meaning the whole image) differ by more than a threshold of @option{lo}.
11072 Default value for @option{hi} is 64*12, default value for @option{lo} is
11073 64*5, and default value for @option{frac} is 0.33.
11079 Negate input video.
11081 It accepts an integer in input; if non-zero it negates the
11082 alpha component (if available). The default value in input is 0.
11086 Denoise frames using Non-Local Means algorithm.
11088 Each pixel is adjusted by looking for other pixels with similar contexts. This
11089 context similarity is defined by comparing their surrounding patches of size
11090 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
11093 Note that the research area defines centers for patches, which means some
11094 patches will be made of pixels outside that research area.
11096 The filter accepts the following options.
11100 Set denoising strength.
11106 Same as @option{p} but for chroma planes.
11108 The default value is @var{0} and means automatic.
11114 Same as @option{r} but for chroma planes.
11116 The default value is @var{0} and means automatic.
11121 Deinterlace video using neural network edge directed interpolation.
11123 This filter accepts the following options:
11127 Mandatory option, without binary file filter can not work.
11128 Currently file can be found here:
11129 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
11132 Set which frames to deinterlace, by default it is @code{all}.
11133 Can be @code{all} or @code{interlaced}.
11136 Set mode of operation.
11138 Can be one of the following:
11142 Use frame flags, both fields.
11144 Use frame flags, single field.
11146 Use top field only.
11148 Use bottom field only.
11150 Use both fields, top first.
11152 Use both fields, bottom first.
11156 Set which planes to process, by default filter process all frames.
11159 Set size of local neighborhood around each pixel, used by the predictor neural
11162 Can be one of the following:
11175 Set the number of neurons in predictor neural network.
11176 Can be one of the following:
11187 Controls the number of different neural network predictions that are blended
11188 together to compute the final output value. Can be @code{fast}, default or
11192 Set which set of weights to use in the predictor.
11193 Can be one of the following:
11197 weights trained to minimize absolute error
11199 weights trained to minimize squared error
11203 Controls whether or not the prescreener neural network is used to decide
11204 which pixels should be processed by the predictor neural network and which
11205 can be handled by simple cubic interpolation.
11206 The prescreener is trained to know whether cubic interpolation will be
11207 sufficient for a pixel or whether it should be predicted by the predictor nn.
11208 The computational complexity of the prescreener nn is much less than that of
11209 the predictor nn. Since most pixels can be handled by cubic interpolation,
11210 using the prescreener generally results in much faster processing.
11211 The prescreener is pretty accurate, so the difference between using it and not
11212 using it is almost always unnoticeable.
11214 Can be one of the following:
11222 Default is @code{new}.
11225 Set various debugging flags.
11230 Force libavfilter not to use any of the specified pixel formats for the
11231 input to the next filter.
11233 It accepts the following parameters:
11237 A '|'-separated list of pixel format names, such as
11238 pix_fmts=yuv420p|monow|rgb24".
11242 @subsection Examples
11246 Force libavfilter to use a format different from @var{yuv420p} for the
11247 input to the vflip filter:
11249 noformat=pix_fmts=yuv420p,vflip
11253 Convert the input video to any of the formats not contained in the list:
11255 noformat=yuv420p|yuv444p|yuv410p
11261 Add noise on video input frame.
11263 The filter accepts the following options:
11271 Set noise seed for specific pixel component or all pixel components in case
11272 of @var{all_seed}. Default value is @code{123457}.
11274 @item all_strength, alls
11275 @item c0_strength, c0s
11276 @item c1_strength, c1s
11277 @item c2_strength, c2s
11278 @item c3_strength, c3s
11279 Set noise strength for specific pixel component or all pixel components in case
11280 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
11282 @item all_flags, allf
11283 @item c0_flags, c0f
11284 @item c1_flags, c1f
11285 @item c2_flags, c2f
11286 @item c3_flags, c3f
11287 Set pixel component flags or set flags for all components if @var{all_flags}.
11288 Available values for component flags are:
11291 averaged temporal noise (smoother)
11293 mix random noise with a (semi)regular pattern
11295 temporal noise (noise pattern changes between frames)
11297 uniform noise (gaussian otherwise)
11301 @subsection Examples
11303 Add temporal and uniform noise to input video:
11305 noise=alls=20:allf=t+u
11310 Normalize RGB video (aka histogram stretching, contrast stretching).
11311 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
11313 For each channel of each frame, the filter computes the input range and maps
11314 it linearly to the user-specified output range. The output range defaults
11315 to the full dynamic range from pure black to pure white.
11317 Temporal smoothing can be used on the input range to reduce flickering (rapid
11318 changes in brightness) caused when small dark or bright objects enter or leave
11319 the scene. This is similar to the auto-exposure (automatic gain control) on a
11320 video camera, and, like a video camera, it may cause a period of over- or
11321 under-exposure of the video.
11323 The R,G,B channels can be normalized independently, which may cause some
11324 color shifting, or linked together as a single channel, which prevents
11325 color shifting. Linked normalization preserves hue. Independent normalization
11326 does not, so it can be used to remove some color casts. Independent and linked
11327 normalization can be combined in any ratio.
11329 The normalize filter accepts the following options:
11334 Colors which define the output range. The minimum input value is mapped to
11335 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
11336 The defaults are black and white respectively. Specifying white for
11337 @var{blackpt} and black for @var{whitept} will give color-inverted,
11338 normalized video. Shades of grey can be used to reduce the dynamic range
11339 (contrast). Specifying saturated colors here can create some interesting
11343 The number of previous frames to use for temporal smoothing. The input range
11344 of each channel is smoothed using a rolling average over the current frame
11345 and the @var{smoothing} previous frames. The default is 0 (no temporal
11349 Controls the ratio of independent (color shifting) channel normalization to
11350 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
11351 independent. Defaults to 1.0 (fully independent).
11354 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
11355 expensive no-op. Defaults to 1.0 (full strength).
11359 @subsection Examples
11361 Stretch video contrast to use the full dynamic range, with no temporal
11362 smoothing; may flicker depending on the source content:
11364 normalize=blackpt=black:whitept=white:smoothing=0
11367 As above, but with 50 frames of temporal smoothing; flicker should be
11368 reduced, depending on the source content:
11370 normalize=blackpt=black:whitept=white:smoothing=50
11373 As above, but with hue-preserving linked channel normalization:
11375 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
11378 As above, but with half strength:
11380 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
11383 Map the darkest input color to red, the brightest input color to cyan:
11385 normalize=blackpt=red:whitept=cyan
11390 Pass the video source unchanged to the output.
11393 Optical Character Recognition
11395 This filter uses Tesseract for optical character recognition.
11397 It accepts the following options:
11401 Set datapath to tesseract data. Default is to use whatever was
11402 set at installation.
11405 Set language, default is "eng".
11408 Set character whitelist.
11411 Set character blacklist.
11414 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
11418 Apply a video transform using libopencv.
11420 To enable this filter, install the libopencv library and headers and
11421 configure FFmpeg with @code{--enable-libopencv}.
11423 It accepts the following parameters:
11428 The name of the libopencv filter to apply.
11430 @item filter_params
11431 The parameters to pass to the libopencv filter. If not specified, the default
11432 values are assumed.
11436 Refer to the official libopencv documentation for more precise
11438 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
11440 Several libopencv filters are supported; see the following subsections.
11445 Dilate an image by using a specific structuring element.
11446 It corresponds to the libopencv function @code{cvDilate}.
11448 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
11450 @var{struct_el} represents a structuring element, and has the syntax:
11451 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
11453 @var{cols} and @var{rows} represent the number of columns and rows of
11454 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
11455 point, and @var{shape} the shape for the structuring element. @var{shape}
11456 must be "rect", "cross", "ellipse", or "custom".
11458 If the value for @var{shape} is "custom", it must be followed by a
11459 string of the form "=@var{filename}". The file with name
11460 @var{filename} is assumed to represent a binary image, with each
11461 printable character corresponding to a bright pixel. When a custom
11462 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
11463 or columns and rows of the read file are assumed instead.
11465 The default value for @var{struct_el} is "3x3+0x0/rect".
11467 @var{nb_iterations} specifies the number of times the transform is
11468 applied to the image, and defaults to 1.
11472 # Use the default values
11475 # Dilate using a structuring element with a 5x5 cross, iterating two times
11476 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
11478 # Read the shape from the file diamond.shape, iterating two times.
11479 # The file diamond.shape may contain a pattern of characters like this
11485 # The specified columns and rows are ignored
11486 # but the anchor point coordinates are not
11487 ocv=dilate:0x0+2x2/custom=diamond.shape|2
11492 Erode an image by using a specific structuring element.
11493 It corresponds to the libopencv function @code{cvErode}.
11495 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
11496 with the same syntax and semantics as the @ref{dilate} filter.
11500 Smooth the input video.
11502 The filter takes the following parameters:
11503 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
11505 @var{type} is the type of smooth filter to apply, and must be one of
11506 the following values: "blur", "blur_no_scale", "median", "gaussian",
11507 or "bilateral". The default value is "gaussian".
11509 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
11510 depend on the smooth type. @var{param1} and
11511 @var{param2} accept integer positive values or 0. @var{param3} and
11512 @var{param4} accept floating point values.
11514 The default value for @var{param1} is 3. The default value for the
11515 other parameters is 0.
11517 These parameters correspond to the parameters assigned to the
11518 libopencv function @code{cvSmooth}.
11520 @section oscilloscope
11522 2D Video Oscilloscope.
11524 Useful to measure spatial impulse, step responses, chroma delays, etc.
11526 It accepts the following parameters:
11530 Set scope center x position.
11533 Set scope center y position.
11536 Set scope size, relative to frame diagonal.
11539 Set scope tilt/rotation.
11545 Set trace center x position.
11548 Set trace center y position.
11551 Set trace width, relative to width of frame.
11554 Set trace height, relative to height of frame.
11557 Set which components to trace. By default it traces first three components.
11560 Draw trace grid. By default is enabled.
11563 Draw some statistics. By default is enabled.
11566 Draw scope. By default is enabled.
11569 @subsection Examples
11573 Inspect full first row of video frame.
11575 oscilloscope=x=0.5:y=0:s=1
11579 Inspect full last row of video frame.
11581 oscilloscope=x=0.5:y=1:s=1
11585 Inspect full 5th line of video frame of height 1080.
11587 oscilloscope=x=0.5:y=5/1080:s=1
11591 Inspect full last column of video frame.
11593 oscilloscope=x=1:y=0.5:s=1:t=1
11601 Overlay one video on top of another.
11603 It takes two inputs and has one output. The first input is the "main"
11604 video on which the second input is overlaid.
11606 It accepts the following parameters:
11608 A description of the accepted options follows.
11613 Set the expression for the x and y coordinates of the overlaid video
11614 on the main video. Default value is "0" for both expressions. In case
11615 the expression is invalid, it is set to a huge value (meaning that the
11616 overlay will not be displayed within the output visible area).
11619 See @ref{framesync}.
11622 Set when the expressions for @option{x}, and @option{y} are evaluated.
11624 It accepts the following values:
11627 only evaluate expressions once during the filter initialization or
11628 when a command is processed
11631 evaluate expressions for each incoming frame
11634 Default value is @samp{frame}.
11637 See @ref{framesync}.
11640 Set the format for the output video.
11642 It accepts the following values:
11645 force YUV420 output
11648 force YUV422 output
11651 force YUV444 output
11654 force packed RGB output
11657 force planar RGB output
11660 automatically pick format
11663 Default value is @samp{yuv420}.
11666 See @ref{framesync}.
11669 Set format of alpha of the overlaid video, it can be @var{straight} or
11670 @var{premultiplied}. Default is @var{straight}.
11673 The @option{x}, and @option{y} expressions can contain the following
11679 The main input width and height.
11683 The overlay input width and height.
11687 The computed values for @var{x} and @var{y}. They are evaluated for
11692 horizontal and vertical chroma subsample values of the output
11693 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
11697 the number of input frame, starting from 0
11700 the position in the file of the input frame, NAN if unknown
11703 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
11707 This filter also supports the @ref{framesync} options.
11709 Note that the @var{n}, @var{pos}, @var{t} variables are available only
11710 when evaluation is done @emph{per frame}, and will evaluate to NAN
11711 when @option{eval} is set to @samp{init}.
11713 Be aware that frames are taken from each input video in timestamp
11714 order, hence, if their initial timestamps differ, it is a good idea
11715 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
11716 have them begin in the same zero timestamp, as the example for
11717 the @var{movie} filter does.
11719 You can chain together more overlays but you should test the
11720 efficiency of such approach.
11722 @subsection Commands
11724 This filter supports the following commands:
11728 Modify the x and y of the overlay input.
11729 The command accepts the same syntax of the corresponding option.
11731 If the specified expression is not valid, it is kept at its current
11735 @subsection Examples
11739 Draw the overlay at 10 pixels from the bottom right corner of the main
11742 overlay=main_w-overlay_w-10:main_h-overlay_h-10
11745 Using named options the example above becomes:
11747 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
11751 Insert a transparent PNG logo in the bottom left corner of the input,
11752 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
11754 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
11758 Insert 2 different transparent PNG logos (second logo on bottom
11759 right corner) using the @command{ffmpeg} tool:
11761 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
11765 Add a transparent color layer on top of the main video; @code{WxH}
11766 must specify the size of the main input to the overlay filter:
11768 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
11772 Play an original video and a filtered version (here with the deshake
11773 filter) side by side using the @command{ffplay} tool:
11775 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
11778 The above command is the same as:
11780 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
11784 Make a sliding overlay appearing from the left to the right top part of the
11785 screen starting since time 2:
11787 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
11791 Compose output by putting two input videos side to side:
11793 ffmpeg -i left.avi -i right.avi -filter_complex "
11794 nullsrc=size=200x100 [background];
11795 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
11796 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
11797 [background][left] overlay=shortest=1 [background+left];
11798 [background+left][right] overlay=shortest=1:x=100 [left+right]
11803 Mask 10-20 seconds of a video by applying the delogo filter to a section
11805 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
11806 -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]'
11811 Chain several overlays in cascade:
11813 nullsrc=s=200x200 [bg];
11814 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
11815 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
11816 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
11817 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
11818 [in3] null, [mid2] overlay=100:100 [out0]
11825 Apply Overcomplete Wavelet denoiser.
11827 The filter accepts the following options:
11833 Larger depth values will denoise lower frequency components more, but
11834 slow down filtering.
11836 Must be an int in the range 8-16, default is @code{8}.
11838 @item luma_strength, ls
11841 Must be a double value in the range 0-1000, default is @code{1.0}.
11843 @item chroma_strength, cs
11844 Set chroma strength.
11846 Must be a double value in the range 0-1000, default is @code{1.0}.
11852 Add paddings to the input image, and place the original input at the
11853 provided @var{x}, @var{y} coordinates.
11855 It accepts the following parameters:
11860 Specify an expression for the size of the output image with the
11861 paddings added. If the value for @var{width} or @var{height} is 0, the
11862 corresponding input size is used for the output.
11864 The @var{width} expression can reference the value set by the
11865 @var{height} expression, and vice versa.
11867 The default value of @var{width} and @var{height} is 0.
11871 Specify the offsets to place the input image at within the padded area,
11872 with respect to the top/left border of the output image.
11874 The @var{x} expression can reference the value set by the @var{y}
11875 expression, and vice versa.
11877 The default value of @var{x} and @var{y} is 0.
11879 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
11880 so the input image is centered on the padded area.
11883 Specify the color of the padded area. For the syntax of this option,
11884 check the "Color" section in the ffmpeg-utils manual.
11886 The default value of @var{color} is "black".
11889 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
11891 It accepts the following values:
11895 Only evaluate expressions once during the filter initialization or when
11896 a command is processed.
11899 Evaluate expressions for each incoming frame.
11903 Default value is @samp{init}.
11906 Pad to aspect instead to a resolution.
11910 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
11911 options are expressions containing the following constants:
11916 The input video width and height.
11920 These are the same as @var{in_w} and @var{in_h}.
11924 The output width and height (the size of the padded area), as
11925 specified by the @var{width} and @var{height} expressions.
11929 These are the same as @var{out_w} and @var{out_h}.
11933 The x and y offsets as specified by the @var{x} and @var{y}
11934 expressions, or NAN if not yet specified.
11937 same as @var{iw} / @var{ih}
11940 input sample aspect ratio
11943 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
11947 The horizontal and vertical chroma subsample values. For example for the
11948 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
11951 @subsection Examples
11955 Add paddings with the color "violet" to the input video. The output video
11956 size is 640x480, and the top-left corner of the input video is placed at
11959 pad=640:480:0:40:violet
11962 The example above is equivalent to the following command:
11964 pad=width=640:height=480:x=0:y=40:color=violet
11968 Pad the input to get an output with dimensions increased by 3/2,
11969 and put the input video at the center of the padded area:
11971 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
11975 Pad the input to get a squared output with size equal to the maximum
11976 value between the input width and height, and put the input video at
11977 the center of the padded area:
11979 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
11983 Pad the input to get a final w/h ratio of 16:9:
11985 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
11989 In case of anamorphic video, in order to set the output display aspect
11990 correctly, it is necessary to use @var{sar} in the expression,
11991 according to the relation:
11993 (ih * X / ih) * sar = output_dar
11994 X = output_dar / sar
11997 Thus the previous example needs to be modified to:
11999 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
12003 Double the output size and put the input video in the bottom-right
12004 corner of the output padded area:
12006 pad="2*iw:2*ih:ow-iw:oh-ih"
12010 @anchor{palettegen}
12011 @section palettegen
12013 Generate one palette for a whole video stream.
12015 It accepts the following options:
12019 Set the maximum number of colors to quantize in the palette.
12020 Note: the palette will still contain 256 colors; the unused palette entries
12023 @item reserve_transparent
12024 Create a palette of 255 colors maximum and reserve the last one for
12025 transparency. Reserving the transparency color is useful for GIF optimization.
12026 If not set, the maximum of colors in the palette will be 256. You probably want
12027 to disable this option for a standalone image.
12030 @item transparency_color
12031 Set the color that will be used as background for transparency.
12034 Set statistics mode.
12036 It accepts the following values:
12039 Compute full frame histograms.
12041 Compute histograms only for the part that differs from previous frame. This
12042 might be relevant to give more importance to the moving part of your input if
12043 the background is static.
12045 Compute new histogram for each frame.
12048 Default value is @var{full}.
12051 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
12052 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
12053 color quantization of the palette. This information is also visible at
12054 @var{info} logging level.
12056 @subsection Examples
12060 Generate a representative palette of a given video using @command{ffmpeg}:
12062 ffmpeg -i input.mkv -vf palettegen palette.png
12066 @section paletteuse
12068 Use a palette to downsample an input video stream.
12070 The filter takes two inputs: one video stream and a palette. The palette must
12071 be a 256 pixels image.
12073 It accepts the following options:
12077 Select dithering mode. Available algorithms are:
12080 Ordered 8x8 bayer dithering (deterministic)
12082 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
12083 Note: this dithering is sometimes considered "wrong" and is included as a
12085 @item floyd_steinberg
12086 Floyd and Steingberg dithering (error diffusion)
12088 Frankie Sierra dithering v2 (error diffusion)
12090 Frankie Sierra dithering v2 "Lite" (error diffusion)
12093 Default is @var{sierra2_4a}.
12096 When @var{bayer} dithering is selected, this option defines the scale of the
12097 pattern (how much the crosshatch pattern is visible). A low value means more
12098 visible pattern for less banding, and higher value means less visible pattern
12099 at the cost of more banding.
12101 The option must be an integer value in the range [0,5]. Default is @var{2}.
12104 If set, define the zone to process
12108 Only the changing rectangle will be reprocessed. This is similar to GIF
12109 cropping/offsetting compression mechanism. This option can be useful for speed
12110 if only a part of the image is changing, and has use cases such as limiting the
12111 scope of the error diffusal @option{dither} to the rectangle that bounds the
12112 moving scene (it leads to more deterministic output if the scene doesn't change
12113 much, and as a result less moving noise and better GIF compression).
12116 Default is @var{none}.
12119 Take new palette for each output frame.
12121 @item alpha_threshold
12122 Sets the alpha threshold for transparency. Alpha values above this threshold
12123 will be treated as completely opaque, and values below this threshold will be
12124 treated as completely transparent.
12126 The option must be an integer value in the range [0,255]. Default is @var{128}.
12129 @subsection Examples
12133 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
12134 using @command{ffmpeg}:
12136 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
12140 @section perspective
12142 Correct perspective of video not recorded perpendicular to the screen.
12144 A description of the accepted parameters follows.
12155 Set coordinates expression for top left, top right, bottom left and bottom right corners.
12156 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
12157 If the @code{sense} option is set to @code{source}, then the specified points will be sent
12158 to the corners of the destination. If the @code{sense} option is set to @code{destination},
12159 then the corners of the source will be sent to the specified coordinates.
12161 The expressions can use the following variables:
12166 the width and height of video frame.
12170 Output frame count.
12173 @item interpolation
12174 Set interpolation for perspective correction.
12176 It accepts the following values:
12182 Default value is @samp{linear}.
12185 Set interpretation of coordinate options.
12187 It accepts the following values:
12191 Send point in the source specified by the given coordinates to
12192 the corners of the destination.
12194 @item 1, destination
12196 Send the corners of the source to the point in the destination specified
12197 by the given coordinates.
12199 Default value is @samp{source}.
12203 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
12205 It accepts the following values:
12208 only evaluate expressions once during the filter initialization or
12209 when a command is processed
12212 evaluate expressions for each incoming frame
12215 Default value is @samp{init}.
12220 Delay interlaced video by one field time so that the field order changes.
12222 The intended use is to fix PAL movies that have been captured with the
12223 opposite field order to the film-to-video transfer.
12225 A description of the accepted parameters follows.
12231 It accepts the following values:
12234 Capture field order top-first, transfer bottom-first.
12235 Filter will delay the bottom field.
12238 Capture field order bottom-first, transfer top-first.
12239 Filter will delay the top field.
12242 Capture and transfer with the same field order. This mode only exists
12243 for the documentation of the other options to refer to, but if you
12244 actually select it, the filter will faithfully do nothing.
12247 Capture field order determined automatically by field flags, transfer
12249 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
12250 basis using field flags. If no field information is available,
12251 then this works just like @samp{u}.
12254 Capture unknown or varying, transfer opposite.
12255 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
12256 analyzing the images and selecting the alternative that produces best
12257 match between the fields.
12260 Capture top-first, transfer unknown or varying.
12261 Filter selects among @samp{t} and @samp{p} using image analysis.
12264 Capture bottom-first, transfer unknown or varying.
12265 Filter selects among @samp{b} and @samp{p} using image analysis.
12268 Capture determined by field flags, transfer unknown or varying.
12269 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
12270 image analysis. If no field information is available, then this works just
12271 like @samp{U}. This is the default mode.
12274 Both capture and transfer unknown or varying.
12275 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
12279 @section pixdesctest
12281 Pixel format descriptor test filter, mainly useful for internal
12282 testing. The output video should be equal to the input video.
12286 format=monow, pixdesctest
12289 can be used to test the monowhite pixel format descriptor definition.
12293 Display sample values of color channels. Mainly useful for checking color
12294 and levels. Minimum supported resolution is 640x480.
12296 The filters accept the following options:
12300 Set scope X position, relative offset on X axis.
12303 Set scope Y position, relative offset on Y axis.
12312 Set window opacity. This window also holds statistics about pixel area.
12315 Set window X position, relative offset on X axis.
12318 Set window Y position, relative offset on Y axis.
12323 Enable the specified chain of postprocessing subfilters using libpostproc. This
12324 library should be automatically selected with a GPL build (@code{--enable-gpl}).
12325 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
12326 Each subfilter and some options have a short and a long name that can be used
12327 interchangeably, i.e. dr/dering are the same.
12329 The filters accept the following options:
12333 Set postprocessing subfilters string.
12336 All subfilters share common options to determine their scope:
12340 Honor the quality commands for this subfilter.
12343 Do chrominance filtering, too (default).
12346 Do luminance filtering only (no chrominance).
12349 Do chrominance filtering only (no luminance).
12352 These options can be appended after the subfilter name, separated by a '|'.
12354 Available subfilters are:
12357 @item hb/hdeblock[|difference[|flatness]]
12358 Horizontal deblocking filter
12361 Difference factor where higher values mean more deblocking (default: @code{32}).
12363 Flatness threshold where lower values mean more deblocking (default: @code{39}).
12366 @item vb/vdeblock[|difference[|flatness]]
12367 Vertical deblocking filter
12370 Difference factor where higher values mean more deblocking (default: @code{32}).
12372 Flatness threshold where lower values mean more deblocking (default: @code{39}).
12375 @item ha/hadeblock[|difference[|flatness]]
12376 Accurate horizontal deblocking filter
12379 Difference factor where higher values mean more deblocking (default: @code{32}).
12381 Flatness threshold where lower values mean more deblocking (default: @code{39}).
12384 @item va/vadeblock[|difference[|flatness]]
12385 Accurate vertical deblocking filter
12388 Difference factor where higher values mean more deblocking (default: @code{32}).
12390 Flatness threshold where lower values mean more deblocking (default: @code{39}).
12394 The horizontal and vertical deblocking filters share the difference and
12395 flatness values so you cannot set different horizontal and vertical
12399 @item h1/x1hdeblock
12400 Experimental horizontal deblocking filter
12402 @item v1/x1vdeblock
12403 Experimental vertical deblocking filter
12408 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
12411 larger -> stronger filtering
12413 larger -> stronger filtering
12415 larger -> stronger filtering
12418 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
12421 Stretch luminance to @code{0-255}.
12424 @item lb/linblenddeint
12425 Linear blend deinterlacing filter that deinterlaces the given block by
12426 filtering all lines with a @code{(1 2 1)} filter.
12428 @item li/linipoldeint
12429 Linear interpolating deinterlacing filter that deinterlaces the given block by
12430 linearly interpolating every second line.
12432 @item ci/cubicipoldeint
12433 Cubic interpolating deinterlacing filter deinterlaces the given block by
12434 cubically interpolating every second line.
12436 @item md/mediandeint
12437 Median deinterlacing filter that deinterlaces the given block by applying a
12438 median filter to every second line.
12440 @item fd/ffmpegdeint
12441 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
12442 second line with a @code{(-1 4 2 4 -1)} filter.
12445 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
12446 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
12448 @item fq/forceQuant[|quantizer]
12449 Overrides the quantizer table from the input with the constant quantizer you
12457 Default pp filter combination (@code{hb|a,vb|a,dr|a})
12460 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
12463 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
12466 @subsection Examples
12470 Apply horizontal and vertical deblocking, deringing and automatic
12471 brightness/contrast:
12477 Apply default filters without brightness/contrast correction:
12483 Apply default filters and temporal denoiser:
12485 pp=default/tmpnoise|1|2|3
12489 Apply deblocking on luminance only, and switch vertical deblocking on or off
12490 automatically depending on available CPU time:
12497 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
12498 similar to spp = 6 with 7 point DCT, where only the center sample is
12501 The filter accepts the following options:
12505 Force a constant quantization parameter. It accepts an integer in range
12506 0 to 63. If not set, the filter will use the QP from the video stream
12510 Set thresholding mode. Available modes are:
12514 Set hard thresholding.
12516 Set soft thresholding (better de-ringing effect, but likely blurrier).
12518 Set medium thresholding (good results, default).
12522 @section premultiply
12523 Apply alpha premultiply effect to input video stream using first plane
12524 of second stream as alpha.
12526 Both streams must have same dimensions and same pixel format.
12528 The filter accepts the following option:
12532 Set which planes will be processed, unprocessed planes will be copied.
12533 By default value 0xf, all planes will be processed.
12536 Do not require 2nd input for processing, instead use alpha plane from input stream.
12540 Apply prewitt operator to input video stream.
12542 The filter accepts the following option:
12546 Set which planes will be processed, unprocessed planes will be copied.
12547 By default value 0xf, all planes will be processed.
12550 Set value which will be multiplied with filtered result.
12553 Set value which will be added to filtered result.
12556 @anchor{program_opencl}
12557 @section program_opencl
12559 Filter video using an OpenCL program.
12564 OpenCL program source file.
12567 Kernel name in program.
12570 Number of inputs to the filter. Defaults to 1.
12573 Size of output frames. Defaults to the same as the first input.
12577 The program source file must contain a kernel function with the given name,
12578 which will be run once for each plane of the output. Each run on a plane
12579 gets enqueued as a separate 2D global NDRange with one work-item for each
12580 pixel to be generated. The global ID offset for each work-item is therefore
12581 the coordinates of a pixel in the destination image.
12583 The kernel function needs to take the following arguments:
12586 Destination image, @var{__write_only image2d_t}.
12588 This image will become the output; the kernel should write all of it.
12590 Frame index, @var{unsigned int}.
12592 This is a counter starting from zero and increasing by one for each frame.
12594 Source images, @var{__read_only image2d_t}.
12596 These are the most recent images on each input. The kernel may read from
12597 them to generate the output, but they can't be written to.
12604 Copy the input to the output (output must be the same size as the input).
12606 __kernel void copy(__write_only image2d_t destination,
12607 unsigned int index,
12608 __read_only image2d_t source)
12610 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
12612 int2 location = (int2)(get_global_id(0), get_global_id(1));
12614 float4 value = read_imagef(source, sampler, location);
12616 write_imagef(destination, location, value);
12621 Apply a simple transformation, rotating the input by an amount increasing
12622 with the index counter. Pixel values are linearly interpolated by the
12623 sampler, and the output need not have the same dimensions as the input.
12625 __kernel void rotate_image(__write_only image2d_t dst,
12626 unsigned int index,
12627 __read_only image2d_t src)
12629 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
12630 CLK_FILTER_LINEAR);
12632 float angle = (float)index / 100.0f;
12634 float2 dst_dim = convert_float2(get_image_dim(dst));
12635 float2 src_dim = convert_float2(get_image_dim(src));
12637 float2 dst_cen = dst_dim / 2.0f;
12638 float2 src_cen = src_dim / 2.0f;
12640 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
12642 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
12644 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
12645 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
12647 src_pos = src_pos * src_dim / dst_dim;
12649 float2 src_loc = src_pos + src_cen;
12651 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
12652 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
12653 write_imagef(dst, dst_loc, 0.5f);
12655 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
12660 Blend two inputs together, with the amount of each input used varying
12661 with the index counter.
12663 __kernel void blend_images(__write_only image2d_t dst,
12664 unsigned int index,
12665 __read_only image2d_t src1,
12666 __read_only image2d_t src2)
12668 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
12669 CLK_FILTER_LINEAR);
12671 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
12673 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
12674 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
12675 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
12677 float4 val1 = read_imagef(src1, sampler, src1_loc);
12678 float4 val2 = read_imagef(src2, sampler, src2_loc);
12680 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
12686 @section pseudocolor
12688 Alter frame colors in video with pseudocolors.
12690 This filter accept the following options:
12694 set pixel first component expression
12697 set pixel second component expression
12700 set pixel third component expression
12703 set pixel fourth component expression, corresponds to the alpha component
12706 set component to use as base for altering colors
12709 Each of them specifies the expression to use for computing the lookup table for
12710 the corresponding pixel component values.
12712 The expressions can contain the following constants and functions:
12717 The input width and height.
12720 The input value for the pixel component.
12722 @item ymin, umin, vmin, amin
12723 The minimum allowed component value.
12725 @item ymax, umax, vmax, amax
12726 The maximum allowed component value.
12729 All expressions default to "val".
12731 @subsection Examples
12735 Change too high luma values to gradient:
12737 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'"
12743 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
12744 Ratio) between two input videos.
12746 This filter takes in input two input videos, the first input is
12747 considered the "main" source and is passed unchanged to the
12748 output. The second input is used as a "reference" video for computing
12751 Both video inputs must have the same resolution and pixel format for
12752 this filter to work correctly. Also it assumes that both inputs
12753 have the same number of frames, which are compared one by one.
12755 The obtained average PSNR is printed through the logging system.
12757 The filter stores the accumulated MSE (mean squared error) of each
12758 frame, and at the end of the processing it is averaged across all frames
12759 equally, and the following formula is applied to obtain the PSNR:
12762 PSNR = 10*log10(MAX^2/MSE)
12765 Where MAX is the average of the maximum values of each component of the
12768 The description of the accepted parameters follows.
12771 @item stats_file, f
12772 If specified the filter will use the named file to save the PSNR of
12773 each individual frame. When filename equals "-" the data is sent to
12776 @item stats_version
12777 Specifies which version of the stats file format to use. Details of
12778 each format are written below.
12779 Default value is 1.
12781 @item stats_add_max
12782 Determines whether the max value is output to the stats log.
12783 Default value is 0.
12784 Requires stats_version >= 2. If this is set and stats_version < 2,
12785 the filter will return an error.
12788 This filter also supports the @ref{framesync} options.
12790 The file printed if @var{stats_file} is selected, contains a sequence of
12791 key/value pairs of the form @var{key}:@var{value} for each compared
12794 If a @var{stats_version} greater than 1 is specified, a header line precedes
12795 the list of per-frame-pair stats, with key value pairs following the frame
12796 format with the following parameters:
12799 @item psnr_log_version
12800 The version of the log file format. Will match @var{stats_version}.
12803 A comma separated list of the per-frame-pair parameters included in
12807 A description of each shown per-frame-pair parameter follows:
12811 sequential number of the input frame, starting from 1
12814 Mean Square Error pixel-by-pixel average difference of the compared
12815 frames, averaged over all the image components.
12817 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
12818 Mean Square Error pixel-by-pixel average difference of the compared
12819 frames for the component specified by the suffix.
12821 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
12822 Peak Signal to Noise ratio of the compared frames for the component
12823 specified by the suffix.
12825 @item max_avg, max_y, max_u, max_v
12826 Maximum allowed value for each channel, and average over all
12832 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
12833 [main][ref] psnr="stats_file=stats.log" [out]
12836 On this example the input file being processed is compared with the
12837 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
12838 is stored in @file{stats.log}.
12843 Pulldown reversal (inverse telecine) filter, capable of handling mixed
12844 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
12847 The pullup filter is designed to take advantage of future context in making
12848 its decisions. This filter is stateless in the sense that it does not lock
12849 onto a pattern to follow, but it instead looks forward to the following
12850 fields in order to identify matches and rebuild progressive frames.
12852 To produce content with an even framerate, insert the fps filter after
12853 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
12854 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
12856 The filter accepts the following options:
12863 These options set the amount of "junk" to ignore at the left, right, top, and
12864 bottom of the image, respectively. Left and right are in units of 8 pixels,
12865 while top and bottom are in units of 2 lines.
12866 The default is 8 pixels on each side.
12869 Set the strict breaks. Setting this option to 1 will reduce the chances of
12870 filter generating an occasional mismatched frame, but it may also cause an
12871 excessive number of frames to be dropped during high motion sequences.
12872 Conversely, setting it to -1 will make filter match fields more easily.
12873 This may help processing of video where there is slight blurring between
12874 the fields, but may also cause there to be interlaced frames in the output.
12875 Default value is @code{0}.
12878 Set the metric plane to use. It accepts the following values:
12884 Use chroma blue plane.
12887 Use chroma red plane.
12890 This option may be set to use chroma plane instead of the default luma plane
12891 for doing filter's computations. This may improve accuracy on very clean
12892 source material, but more likely will decrease accuracy, especially if there
12893 is chroma noise (rainbow effect) or any grayscale video.
12894 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
12895 load and make pullup usable in realtime on slow machines.
12898 For best results (without duplicated frames in the output file) it is
12899 necessary to change the output frame rate. For example, to inverse
12900 telecine NTSC input:
12902 ffmpeg -i input -vf pullup -r 24000/1001 ...
12907 Change video quantization parameters (QP).
12909 The filter accepts the following option:
12913 Set expression for quantization parameter.
12916 The expression is evaluated through the eval API and can contain, among others,
12917 the following constants:
12921 1 if index is not 129, 0 otherwise.
12924 Sequential index starting from -129 to 128.
12927 @subsection Examples
12931 Some equation like:
12939 Flush video frames from internal cache of frames into a random order.
12940 No frame is discarded.
12941 Inspired by @ref{frei0r} nervous filter.
12945 Set size in number of frames of internal cache, in range from @code{2} to
12946 @code{512}. Default is @code{30}.
12949 Set seed for random number generator, must be an integer included between
12950 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
12951 less than @code{0}, the filter will try to use a good random seed on a
12955 @section readeia608
12957 Read closed captioning (EIA-608) information from the top lines of a video frame.
12959 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
12960 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
12961 with EIA-608 data (starting from 0). A description of each metadata value follows:
12964 @item lavfi.readeia608.X.cc
12965 The two bytes stored as EIA-608 data (printed in hexadecimal).
12967 @item lavfi.readeia608.X.line
12968 The number of the line on which the EIA-608 data was identified and read.
12971 This filter accepts the following options:
12975 Set the line to start scanning for EIA-608 data. Default is @code{0}.
12978 Set the line to end scanning for EIA-608 data. Default is @code{29}.
12981 Set minimal acceptable amplitude change for sync codes detection.
12982 Default is @code{0.2}. Allowed range is @code{[0.001 - 1]}.
12985 Set the ratio of width reserved for sync code detection.
12986 Default is @code{0.27}. Allowed range is @code{[0.01 - 0.7]}.
12989 Set the max peaks height difference for sync code detection.
12990 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
12993 Set max peaks period difference for sync code detection.
12994 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
12997 Set the first two max start code bits differences.
12998 Default is @code{0.02}. Allowed range is @code{[0.0 - 0.5]}.
13001 Set the minimum ratio of bits height compared to 3rd start code bit.
13002 Default is @code{0.75}. Allowed range is @code{[0.01 - 1]}.
13005 Set the white color threshold. Default is @code{0.35}. Allowed range is @code{[0.1 - 1]}.
13008 Set the black color threshold. Default is @code{0.15}. Allowed range is @code{[0.0 - 0.5]}.
13011 Enable checking the parity bit. In the event of a parity error, the filter will output
13012 @code{0x00} for that character. Default is false.
13015 @subsection Examples
13019 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
13021 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
13027 Read vertical interval timecode (VITC) information from the top lines of a
13030 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
13031 timecode value, if a valid timecode has been detected. Further metadata key
13032 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
13033 timecode data has been found or not.
13035 This filter accepts the following options:
13039 Set the maximum number of lines to scan for VITC data. If the value is set to
13040 @code{-1} the full video frame is scanned. Default is @code{45}.
13043 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
13044 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
13047 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
13048 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
13051 @subsection Examples
13055 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
13056 draw @code{--:--:--:--} as a placeholder:
13058 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
13064 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
13066 Destination pixel at position (X, Y) will be picked from source (x, y) position
13067 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
13068 value for pixel will be used for destination pixel.
13070 Xmap and Ymap input video streams must be of same dimensions. Output video stream
13071 will have Xmap/Ymap video stream dimensions.
13072 Xmap and Ymap input video streams are 16bit depth, single channel.
13074 @section removegrain
13076 The removegrain filter is a spatial denoiser for progressive video.
13080 Set mode for the first plane.
13083 Set mode for the second plane.
13086 Set mode for the third plane.
13089 Set mode for the fourth plane.
13092 Range of mode is from 0 to 24. Description of each mode follows:
13096 Leave input plane unchanged. Default.
13099 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
13102 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
13105 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
13108 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
13109 This is equivalent to a median filter.
13112 Line-sensitive clipping giving the minimal change.
13115 Line-sensitive clipping, intermediate.
13118 Line-sensitive clipping, intermediate.
13121 Line-sensitive clipping, intermediate.
13124 Line-sensitive clipping on a line where the neighbours pixels are the closest.
13127 Replaces the target pixel with the closest neighbour.
13130 [1 2 1] horizontal and vertical kernel blur.
13136 Bob mode, interpolates top field from the line where the neighbours
13137 pixels are the closest.
13140 Bob mode, interpolates bottom field from the line where the neighbours
13141 pixels are the closest.
13144 Bob mode, interpolates top field. Same as 13 but with a more complicated
13145 interpolation formula.
13148 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
13149 interpolation formula.
13152 Clips the pixel with the minimum and maximum of respectively the maximum and
13153 minimum of each pair of opposite neighbour pixels.
13156 Line-sensitive clipping using opposite neighbours whose greatest distance from
13157 the current pixel is minimal.
13160 Replaces the pixel with the average of its 8 neighbours.
13163 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
13166 Clips pixels using the averages of opposite neighbour.
13169 Same as mode 21 but simpler and faster.
13172 Small edge and halo removal, but reputed useless.
13178 @section removelogo
13180 Suppress a TV station logo, using an image file to determine which
13181 pixels comprise the logo. It works by filling in the pixels that
13182 comprise the logo with neighboring pixels.
13184 The filter accepts the following options:
13188 Set the filter bitmap file, which can be any image format supported by
13189 libavformat. The width and height of the image file must match those of the
13190 video stream being processed.
13193 Pixels in the provided bitmap image with a value of zero are not
13194 considered part of the logo, non-zero pixels are considered part of
13195 the logo. If you use white (255) for the logo and black (0) for the
13196 rest, you will be safe. For making the filter bitmap, it is
13197 recommended to take a screen capture of a black frame with the logo
13198 visible, and then using a threshold filter followed by the erode
13199 filter once or twice.
13201 If needed, little splotches can be fixed manually. Remember that if
13202 logo pixels are not covered, the filter quality will be much
13203 reduced. Marking too many pixels as part of the logo does not hurt as
13204 much, but it will increase the amount of blurring needed to cover over
13205 the image and will destroy more information than necessary, and extra
13206 pixels will slow things down on a large logo.
13208 @section repeatfields
13210 This filter uses the repeat_field flag from the Video ES headers and hard repeats
13211 fields based on its value.
13215 Reverse a video clip.
13217 Warning: This filter requires memory to buffer the entire clip, so trimming
13220 @subsection Examples
13224 Take the first 5 seconds of a clip, and reverse it.
13231 Apply roberts cross operator to input video stream.
13233 The filter accepts the following option:
13237 Set which planes will be processed, unprocessed planes will be copied.
13238 By default value 0xf, all planes will be processed.
13241 Set value which will be multiplied with filtered result.
13244 Set value which will be added to filtered result.
13249 Rotate video by an arbitrary angle expressed in radians.
13251 The filter accepts the following options:
13253 A description of the optional parameters follows.
13256 Set an expression for the angle by which to rotate the input video
13257 clockwise, expressed as a number of radians. A negative value will
13258 result in a counter-clockwise rotation. By default it is set to "0".
13260 This expression is evaluated for each frame.
13263 Set the output width expression, default value is "iw".
13264 This expression is evaluated just once during configuration.
13267 Set the output height expression, default value is "ih".
13268 This expression is evaluated just once during configuration.
13271 Enable bilinear interpolation if set to 1, a value of 0 disables
13272 it. Default value is 1.
13275 Set the color used to fill the output area not covered by the rotated
13276 image. For the general syntax of this option, check the "Color" section in the
13277 ffmpeg-utils manual. If the special value "none" is selected then no
13278 background is printed (useful for example if the background is never shown).
13280 Default value is "black".
13283 The expressions for the angle and the output size can contain the
13284 following constants and functions:
13288 sequential number of the input frame, starting from 0. It is always NAN
13289 before the first frame is filtered.
13292 time in seconds of the input frame, it is set to 0 when the filter is
13293 configured. It is always NAN before the first frame is filtered.
13297 horizontal and vertical chroma subsample values. For example for the
13298 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
13302 the input video width and height
13306 the output width and height, that is the size of the padded area as
13307 specified by the @var{width} and @var{height} expressions
13311 the minimal width/height required for completely containing the input
13312 video rotated by @var{a} radians.
13314 These are only available when computing the @option{out_w} and
13315 @option{out_h} expressions.
13318 @subsection Examples
13322 Rotate the input by PI/6 radians clockwise:
13328 Rotate the input by PI/6 radians counter-clockwise:
13334 Rotate the input by 45 degrees clockwise:
13340 Apply a constant rotation with period T, starting from an angle of PI/3:
13342 rotate=PI/3+2*PI*t/T
13346 Make the input video rotation oscillating with a period of T
13347 seconds and an amplitude of A radians:
13349 rotate=A*sin(2*PI/T*t)
13353 Rotate the video, output size is chosen so that the whole rotating
13354 input video is always completely contained in the output:
13356 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
13360 Rotate the video, reduce the output size so that no background is ever
13363 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
13367 @subsection Commands
13369 The filter supports the following commands:
13373 Set the angle expression.
13374 The command accepts the same syntax of the corresponding option.
13376 If the specified expression is not valid, it is kept at its current
13382 Apply Shape Adaptive Blur.
13384 The filter accepts the following options:
13387 @item luma_radius, lr
13388 Set luma blur filter strength, must be a value in range 0.1-4.0, default
13389 value is 1.0. A greater value will result in a more blurred image, and
13390 in slower processing.
13392 @item luma_pre_filter_radius, lpfr
13393 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
13396 @item luma_strength, ls
13397 Set luma maximum difference between pixels to still be considered, must
13398 be a value in the 0.1-100.0 range, default value is 1.0.
13400 @item chroma_radius, cr
13401 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
13402 greater value will result in a more blurred image, and in slower
13405 @item chroma_pre_filter_radius, cpfr
13406 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
13408 @item chroma_strength, cs
13409 Set chroma maximum difference between pixels to still be considered,
13410 must be a value in the -0.9-100.0 range.
13413 Each chroma option value, if not explicitly specified, is set to the
13414 corresponding luma option value.
13419 Scale (resize) the input video, using the libswscale library.
13421 The scale filter forces the output display aspect ratio to be the same
13422 of the input, by changing the output sample aspect ratio.
13424 If the input image format is different from the format requested by
13425 the next filter, the scale filter will convert the input to the
13428 @subsection Options
13429 The filter accepts the following options, or any of the options
13430 supported by the libswscale scaler.
13432 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
13433 the complete list of scaler options.
13438 Set the output video dimension expression. Default value is the input
13441 If the @var{width} or @var{w} value is 0, the input width is used for
13442 the output. If the @var{height} or @var{h} value is 0, the input height
13443 is used for the output.
13445 If one and only one of the values is -n with n >= 1, the scale filter
13446 will use a value that maintains the aspect ratio of the input image,
13447 calculated from the other specified dimension. After that it will,
13448 however, make sure that the calculated dimension is divisible by n and
13449 adjust the value if necessary.
13451 If both values are -n with n >= 1, the behavior will be identical to
13452 both values being set to 0 as previously detailed.
13454 See below for the list of accepted constants for use in the dimension
13458 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
13462 Only evaluate expressions once during the filter initialization or when a command is processed.
13465 Evaluate expressions for each incoming frame.
13469 Default value is @samp{init}.
13473 Set the interlacing mode. It accepts the following values:
13477 Force interlaced aware scaling.
13480 Do not apply interlaced scaling.
13483 Select interlaced aware scaling depending on whether the source frames
13484 are flagged as interlaced or not.
13487 Default value is @samp{0}.
13490 Set libswscale scaling flags. See
13491 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
13492 complete list of values. If not explicitly specified the filter applies
13496 @item param0, param1
13497 Set libswscale input parameters for scaling algorithms that need them. See
13498 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
13499 complete documentation. If not explicitly specified the filter applies
13505 Set the video size. For the syntax of this option, check the
13506 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
13508 @item in_color_matrix
13509 @item out_color_matrix
13510 Set in/output YCbCr color space type.
13512 This allows the autodetected value to be overridden as well as allows forcing
13513 a specific value used for the output and encoder.
13515 If not specified, the color space type depends on the pixel format.
13521 Choose automatically.
13524 Format conforming to International Telecommunication Union (ITU)
13525 Recommendation BT.709.
13528 Set color space conforming to the United States Federal Communications
13529 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
13532 Set color space conforming to:
13536 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
13539 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
13542 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
13547 Set color space conforming to SMPTE ST 240:1999.
13552 Set in/output YCbCr sample range.
13554 This allows the autodetected value to be overridden as well as allows forcing
13555 a specific value used for the output and encoder. If not specified, the
13556 range depends on the pixel format. Possible values:
13560 Choose automatically.
13563 Set full range (0-255 in case of 8-bit luma).
13565 @item mpeg/limited/tv
13566 Set "MPEG" range (16-235 in case of 8-bit luma).
13569 @item force_original_aspect_ratio
13570 Enable decreasing or increasing output video width or height if necessary to
13571 keep the original aspect ratio. Possible values:
13575 Scale the video as specified and disable this feature.
13578 The output video dimensions will automatically be decreased if needed.
13581 The output video dimensions will automatically be increased if needed.
13585 One useful instance of this option is that when you know a specific device's
13586 maximum allowed resolution, you can use this to limit the output video to
13587 that, while retaining the aspect ratio. For example, device A allows
13588 1280x720 playback, and your video is 1920x800. Using this option (set it to
13589 decrease) and specifying 1280x720 to the command line makes the output
13592 Please note that this is a different thing than specifying -1 for @option{w}
13593 or @option{h}, you still need to specify the output resolution for this option
13598 The values of the @option{w} and @option{h} options are expressions
13599 containing the following constants:
13604 The input width and height
13608 These are the same as @var{in_w} and @var{in_h}.
13612 The output (scaled) width and height
13616 These are the same as @var{out_w} and @var{out_h}
13619 The same as @var{iw} / @var{ih}
13622 input sample aspect ratio
13625 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
13629 horizontal and vertical input chroma subsample values. For example for the
13630 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
13634 horizontal and vertical output chroma subsample values. For example for the
13635 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
13638 @subsection Examples
13642 Scale the input video to a size of 200x100
13647 This is equivalent to:
13658 Specify a size abbreviation for the output size:
13663 which can also be written as:
13669 Scale the input to 2x:
13671 scale=w=2*iw:h=2*ih
13675 The above is the same as:
13677 scale=2*in_w:2*in_h
13681 Scale the input to 2x with forced interlaced scaling:
13683 scale=2*iw:2*ih:interl=1
13687 Scale the input to half size:
13689 scale=w=iw/2:h=ih/2
13693 Increase the width, and set the height to the same size:
13699 Seek Greek harmony:
13706 Increase the height, and set the width to 3/2 of the height:
13708 scale=w=3/2*oh:h=3/5*ih
13712 Increase the size, making the size a multiple of the chroma
13715 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
13719 Increase the width to a maximum of 500 pixels,
13720 keeping the same aspect ratio as the input:
13722 scale=w='min(500\, iw*3/2):h=-1'
13726 @subsection Commands
13728 This filter supports the following commands:
13732 Set the output video dimension expression.
13733 The command accepts the same syntax of the corresponding option.
13735 If the specified expression is not valid, it is kept at its current
13741 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
13742 format conversion on CUDA video frames. Setting the output width and height
13743 works in the same way as for the @var{scale} filter.
13745 The following additional options are accepted:
13748 The pixel format of the output CUDA frames. If set to the string "same" (the
13749 default), the input format will be kept. Note that automatic format negotiation
13750 and conversion is not yet supported for hardware frames
13753 The interpolation algorithm used for resizing. One of the following:
13760 @item cubic2p_bspline
13761 2-parameter cubic (B=1, C=0)
13763 @item cubic2p_catmullrom
13764 2-parameter cubic (B=0, C=1/2)
13766 @item cubic2p_b05c03
13767 2-parameter cubic (B=1/2, C=3/10)
13779 Scale (resize) the input video, based on a reference video.
13781 See the scale filter for available options, scale2ref supports the same but
13782 uses the reference video instead of the main input as basis. scale2ref also
13783 supports the following additional constants for the @option{w} and
13784 @option{h} options:
13789 The main input video's width and height
13792 The same as @var{main_w} / @var{main_h}
13795 The main input video's sample aspect ratio
13797 @item main_dar, mdar
13798 The main input video's display aspect ratio. Calculated from
13799 @code{(main_w / main_h) * main_sar}.
13803 The main input video's horizontal and vertical chroma subsample values.
13804 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
13808 @subsection Examples
13812 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
13814 'scale2ref[b][a];[a][b]overlay'
13818 @anchor{selectivecolor}
13819 @section selectivecolor
13821 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
13822 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
13823 by the "purity" of the color (that is, how saturated it already is).
13825 This filter is similar to the Adobe Photoshop Selective Color tool.
13827 The filter accepts the following options:
13830 @item correction_method
13831 Select color correction method.
13833 Available values are:
13836 Specified adjustments are applied "as-is" (added/subtracted to original pixel
13839 Specified adjustments are relative to the original component value.
13841 Default is @code{absolute}.
13843 Adjustments for red pixels (pixels where the red component is the maximum)
13845 Adjustments for yellow pixels (pixels where the blue component is the minimum)
13847 Adjustments for green pixels (pixels where the green component is the maximum)
13849 Adjustments for cyan pixels (pixels where the red component is the minimum)
13851 Adjustments for blue pixels (pixels where the blue component is the maximum)
13853 Adjustments for magenta pixels (pixels where the green component is the minimum)
13855 Adjustments for white pixels (pixels where all components are greater than 128)
13857 Adjustments for all pixels except pure black and pure white
13859 Adjustments for black pixels (pixels where all components are lesser than 128)
13861 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
13864 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
13865 4 space separated floating point adjustment values in the [-1,1] range,
13866 respectively to adjust the amount of cyan, magenta, yellow and black for the
13867 pixels of its range.
13869 @subsection Examples
13873 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
13874 increase magenta by 27% in blue areas:
13876 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
13880 Use a Photoshop selective color preset:
13882 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
13886 @anchor{separatefields}
13887 @section separatefields
13889 The @code{separatefields} takes a frame-based video input and splits
13890 each frame into its components fields, producing a new half height clip
13891 with twice the frame rate and twice the frame count.
13893 This filter use field-dominance information in frame to decide which
13894 of each pair of fields to place first in the output.
13895 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
13897 @section setdar, setsar
13899 The @code{setdar} filter sets the Display Aspect Ratio for the filter
13902 This is done by changing the specified Sample (aka Pixel) Aspect
13903 Ratio, according to the following equation:
13905 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
13908 Keep in mind that the @code{setdar} filter does not modify the pixel
13909 dimensions of the video frame. Also, the display aspect ratio set by
13910 this filter may be changed by later filters in the filterchain,
13911 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
13914 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
13915 the filter output video.
13917 Note that as a consequence of the application of this filter, the
13918 output display aspect ratio will change according to the equation
13921 Keep in mind that the sample aspect ratio set by the @code{setsar}
13922 filter may be changed by later filters in the filterchain, e.g. if
13923 another "setsar" or a "setdar" filter is applied.
13925 It accepts the following parameters:
13928 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
13929 Set the aspect ratio used by the filter.
13931 The parameter can be a floating point number string, an expression, or
13932 a string of the form @var{num}:@var{den}, where @var{num} and
13933 @var{den} are the numerator and denominator of the aspect ratio. If
13934 the parameter is not specified, it is assumed the value "0".
13935 In case the form "@var{num}:@var{den}" is used, the @code{:} character
13939 Set the maximum integer value to use for expressing numerator and
13940 denominator when reducing the expressed aspect ratio to a rational.
13941 Default value is @code{100}.
13945 The parameter @var{sar} is an expression containing
13946 the following constants:
13950 These are approximated values for the mathematical constants e
13951 (Euler's number), pi (Greek pi), and phi (the golden ratio).
13954 The input width and height.
13957 These are the same as @var{w} / @var{h}.
13960 The input sample aspect ratio.
13963 The input display aspect ratio. It is the same as
13964 (@var{w} / @var{h}) * @var{sar}.
13967 Horizontal and vertical chroma subsample values. For example, for the
13968 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
13971 @subsection Examples
13976 To change the display aspect ratio to 16:9, specify one of the following:
13983 To change the sample aspect ratio to 10:11, specify:
13989 To set a display aspect ratio of 16:9, and specify a maximum integer value of
13990 1000 in the aspect ratio reduction, use the command:
13992 setdar=ratio=16/9:max=1000
14000 Force field for the output video frame.
14002 The @code{setfield} filter marks the interlace type field for the
14003 output frames. It does not change the input frame, but only sets the
14004 corresponding property, which affects how the frame is treated by
14005 following filters (e.g. @code{fieldorder} or @code{yadif}).
14007 The filter accepts the following options:
14012 Available values are:
14016 Keep the same field property.
14019 Mark the frame as bottom-field-first.
14022 Mark the frame as top-field-first.
14025 Mark the frame as progressive.
14031 Show a line containing various information for each input video frame.
14032 The input video is not modified.
14034 The shown line contains a sequence of key/value pairs of the form
14035 @var{key}:@var{value}.
14037 The following values are shown in the output:
14041 The (sequential) number of the input frame, starting from 0.
14044 The Presentation TimeStamp of the input frame, expressed as a number of
14045 time base units. The time base unit depends on the filter input pad.
14048 The Presentation TimeStamp of the input frame, expressed as a number of
14052 The position of the frame in the input stream, or -1 if this information is
14053 unavailable and/or meaningless (for example in case of synthetic video).
14056 The pixel format name.
14059 The sample aspect ratio of the input frame, expressed in the form
14060 @var{num}/@var{den}.
14063 The size of the input frame. For the syntax of this option, check the
14064 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
14067 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
14068 for bottom field first).
14071 This is 1 if the frame is a key frame, 0 otherwise.
14074 The picture type of the input frame ("I" for an I-frame, "P" for a
14075 P-frame, "B" for a B-frame, or "?" for an unknown type).
14076 Also refer to the documentation of the @code{AVPictureType} enum and of
14077 the @code{av_get_picture_type_char} function defined in
14078 @file{libavutil/avutil.h}.
14081 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
14083 @item plane_checksum
14084 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
14085 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
14088 @section showpalette
14090 Displays the 256 colors palette of each frame. This filter is only relevant for
14091 @var{pal8} pixel format frames.
14093 It accepts the following option:
14097 Set the size of the box used to represent one palette color entry. Default is
14098 @code{30} (for a @code{30x30} pixel box).
14101 @section shuffleframes
14103 Reorder and/or duplicate and/or drop video frames.
14105 It accepts the following parameters:
14109 Set the destination indexes of input frames.
14110 This is space or '|' separated list of indexes that maps input frames to output
14111 frames. Number of indexes also sets maximal value that each index may have.
14112 '-1' index have special meaning and that is to drop frame.
14115 The first frame has the index 0. The default is to keep the input unchanged.
14117 @subsection Examples
14121 Swap second and third frame of every three frames of the input:
14123 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
14127 Swap 10th and 1st frame of every ten frames of the input:
14129 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
14133 @section shuffleplanes
14135 Reorder and/or duplicate video planes.
14137 It accepts the following parameters:
14142 The index of the input plane to be used as the first output plane.
14145 The index of the input plane to be used as the second output plane.
14148 The index of the input plane to be used as the third output plane.
14151 The index of the input plane to be used as the fourth output plane.
14155 The first plane has the index 0. The default is to keep the input unchanged.
14157 @subsection Examples
14161 Swap the second and third planes of the input:
14163 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
14167 @anchor{signalstats}
14168 @section signalstats
14169 Evaluate various visual metrics that assist in determining issues associated
14170 with the digitization of analog video media.
14172 By default the filter will log these metadata values:
14176 Display the minimal Y value contained within the input frame. Expressed in
14180 Display the Y value at the 10% percentile within the input frame. Expressed in
14184 Display the average Y value within the input frame. Expressed in range of
14188 Display the Y value at the 90% percentile within the input frame. Expressed in
14192 Display the maximum Y value contained within the input frame. Expressed in
14196 Display the minimal U value contained within the input frame. Expressed in
14200 Display the U value at the 10% percentile within the input frame. Expressed in
14204 Display the average U value within the input frame. Expressed in range of
14208 Display the U value at the 90% percentile within the input frame. Expressed in
14212 Display the maximum U value contained within the input frame. Expressed in
14216 Display the minimal V value contained within the input frame. Expressed in
14220 Display the V value at the 10% percentile within the input frame. Expressed in
14224 Display the average V value within the input frame. Expressed in range of
14228 Display the V value at the 90% percentile within the input frame. Expressed in
14232 Display the maximum V value contained within the input frame. Expressed in
14236 Display the minimal saturation value contained within the input frame.
14237 Expressed in range of [0-~181.02].
14240 Display the saturation value at the 10% percentile within the input frame.
14241 Expressed in range of [0-~181.02].
14244 Display the average saturation value within the input frame. Expressed in range
14248 Display the saturation value at the 90% percentile within the input frame.
14249 Expressed in range of [0-~181.02].
14252 Display the maximum saturation value contained within the input frame.
14253 Expressed in range of [0-~181.02].
14256 Display the median value for hue within the input frame. Expressed in range of
14260 Display the average value for hue within the input frame. Expressed in range of
14264 Display the average of sample value difference between all values of the Y
14265 plane in the current frame and corresponding values of the previous input frame.
14266 Expressed in range of [0-255].
14269 Display the average of sample value difference between all values of the U
14270 plane in the current frame and corresponding values of the previous input frame.
14271 Expressed in range of [0-255].
14274 Display the average of sample value difference between all values of the V
14275 plane in the current frame and corresponding values of the previous input frame.
14276 Expressed in range of [0-255].
14279 Display bit depth of Y plane in current frame.
14280 Expressed in range of [0-16].
14283 Display bit depth of U plane in current frame.
14284 Expressed in range of [0-16].
14287 Display bit depth of V plane in current frame.
14288 Expressed in range of [0-16].
14291 The filter accepts the following options:
14297 @option{stat} specify an additional form of image analysis.
14298 @option{out} output video with the specified type of pixel highlighted.
14300 Both options accept the following values:
14304 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
14305 unlike the neighboring pixels of the same field. Examples of temporal outliers
14306 include the results of video dropouts, head clogs, or tape tracking issues.
14309 Identify @var{vertical line repetition}. Vertical line repetition includes
14310 similar rows of pixels within a frame. In born-digital video vertical line
14311 repetition is common, but this pattern is uncommon in video digitized from an
14312 analog source. When it occurs in video that results from the digitization of an
14313 analog source it can indicate concealment from a dropout compensator.
14316 Identify pixels that fall outside of legal broadcast range.
14320 Set the highlight color for the @option{out} option. The default color is
14324 @subsection Examples
14328 Output data of various video metrics:
14330 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
14334 Output specific data about the minimum and maximum values of the Y plane per frame:
14336 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
14340 Playback video while highlighting pixels that are outside of broadcast range in red.
14342 ffplay example.mov -vf signalstats="out=brng:color=red"
14346 Playback video with signalstats metadata drawn over the frame.
14348 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
14351 The contents of signalstat_drawtext.txt used in the command are:
14354 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
14355 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
14356 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
14357 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
14365 Calculates the MPEG-7 Video Signature. The filter can handle more than one
14366 input. In this case the matching between the inputs can be calculated additionally.
14367 The filter always passes through the first input. The signature of each stream can
14368 be written into a file.
14370 It accepts the following options:
14374 Enable or disable the matching process.
14376 Available values are:
14380 Disable the calculation of a matching (default).
14382 Calculate the matching for the whole video and output whether the whole video
14383 matches or only parts.
14385 Calculate only until a matching is found or the video ends. Should be faster in
14390 Set the number of inputs. The option value must be a non negative integer.
14391 Default value is 1.
14394 Set the path to which the output is written. If there is more than one input,
14395 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
14396 integer), that will be replaced with the input number. If no filename is
14397 specified, no output will be written. This is the default.
14400 Choose the output format.
14402 Available values are:
14406 Use the specified binary representation (default).
14408 Use the specified xml representation.
14412 Set threshold to detect one word as similar. The option value must be an integer
14413 greater than zero. The default value is 9000.
14416 Set threshold to detect all words as similar. The option value must be an integer
14417 greater than zero. The default value is 60000.
14420 Set threshold to detect frames as similar. The option value must be an integer
14421 greater than zero. The default value is 116.
14424 Set the minimum length of a sequence in frames to recognize it as matching
14425 sequence. The option value must be a non negative integer value.
14426 The default value is 0.
14429 Set the minimum relation, that matching frames to all frames must have.
14430 The option value must be a double value between 0 and 1. The default value is 0.5.
14433 @subsection Examples
14437 To calculate the signature of an input video and store it in signature.bin:
14439 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
14443 To detect whether two videos match and store the signatures in XML format in
14444 signature0.xml and signature1.xml:
14446 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 -
14454 Blur the input video without impacting the outlines.
14456 It accepts the following options:
14459 @item luma_radius, lr
14460 Set the luma radius. The option value must be a float number in
14461 the range [0.1,5.0] that specifies the variance of the gaussian filter
14462 used to blur the image (slower if larger). Default value is 1.0.
14464 @item luma_strength, ls
14465 Set the luma strength. The option value must be a float number
14466 in the range [-1.0,1.0] that configures the blurring. A value included
14467 in [0.0,1.0] will blur the image whereas a value included in
14468 [-1.0,0.0] will sharpen the image. Default value is 1.0.
14470 @item luma_threshold, lt
14471 Set the luma threshold used as a coefficient to determine
14472 whether a pixel should be blurred or not. The option value must be an
14473 integer in the range [-30,30]. A value of 0 will filter all the image,
14474 a value included in [0,30] will filter flat areas and a value included
14475 in [-30,0] will filter edges. Default value is 0.
14477 @item chroma_radius, cr
14478 Set the chroma radius. The option value must be a float number in
14479 the range [0.1,5.0] that specifies the variance of the gaussian filter
14480 used to blur the image (slower if larger). Default value is @option{luma_radius}.
14482 @item chroma_strength, cs
14483 Set the chroma strength. The option value must be a float number
14484 in the range [-1.0,1.0] that configures the blurring. A value included
14485 in [0.0,1.0] will blur the image whereas a value included in
14486 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
14488 @item chroma_threshold, ct
14489 Set the chroma threshold used as a coefficient to determine
14490 whether a pixel should be blurred or not. The option value must be an
14491 integer in the range [-30,30]. A value of 0 will filter all the image,
14492 a value included in [0,30] will filter flat areas and a value included
14493 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
14496 If a chroma option is not explicitly set, the corresponding luma value
14501 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
14503 This filter takes in input two input videos, the first input is
14504 considered the "main" source and is passed unchanged to the
14505 output. The second input is used as a "reference" video for computing
14508 Both video inputs must have the same resolution and pixel format for
14509 this filter to work correctly. Also it assumes that both inputs
14510 have the same number of frames, which are compared one by one.
14512 The filter stores the calculated SSIM of each frame.
14514 The description of the accepted parameters follows.
14517 @item stats_file, f
14518 If specified the filter will use the named file to save the SSIM of
14519 each individual frame. When filename equals "-" the data is sent to
14523 The file printed if @var{stats_file} is selected, contains a sequence of
14524 key/value pairs of the form @var{key}:@var{value} for each compared
14527 A description of each shown parameter follows:
14531 sequential number of the input frame, starting from 1
14533 @item Y, U, V, R, G, B
14534 SSIM of the compared frames for the component specified by the suffix.
14537 SSIM of the compared frames for the whole frame.
14540 Same as above but in dB representation.
14543 This filter also supports the @ref{framesync} options.
14547 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
14548 [main][ref] ssim="stats_file=stats.log" [out]
14551 On this example the input file being processed is compared with the
14552 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
14553 is stored in @file{stats.log}.
14555 Another example with both psnr and ssim at same time:
14557 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
14562 Convert between different stereoscopic image formats.
14564 The filters accept the following options:
14568 Set stereoscopic image format of input.
14570 Available values for input image formats are:
14573 side by side parallel (left eye left, right eye right)
14576 side by side crosseye (right eye left, left eye right)
14579 side by side parallel with half width resolution
14580 (left eye left, right eye right)
14583 side by side crosseye with half width resolution
14584 (right eye left, left eye right)
14587 above-below (left eye above, right eye below)
14590 above-below (right eye above, left eye below)
14593 above-below with half height resolution
14594 (left eye above, right eye below)
14597 above-below with half height resolution
14598 (right eye above, left eye below)
14601 alternating frames (left eye first, right eye second)
14604 alternating frames (right eye first, left eye second)
14607 interleaved rows (left eye has top row, right eye starts on next row)
14610 interleaved rows (right eye has top row, left eye starts on next row)
14613 interleaved columns, left eye first
14616 interleaved columns, right eye first
14618 Default value is @samp{sbsl}.
14622 Set stereoscopic image format of output.
14626 side by side parallel (left eye left, right eye right)
14629 side by side crosseye (right eye left, left eye right)
14632 side by side parallel with half width resolution
14633 (left eye left, right eye right)
14636 side by side crosseye with half width resolution
14637 (right eye left, left eye right)
14640 above-below (left eye above, right eye below)
14643 above-below (right eye above, left eye below)
14646 above-below with half height resolution
14647 (left eye above, right eye below)
14650 above-below with half height resolution
14651 (right eye above, left eye below)
14654 alternating frames (left eye first, right eye second)
14657 alternating frames (right eye first, left eye second)
14660 interleaved rows (left eye has top row, right eye starts on next row)
14663 interleaved rows (right eye has top row, left eye starts on next row)
14666 anaglyph red/blue gray
14667 (red filter on left eye, blue filter on right eye)
14670 anaglyph red/green gray
14671 (red filter on left eye, green filter on right eye)
14674 anaglyph red/cyan gray
14675 (red filter on left eye, cyan filter on right eye)
14678 anaglyph red/cyan half colored
14679 (red filter on left eye, cyan filter on right eye)
14682 anaglyph red/cyan color
14683 (red filter on left eye, cyan filter on right eye)
14686 anaglyph red/cyan color optimized with the least squares projection of dubois
14687 (red filter on left eye, cyan filter on right eye)
14690 anaglyph green/magenta gray
14691 (green filter on left eye, magenta filter on right eye)
14694 anaglyph green/magenta half colored
14695 (green filter on left eye, magenta filter on right eye)
14698 anaglyph green/magenta colored
14699 (green filter on left eye, magenta filter on right eye)
14702 anaglyph green/magenta color optimized with the least squares projection of dubois
14703 (green filter on left eye, magenta filter on right eye)
14706 anaglyph yellow/blue gray
14707 (yellow filter on left eye, blue filter on right eye)
14710 anaglyph yellow/blue half colored
14711 (yellow filter on left eye, blue filter on right eye)
14714 anaglyph yellow/blue colored
14715 (yellow filter on left eye, blue filter on right eye)
14718 anaglyph yellow/blue color optimized with the least squares projection of dubois
14719 (yellow filter on left eye, blue filter on right eye)
14722 mono output (left eye only)
14725 mono output (right eye only)
14728 checkerboard, left eye first
14731 checkerboard, right eye first
14734 interleaved columns, left eye first
14737 interleaved columns, right eye first
14743 Default value is @samp{arcd}.
14746 @subsection Examples
14750 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
14756 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
14762 @section streamselect, astreamselect
14763 Select video or audio streams.
14765 The filter accepts the following options:
14769 Set number of inputs. Default is 2.
14772 Set input indexes to remap to outputs.
14775 @subsection Commands
14777 The @code{streamselect} and @code{astreamselect} filter supports the following
14782 Set input indexes to remap to outputs.
14785 @subsection Examples
14789 Select first 5 seconds 1st stream and rest of time 2nd stream:
14791 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
14795 Same as above, but for audio:
14797 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
14802 Apply sobel operator to input video stream.
14804 The filter accepts the following option:
14808 Set which planes will be processed, unprocessed planes will be copied.
14809 By default value 0xf, all planes will be processed.
14812 Set value which will be multiplied with filtered result.
14815 Set value which will be added to filtered result.
14821 Apply a simple postprocessing filter that compresses and decompresses the image
14822 at several (or - in the case of @option{quality} level @code{6} - all) shifts
14823 and average the results.
14825 The filter accepts the following options:
14829 Set quality. This option defines the number of levels for averaging. It accepts
14830 an integer in the range 0-6. If set to @code{0}, the filter will have no
14831 effect. A value of @code{6} means the higher quality. For each increment of
14832 that value the speed drops by a factor of approximately 2. Default value is
14836 Force a constant quantization parameter. If not set, the filter will use the QP
14837 from the video stream (if available).
14840 Set thresholding mode. Available modes are:
14844 Set hard thresholding (default).
14846 Set soft thresholding (better de-ringing effect, but likely blurrier).
14849 @item use_bframe_qp
14850 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
14851 option may cause flicker since the B-Frames have often larger QP. Default is
14852 @code{0} (not enabled).
14858 Draw subtitles on top of input video using the libass library.
14860 To enable compilation of this filter you need to configure FFmpeg with
14861 @code{--enable-libass}. This filter also requires a build with libavcodec and
14862 libavformat to convert the passed subtitles file to ASS (Advanced Substation
14863 Alpha) subtitles format.
14865 The filter accepts the following options:
14869 Set the filename of the subtitle file to read. It must be specified.
14871 @item original_size
14872 Specify the size of the original video, the video for which the ASS file
14873 was composed. For the syntax of this option, check the
14874 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
14875 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
14876 correctly scale the fonts if the aspect ratio has been changed.
14879 Set a directory path containing fonts that can be used by the filter.
14880 These fonts will be used in addition to whatever the font provider uses.
14883 Process alpha channel, by default alpha channel is untouched.
14886 Set subtitles input character encoding. @code{subtitles} filter only. Only
14887 useful if not UTF-8.
14889 @item stream_index, si
14890 Set subtitles stream index. @code{subtitles} filter only.
14893 Override default style or script info parameters of the subtitles. It accepts a
14894 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
14897 If the first key is not specified, it is assumed that the first value
14898 specifies the @option{filename}.
14900 For example, to render the file @file{sub.srt} on top of the input
14901 video, use the command:
14906 which is equivalent to:
14908 subtitles=filename=sub.srt
14911 To render the default subtitles stream from file @file{video.mkv}, use:
14913 subtitles=video.mkv
14916 To render the second subtitles stream from that file, use:
14918 subtitles=video.mkv:si=1
14921 To make the subtitles stream from @file{sub.srt} appear in transparent green
14922 @code{DejaVu Serif}, use:
14924 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HAA00FF00'
14927 @section super2xsai
14929 Scale the input by 2x and smooth using the Super2xSaI (Scale and
14930 Interpolate) pixel art scaling algorithm.
14932 Useful for enlarging pixel art images without reducing sharpness.
14936 Swap two rectangular objects in video.
14938 This filter accepts the following options:
14948 Set 1st rect x coordinate.
14951 Set 1st rect y coordinate.
14954 Set 2nd rect x coordinate.
14957 Set 2nd rect y coordinate.
14959 All expressions are evaluated once for each frame.
14962 The all options are expressions containing the following constants:
14967 The input width and height.
14970 same as @var{w} / @var{h}
14973 input sample aspect ratio
14976 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
14979 The number of the input frame, starting from 0.
14982 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
14985 the position in the file of the input frame, NAN if unknown
14993 Apply telecine process to the video.
14995 This filter accepts the following options:
15004 The default value is @code{top}.
15008 A string of numbers representing the pulldown pattern you wish to apply.
15009 The default value is @code{23}.
15013 Some typical patterns:
15018 24p: 2332 (preferred)
15025 24p: 222222222223 ("Euro pulldown")
15032 Apply threshold effect to video stream.
15034 This filter needs four video streams to perform thresholding.
15035 First stream is stream we are filtering.
15036 Second stream is holding threshold values, third stream is holding min values,
15037 and last, fourth stream is holding max values.
15039 The filter accepts the following option:
15043 Set which planes will be processed, unprocessed planes will be copied.
15044 By default value 0xf, all planes will be processed.
15047 For example if first stream pixel's component value is less then threshold value
15048 of pixel component from 2nd threshold stream, third stream value will picked,
15049 otherwise fourth stream pixel component value will be picked.
15051 Using color source filter one can perform various types of thresholding:
15053 @subsection Examples
15057 Binary threshold, using gray color as threshold:
15059 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
15063 Inverted binary threshold, using gray color as threshold:
15065 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
15069 Truncate binary threshold, using gray color as threshold:
15071 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
15075 Threshold to zero, using gray color as threshold:
15077 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
15081 Inverted threshold to zero, using gray color as threshold:
15083 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
15088 Select the most representative frame in a given sequence of consecutive frames.
15090 The filter accepts the following options:
15094 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
15095 will pick one of them, and then handle the next batch of @var{n} frames until
15096 the end. Default is @code{100}.
15099 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
15100 value will result in a higher memory usage, so a high value is not recommended.
15102 @subsection Examples
15106 Extract one picture each 50 frames:
15112 Complete example of a thumbnail creation with @command{ffmpeg}:
15114 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
15120 Tile several successive frames together.
15122 The filter accepts the following options:
15127 Set the grid size (i.e. the number of lines and columns). For the syntax of
15128 this option, check the
15129 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15132 Set the maximum number of frames to render in the given area. It must be less
15133 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
15134 the area will be used.
15137 Set the outer border margin in pixels.
15140 Set the inner border thickness (i.e. the number of pixels between frames). For
15141 more advanced padding options (such as having different values for the edges),
15142 refer to the pad video filter.
15145 Specify the color of the unused area. For the syntax of this option, check the
15146 "Color" section in the ffmpeg-utils manual. The default value of @var{color}
15150 Set the number of frames to overlap when tiling several successive frames together.
15151 The value must be between @code{0} and @var{nb_frames - 1}.
15154 Set the number of frames to initially be empty before displaying first output frame.
15155 This controls how soon will one get first output frame.
15156 The value must be between @code{0} and @var{nb_frames - 1}.
15159 @subsection Examples
15163 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
15165 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
15167 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
15168 duplicating each output frame to accommodate the originally detected frame
15172 Display @code{5} pictures in an area of @code{3x2} frames,
15173 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
15174 mixed flat and named options:
15176 tile=3x2:nb_frames=5:padding=7:margin=2
15180 @section tinterlace
15182 Perform various types of temporal field interlacing.
15184 Frames are counted starting from 1, so the first input frame is
15187 The filter accepts the following options:
15192 Specify the mode of the interlacing. This option can also be specified
15193 as a value alone. See below for a list of values for this option.
15195 Available values are:
15199 Move odd frames into the upper field, even into the lower field,
15200 generating a double height frame at half frame rate.
15204 Frame 1 Frame 2 Frame 3 Frame 4
15206 11111 22222 33333 44444
15207 11111 22222 33333 44444
15208 11111 22222 33333 44444
15209 11111 22222 33333 44444
15223 Only output odd frames, even frames are dropped, generating a frame with
15224 unchanged height at half frame rate.
15229 Frame 1 Frame 2 Frame 3 Frame 4
15231 11111 22222 33333 44444
15232 11111 22222 33333 44444
15233 11111 22222 33333 44444
15234 11111 22222 33333 44444
15244 Only output even frames, odd frames are dropped, generating a frame with
15245 unchanged height at half frame rate.
15250 Frame 1 Frame 2 Frame 3 Frame 4
15252 11111 22222 33333 44444
15253 11111 22222 33333 44444
15254 11111 22222 33333 44444
15255 11111 22222 33333 44444
15265 Expand each frame to full height, but pad alternate lines with black,
15266 generating a frame with double height at the same input frame rate.
15271 Frame 1 Frame 2 Frame 3 Frame 4
15273 11111 22222 33333 44444
15274 11111 22222 33333 44444
15275 11111 22222 33333 44444
15276 11111 22222 33333 44444
15279 11111 ..... 33333 .....
15280 ..... 22222 ..... 44444
15281 11111 ..... 33333 .....
15282 ..... 22222 ..... 44444
15283 11111 ..... 33333 .....
15284 ..... 22222 ..... 44444
15285 11111 ..... 33333 .....
15286 ..... 22222 ..... 44444
15290 @item interleave_top, 4
15291 Interleave the upper field from odd frames with the lower field from
15292 even frames, generating a frame with unchanged height at half frame rate.
15297 Frame 1 Frame 2 Frame 3 Frame 4
15299 11111<- 22222 33333<- 44444
15300 11111 22222<- 33333 44444<-
15301 11111<- 22222 33333<- 44444
15302 11111 22222<- 33333 44444<-
15312 @item interleave_bottom, 5
15313 Interleave the lower field from odd frames with the upper field from
15314 even frames, generating a frame with unchanged height at half frame rate.
15319 Frame 1 Frame 2 Frame 3 Frame 4
15321 11111 22222<- 33333 44444<-
15322 11111<- 22222 33333<- 44444
15323 11111 22222<- 33333 44444<-
15324 11111<- 22222 33333<- 44444
15334 @item interlacex2, 6
15335 Double frame rate with unchanged height. Frames are inserted each
15336 containing the second temporal field from the previous input frame and
15337 the first temporal field from the next input frame. This mode relies on
15338 the top_field_first flag. Useful for interlaced video displays with no
15339 field synchronisation.
15344 Frame 1 Frame 2 Frame 3 Frame 4
15346 11111 22222 33333 44444
15347 11111 22222 33333 44444
15348 11111 22222 33333 44444
15349 11111 22222 33333 44444
15352 11111 22222 22222 33333 33333 44444 44444
15353 11111 11111 22222 22222 33333 33333 44444
15354 11111 22222 22222 33333 33333 44444 44444
15355 11111 11111 22222 22222 33333 33333 44444
15360 Move odd frames into the upper field, even into the lower field,
15361 generating a double height frame at same frame rate.
15366 Frame 1 Frame 2 Frame 3 Frame 4
15368 11111 22222 33333 44444
15369 11111 22222 33333 44444
15370 11111 22222 33333 44444
15371 11111 22222 33333 44444
15374 11111 33333 33333 55555
15375 22222 22222 44444 44444
15376 11111 33333 33333 55555
15377 22222 22222 44444 44444
15378 11111 33333 33333 55555
15379 22222 22222 44444 44444
15380 11111 33333 33333 55555
15381 22222 22222 44444 44444
15386 Numeric values are deprecated but are accepted for backward
15387 compatibility reasons.
15389 Default mode is @code{merge}.
15392 Specify flags influencing the filter process.
15394 Available value for @var{flags} is:
15397 @item low_pass_filter, vlfp
15398 Enable linear vertical low-pass filtering in the filter.
15399 Vertical low-pass filtering is required when creating an interlaced
15400 destination from a progressive source which contains high-frequency
15401 vertical detail. Filtering will reduce interlace 'twitter' and Moire
15404 @item complex_filter, cvlfp
15405 Enable complex vertical low-pass filtering.
15406 This will slightly less reduce interlace 'twitter' and Moire
15407 patterning but better retain detail and subjective sharpness impression.
15411 Vertical low-pass filtering can only be enabled for @option{mode}
15412 @var{interleave_top} and @var{interleave_bottom}.
15417 Tone map colors from different dynamic ranges.
15419 This filter expects data in single precision floating point, as it needs to
15420 operate on (and can output) out-of-range values. Another filter, such as
15421 @ref{zscale}, is needed to convert the resulting frame to a usable format.
15423 The tonemapping algorithms implemented only work on linear light, so input
15424 data should be linearized beforehand (and possibly correctly tagged).
15427 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
15430 @subsection Options
15431 The filter accepts the following options.
15435 Set the tone map algorithm to use.
15437 Possible values are:
15440 Do not apply any tone map, only desaturate overbright pixels.
15443 Hard-clip any out-of-range values. Use it for perfect color accuracy for
15444 in-range values, while distorting out-of-range values.
15447 Stretch the entire reference gamut to a linear multiple of the display.
15450 Fit a logarithmic transfer between the tone curves.
15453 Preserve overall image brightness with a simple curve, using nonlinear
15454 contrast, which results in flattening details and degrading color accuracy.
15457 Preserve both dark and bright details better than @var{reinhard}, at the cost
15458 of slightly darkening everything. Use it when detail preservation is more
15459 important than color and brightness accuracy.
15462 Smoothly map out-of-range values, while retaining contrast and colors for
15463 in-range material as much as possible. Use it when color accuracy is more
15464 important than detail preservation.
15470 Tune the tone mapping algorithm.
15472 This affects the following algorithms:
15478 Specifies the scale factor to use while stretching.
15482 Specifies the exponent of the function.
15486 Specify an extra linear coefficient to multiply into the signal before clipping.
15490 Specify the local contrast coefficient at the display peak.
15491 Default to 0.5, which means that in-gamut values will be about half as bright
15498 Specify the transition point from linear to mobius transform. Every value
15499 below this point is guaranteed to be mapped 1:1. The higher the value, the
15500 more accurate the result will be, at the cost of losing bright details.
15501 Default to 0.3, which due to the steep initial slope still preserves in-range
15502 colors fairly accurately.
15506 Apply desaturation for highlights that exceed this level of brightness. The
15507 higher the parameter, the more color information will be preserved. This
15508 setting helps prevent unnaturally blown-out colors for super-highlights, by
15509 (smoothly) turning into white instead. This makes images feel more natural,
15510 at the cost of reducing information about out-of-range colors.
15512 The default of 2.0 is somewhat conservative and will mostly just apply to
15513 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
15515 This option works only if the input frame has a supported color tag.
15518 Override signal/nominal/reference peak with this value. Useful when the
15519 embedded peak information in display metadata is not reliable or when tone
15520 mapping from a lower range to a higher range.
15525 Transpose rows with columns in the input video and optionally flip it.
15527 It accepts the following parameters:
15532 Specify the transposition direction.
15534 Can assume the following values:
15536 @item 0, 4, cclock_flip
15537 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
15545 Rotate by 90 degrees clockwise, that is:
15553 Rotate by 90 degrees counterclockwise, that is:
15560 @item 3, 7, clock_flip
15561 Rotate by 90 degrees clockwise and vertically flip, that is:
15569 For values between 4-7, the transposition is only done if the input
15570 video geometry is portrait and not landscape. These values are
15571 deprecated, the @code{passthrough} option should be used instead.
15573 Numerical values are deprecated, and should be dropped in favor of
15574 symbolic constants.
15577 Do not apply the transposition if the input geometry matches the one
15578 specified by the specified value. It accepts the following values:
15581 Always apply transposition.
15583 Preserve portrait geometry (when @var{height} >= @var{width}).
15585 Preserve landscape geometry (when @var{width} >= @var{height}).
15588 Default value is @code{none}.
15591 For example to rotate by 90 degrees clockwise and preserve portrait
15594 transpose=dir=1:passthrough=portrait
15597 The command above can also be specified as:
15599 transpose=1:portrait
15603 Trim the input so that the output contains one continuous subpart of the input.
15605 It accepts the following parameters:
15608 Specify the time of the start of the kept section, i.e. the frame with the
15609 timestamp @var{start} will be the first frame in the output.
15612 Specify the time of the first frame that will be dropped, i.e. the frame
15613 immediately preceding the one with the timestamp @var{end} will be the last
15614 frame in the output.
15617 This is the same as @var{start}, except this option sets the start timestamp
15618 in timebase units instead of seconds.
15621 This is the same as @var{end}, except this option sets the end timestamp
15622 in timebase units instead of seconds.
15625 The maximum duration of the output in seconds.
15628 The number of the first frame that should be passed to the output.
15631 The number of the first frame that should be dropped.
15634 @option{start}, @option{end}, and @option{duration} are expressed as time
15635 duration specifications; see
15636 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
15637 for the accepted syntax.
15639 Note that the first two sets of the start/end options and the @option{duration}
15640 option look at the frame timestamp, while the _frame variants simply count the
15641 frames that pass through the filter. Also note that this filter does not modify
15642 the timestamps. If you wish for the output timestamps to start at zero, insert a
15643 setpts filter after the trim filter.
15645 If multiple start or end options are set, this filter tries to be greedy and
15646 keep all the frames that match at least one of the specified constraints. To keep
15647 only the part that matches all the constraints at once, chain multiple trim
15650 The defaults are such that all the input is kept. So it is possible to set e.g.
15651 just the end values to keep everything before the specified time.
15656 Drop everything except the second minute of input:
15658 ffmpeg -i INPUT -vf trim=60:120
15662 Keep only the first second:
15664 ffmpeg -i INPUT -vf trim=duration=1
15669 @section unpremultiply
15670 Apply alpha unpremultiply effect to input video stream using first plane
15671 of second stream as alpha.
15673 Both streams must have same dimensions and same pixel format.
15675 The filter accepts the following option:
15679 Set which planes will be processed, unprocessed planes will be copied.
15680 By default value 0xf, all planes will be processed.
15682 If the format has 1 or 2 components, then luma is bit 0.
15683 If the format has 3 or 4 components:
15684 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
15685 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
15686 If present, the alpha channel is always the last bit.
15689 Do not require 2nd input for processing, instead use alpha plane from input stream.
15695 Sharpen or blur the input video.
15697 It accepts the following parameters:
15700 @item luma_msize_x, lx
15701 Set the luma matrix horizontal size. It must be an odd integer between
15702 3 and 23. The default value is 5.
15704 @item luma_msize_y, ly
15705 Set the luma matrix vertical size. It must be an odd integer between 3
15706 and 23. The default value is 5.
15708 @item luma_amount, la
15709 Set the luma effect strength. It must be a floating point number, reasonable
15710 values lay between -1.5 and 1.5.
15712 Negative values will blur the input video, while positive values will
15713 sharpen it, a value of zero will disable the effect.
15715 Default value is 1.0.
15717 @item chroma_msize_x, cx
15718 Set the chroma matrix horizontal size. It must be an odd integer
15719 between 3 and 23. The default value is 5.
15721 @item chroma_msize_y, cy
15722 Set the chroma matrix vertical size. It must be an odd integer
15723 between 3 and 23. The default value is 5.
15725 @item chroma_amount, ca
15726 Set the chroma effect strength. It must be a floating point number, reasonable
15727 values lay between -1.5 and 1.5.
15729 Negative values will blur the input video, while positive values will
15730 sharpen it, a value of zero will disable the effect.
15732 Default value is 0.0.
15736 All parameters are optional and default to the equivalent of the
15737 string '5:5:1.0:5:5:0.0'.
15739 @subsection Examples
15743 Apply strong luma sharpen effect:
15745 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
15749 Apply a strong blur of both luma and chroma parameters:
15751 unsharp=7:7:-2:7:7:-2
15757 Apply ultra slow/simple postprocessing filter that compresses and decompresses
15758 the image at several (or - in the case of @option{quality} level @code{8} - all)
15759 shifts and average the results.
15761 The way this differs from the behavior of spp is that uspp actually encodes &
15762 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
15763 DCT similar to MJPEG.
15765 The filter accepts the following options:
15769 Set quality. This option defines the number of levels for averaging. It accepts
15770 an integer in the range 0-8. If set to @code{0}, the filter will have no
15771 effect. A value of @code{8} means the higher quality. For each increment of
15772 that value the speed drops by a factor of approximately 2. Default value is
15776 Force a constant quantization parameter. If not set, the filter will use the QP
15777 from the video stream (if available).
15780 @section vaguedenoiser
15782 Apply a wavelet based denoiser.
15784 It transforms each frame from the video input into the wavelet domain,
15785 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
15786 the obtained coefficients. It does an inverse wavelet transform after.
15787 Due to wavelet properties, it should give a nice smoothed result, and
15788 reduced noise, without blurring picture features.
15790 This filter accepts the following options:
15794 The filtering strength. The higher, the more filtered the video will be.
15795 Hard thresholding can use a higher threshold than soft thresholding
15796 before the video looks overfiltered. Default value is 2.
15799 The filtering method the filter will use.
15801 It accepts the following values:
15804 All values under the threshold will be zeroed.
15807 All values under the threshold will be zeroed. All values above will be
15808 reduced by the threshold.
15811 Scales or nullifies coefficients - intermediary between (more) soft and
15812 (less) hard thresholding.
15815 Default is garrote.
15818 Number of times, the wavelet will decompose the picture. Picture can't
15819 be decomposed beyond a particular point (typically, 8 for a 640x480
15820 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
15823 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
15826 A list of the planes to process. By default all planes are processed.
15829 @section vectorscope
15831 Display 2 color component values in the two dimensional graph (which is called
15834 This filter accepts the following options:
15838 Set vectorscope mode.
15840 It accepts the following values:
15843 Gray values are displayed on graph, higher brightness means more pixels have
15844 same component color value on location in graph. This is the default mode.
15847 Gray values are displayed on graph. Surrounding pixels values which are not
15848 present in video frame are drawn in gradient of 2 color components which are
15849 set by option @code{x} and @code{y}. The 3rd color component is static.
15852 Actual color components values present in video frame are displayed on graph.
15855 Similar as color2 but higher frequency of same values @code{x} and @code{y}
15856 on graph increases value of another color component, which is luminance by
15857 default values of @code{x} and @code{y}.
15860 Actual colors present in video frame are displayed on graph. If two different
15861 colors map to same position on graph then color with higher value of component
15862 not present in graph is picked.
15865 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
15866 component picked from radial gradient.
15870 Set which color component will be represented on X-axis. Default is @code{1}.
15873 Set which color component will be represented on Y-axis. Default is @code{2}.
15876 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
15877 of color component which represents frequency of (X, Y) location in graph.
15882 No envelope, this is default.
15885 Instant envelope, even darkest single pixel will be clearly highlighted.
15888 Hold maximum and minimum values presented in graph over time. This way you
15889 can still spot out of range values without constantly looking at vectorscope.
15892 Peak and instant envelope combined together.
15896 Set what kind of graticule to draw.
15904 Set graticule opacity.
15907 Set graticule flags.
15911 Draw graticule for white point.
15914 Draw graticule for black point.
15917 Draw color points short names.
15921 Set background opacity.
15923 @item lthreshold, l
15924 Set low threshold for color component not represented on X or Y axis.
15925 Values lower than this value will be ignored. Default is 0.
15926 Note this value is multiplied with actual max possible value one pixel component
15927 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
15930 @item hthreshold, h
15931 Set high threshold for color component not represented on X or Y axis.
15932 Values higher than this value will be ignored. Default is 1.
15933 Note this value is multiplied with actual max possible value one pixel component
15934 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
15935 is 0.9 * 255 = 230.
15937 @item colorspace, c
15938 Set what kind of colorspace to use when drawing graticule.
15947 @anchor{vidstabdetect}
15948 @section vidstabdetect
15950 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
15951 @ref{vidstabtransform} for pass 2.
15953 This filter generates a file with relative translation and rotation
15954 transform information about subsequent frames, which is then used by
15955 the @ref{vidstabtransform} filter.
15957 To enable compilation of this filter you need to configure FFmpeg with
15958 @code{--enable-libvidstab}.
15960 This filter accepts the following options:
15964 Set the path to the file used to write the transforms information.
15965 Default value is @file{transforms.trf}.
15968 Set how shaky the video is and how quick the camera is. It accepts an
15969 integer in the range 1-10, a value of 1 means little shakiness, a
15970 value of 10 means strong shakiness. Default value is 5.
15973 Set the accuracy of the detection process. It must be a value in the
15974 range 1-15. A value of 1 means low accuracy, a value of 15 means high
15975 accuracy. Default value is 15.
15978 Set stepsize of the search process. The region around minimum is
15979 scanned with 1 pixel resolution. Default value is 6.
15982 Set minimum contrast. Below this value a local measurement field is
15983 discarded. Must be a floating point value in the range 0-1. Default
15987 Set reference frame number for tripod mode.
15989 If enabled, the motion of the frames is compared to a reference frame
15990 in the filtered stream, identified by the specified number. The idea
15991 is to compensate all movements in a more-or-less static scene and keep
15992 the camera view absolutely still.
15994 If set to 0, it is disabled. The frames are counted starting from 1.
15997 Show fields and transforms in the resulting frames. It accepts an
15998 integer in the range 0-2. Default value is 0, which disables any
16002 @subsection Examples
16006 Use default values:
16012 Analyze strongly shaky movie and put the results in file
16013 @file{mytransforms.trf}:
16015 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
16019 Visualize the result of internal transformations in the resulting
16022 vidstabdetect=show=1
16026 Analyze a video with medium shakiness using @command{ffmpeg}:
16028 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
16032 @anchor{vidstabtransform}
16033 @section vidstabtransform
16035 Video stabilization/deshaking: pass 2 of 2,
16036 see @ref{vidstabdetect} for pass 1.
16038 Read a file with transform information for each frame and
16039 apply/compensate them. Together with the @ref{vidstabdetect}
16040 filter this can be used to deshake videos. See also
16041 @url{http://public.hronopik.de/vid.stab}. It is important to also use
16042 the @ref{unsharp} filter, see below.
16044 To enable compilation of this filter you need to configure FFmpeg with
16045 @code{--enable-libvidstab}.
16047 @subsection Options
16051 Set path to the file used to read the transforms. Default value is
16052 @file{transforms.trf}.
16055 Set the number of frames (value*2 + 1) used for lowpass filtering the
16056 camera movements. Default value is 10.
16058 For example a number of 10 means that 21 frames are used (10 in the
16059 past and 10 in the future) to smoothen the motion in the video. A
16060 larger value leads to a smoother video, but limits the acceleration of
16061 the camera (pan/tilt movements). 0 is a special case where a static
16062 camera is simulated.
16065 Set the camera path optimization algorithm.
16067 Accepted values are:
16070 gaussian kernel low-pass filter on camera motion (default)
16072 averaging on transformations
16076 Set maximal number of pixels to translate frames. Default value is -1,
16080 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
16081 value is -1, meaning no limit.
16084 Specify how to deal with borders that may be visible due to movement
16087 Available values are:
16090 keep image information from previous frame (default)
16092 fill the border black
16096 Invert transforms if set to 1. Default value is 0.
16099 Consider transforms as relative to previous frame if set to 1,
16100 absolute if set to 0. Default value is 0.
16103 Set percentage to zoom. A positive value will result in a zoom-in
16104 effect, a negative value in a zoom-out effect. Default value is 0 (no
16108 Set optimal zooming to avoid borders.
16110 Accepted values are:
16115 optimal static zoom value is determined (only very strong movements
16116 will lead to visible borders) (default)
16118 optimal adaptive zoom value is determined (no borders will be
16119 visible), see @option{zoomspeed}
16122 Note that the value given at zoom is added to the one calculated here.
16125 Set percent to zoom maximally each frame (enabled when
16126 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
16130 Specify type of interpolation.
16132 Available values are:
16137 linear only horizontal
16139 linear in both directions (default)
16141 cubic in both directions (slow)
16145 Enable virtual tripod mode if set to 1, which is equivalent to
16146 @code{relative=0:smoothing=0}. Default value is 0.
16148 Use also @code{tripod} option of @ref{vidstabdetect}.
16151 Increase log verbosity if set to 1. Also the detected global motions
16152 are written to the temporary file @file{global_motions.trf}. Default
16156 @subsection Examples
16160 Use @command{ffmpeg} for a typical stabilization with default values:
16162 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
16165 Note the use of the @ref{unsharp} filter which is always recommended.
16168 Zoom in a bit more and load transform data from a given file:
16170 vidstabtransform=zoom=5:input="mytransforms.trf"
16174 Smoothen the video even more:
16176 vidstabtransform=smoothing=30
16182 Flip the input video vertically.
16184 For example, to vertically flip a video with @command{ffmpeg}:
16186 ffmpeg -i in.avi -vf "vflip" out.avi
16192 Make or reverse a natural vignetting effect.
16194 The filter accepts the following options:
16198 Set lens angle expression as a number of radians.
16200 The value is clipped in the @code{[0,PI/2]} range.
16202 Default value: @code{"PI/5"}
16206 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
16210 Set forward/backward mode.
16212 Available modes are:
16215 The larger the distance from the central point, the darker the image becomes.
16218 The larger the distance from the central point, the brighter the image becomes.
16219 This can be used to reverse a vignette effect, though there is no automatic
16220 detection to extract the lens @option{angle} and other settings (yet). It can
16221 also be used to create a burning effect.
16224 Default value is @samp{forward}.
16227 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
16229 It accepts the following values:
16232 Evaluate expressions only once during the filter initialization.
16235 Evaluate expressions for each incoming frame. This is way slower than the
16236 @samp{init} mode since it requires all the scalers to be re-computed, but it
16237 allows advanced dynamic expressions.
16240 Default value is @samp{init}.
16243 Set dithering to reduce the circular banding effects. Default is @code{1}
16247 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
16248 Setting this value to the SAR of the input will make a rectangular vignetting
16249 following the dimensions of the video.
16251 Default is @code{1/1}.
16254 @subsection Expressions
16256 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
16257 following parameters.
16262 input width and height
16265 the number of input frame, starting from 0
16268 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
16269 @var{TB} units, NAN if undefined
16272 frame rate of the input video, NAN if the input frame rate is unknown
16275 the PTS (Presentation TimeStamp) of the filtered video frame,
16276 expressed in seconds, NAN if undefined
16279 time base of the input video
16283 @subsection Examples
16287 Apply simple strong vignetting effect:
16293 Make a flickering vignetting:
16295 vignette='PI/4+random(1)*PI/50':eval=frame
16300 @section vmafmotion
16302 Obtain the average vmaf motion score of a video.
16303 It is one of the component filters of VMAF.
16305 The obtained average motion score is printed through the logging system.
16307 In the below example the input file @file{ref.mpg} is being processed and score
16311 ffmpeg -i ref.mpg -lavfi vmafmotion -f null -
16315 Stack input videos vertically.
16317 All streams must be of same pixel format and of same width.
16319 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
16320 to create same output.
16322 The filter accept the following option:
16326 Set number of input streams. Default is 2.
16329 If set to 1, force the output to terminate when the shortest input
16330 terminates. Default value is 0.
16335 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
16336 Deinterlacing Filter").
16338 Based on the process described by Martin Weston for BBC R&D, and
16339 implemented based on the de-interlace algorithm written by Jim
16340 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
16341 uses filter coefficients calculated by BBC R&D.
16343 There are two sets of filter coefficients, so called "simple":
16344 and "complex". Which set of filter coefficients is used can
16345 be set by passing an optional parameter:
16349 Set the interlacing filter coefficients. Accepts one of the following values:
16353 Simple filter coefficient set.
16355 More-complex filter coefficient set.
16357 Default value is @samp{complex}.
16360 Specify which frames to deinterlace. Accept one of the following values:
16364 Deinterlace all frames,
16366 Only deinterlace frames marked as interlaced.
16369 Default value is @samp{all}.
16373 Video waveform monitor.
16375 The waveform monitor plots color component intensity. By default luminance
16376 only. Each column of the waveform corresponds to a column of pixels in the
16379 It accepts the following options:
16383 Can be either @code{row}, or @code{column}. Default is @code{column}.
16384 In row mode, the graph on the left side represents color component value 0 and
16385 the right side represents value = 255. In column mode, the top side represents
16386 color component value = 0 and bottom side represents value = 255.
16389 Set intensity. Smaller values are useful to find out how many values of the same
16390 luminance are distributed across input rows/columns.
16391 Default value is @code{0.04}. Allowed range is [0, 1].
16394 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
16395 In mirrored mode, higher values will be represented on the left
16396 side for @code{row} mode and at the top for @code{column} mode. Default is
16397 @code{1} (mirrored).
16401 It accepts the following values:
16404 Presents information identical to that in the @code{parade}, except
16405 that the graphs representing color components are superimposed directly
16408 This display mode makes it easier to spot relative differences or similarities
16409 in overlapping areas of the color components that are supposed to be identical,
16410 such as neutral whites, grays, or blacks.
16413 Display separate graph for the color components side by side in
16414 @code{row} mode or one below the other in @code{column} mode.
16417 Display separate graph for the color components side by side in
16418 @code{column} mode or one below the other in @code{row} mode.
16420 Using this display mode makes it easy to spot color casts in the highlights
16421 and shadows of an image, by comparing the contours of the top and the bottom
16422 graphs of each waveform. Since whites, grays, and blacks are characterized
16423 by exactly equal amounts of red, green, and blue, neutral areas of the picture
16424 should display three waveforms of roughly equal width/height. If not, the
16425 correction is easy to perform by making level adjustments the three waveforms.
16427 Default is @code{stack}.
16429 @item components, c
16430 Set which color components to display. Default is 1, which means only luminance
16431 or red color component if input is in RGB colorspace. If is set for example to
16432 7 it will display all 3 (if) available color components.
16437 No envelope, this is default.
16440 Instant envelope, minimum and maximum values presented in graph will be easily
16441 visible even with small @code{step} value.
16444 Hold minimum and maximum values presented in graph across time. This way you
16445 can still spot out of range values without constantly looking at waveforms.
16448 Peak and instant envelope combined together.
16454 No filtering, this is default.
16457 Luma and chroma combined together.
16460 Similar as above, but shows difference between blue and red chroma.
16463 Displays only chroma.
16466 Displays actual color value on waveform.
16469 Similar as above, but with luma showing frequency of chroma values.
16473 Set which graticule to display.
16477 Do not display graticule.
16480 Display green graticule showing legal broadcast ranges.
16484 Set graticule opacity.
16487 Set graticule flags.
16491 Draw numbers above lines. By default enabled.
16494 Draw dots instead of lines.
16498 Set scale used for displaying graticule.
16505 Default is digital.
16508 Set background opacity.
16511 @section weave, doubleweave
16513 The @code{weave} takes a field-based video input and join
16514 each two sequential fields into single frame, producing a new double
16515 height clip with half the frame rate and half the frame count.
16517 The @code{doubleweave} works same as @code{weave} but without
16518 halving frame rate and frame count.
16520 It accepts the following option:
16524 Set first field. Available values are:
16528 Set the frame as top-field-first.
16531 Set the frame as bottom-field-first.
16535 @subsection Examples
16539 Interlace video using @ref{select} and @ref{separatefields} filter:
16541 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
16546 Apply the xBR high-quality magnification filter which is designed for pixel
16547 art. It follows a set of edge-detection rules, see
16548 @url{http://www.libretro.com/forums/viewtopic.php?f=6&t=134}.
16550 It accepts the following option:
16554 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
16555 @code{3xBR} and @code{4} for @code{4xBR}.
16556 Default is @code{3}.
16562 Deinterlace the input video ("yadif" means "yet another deinterlacing
16565 It accepts the following parameters:
16571 The interlacing mode to adopt. It accepts one of the following values:
16574 @item 0, send_frame
16575 Output one frame for each frame.
16576 @item 1, send_field
16577 Output one frame for each field.
16578 @item 2, send_frame_nospatial
16579 Like @code{send_frame}, but it skips the spatial interlacing check.
16580 @item 3, send_field_nospatial
16581 Like @code{send_field}, but it skips the spatial interlacing check.
16584 The default value is @code{send_frame}.
16587 The picture field parity assumed for the input interlaced video. It accepts one
16588 of the following values:
16592 Assume the top field is first.
16594 Assume the bottom field is first.
16596 Enable automatic detection of field parity.
16599 The default value is @code{auto}.
16600 If the interlacing is unknown or the decoder does not export this information,
16601 top field first will be assumed.
16604 Specify which frames to deinterlace. Accept one of the following
16609 Deinterlace all frames.
16610 @item 1, interlaced
16611 Only deinterlace frames marked as interlaced.
16614 The default value is @code{all}.
16619 Apply Zoom & Pan effect.
16621 This filter accepts the following options:
16625 Set the zoom expression. Default is 1.
16629 Set the x and y expression. Default is 0.
16632 Set the duration expression in number of frames.
16633 This sets for how many number of frames effect will last for
16634 single input image.
16637 Set the output image size, default is 'hd720'.
16640 Set the output frame rate, default is '25'.
16643 Each expression can contain the following constants:
16662 Output frame count.
16666 Last calculated 'x' and 'y' position from 'x' and 'y' expression
16667 for current input frame.
16671 'x' and 'y' of last output frame of previous input frame or 0 when there was
16672 not yet such frame (first input frame).
16675 Last calculated zoom from 'z' expression for current input frame.
16678 Last calculated zoom of last output frame of previous input frame.
16681 Number of output frames for current input frame. Calculated from 'd' expression
16682 for each input frame.
16685 number of output frames created for previous input frame
16688 Rational number: input width / input height
16691 sample aspect ratio
16694 display aspect ratio
16698 @subsection Examples
16702 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
16704 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
16708 Zoom-in up to 1.5 and pan always at center of picture:
16710 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
16714 Same as above but without pausing:
16716 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
16722 Scale (resize) the input video, using the z.lib library:
16723 https://github.com/sekrit-twc/zimg.
16725 The zscale filter forces the output display aspect ratio to be the same
16726 as the input, by changing the output sample aspect ratio.
16728 If the input image format is different from the format requested by
16729 the next filter, the zscale filter will convert the input to the
16732 @subsection Options
16733 The filter accepts the following options.
16738 Set the output video dimension expression. Default value is the input
16741 If the @var{width} or @var{w} value is 0, the input width is used for
16742 the output. If the @var{height} or @var{h} value is 0, the input height
16743 is used for the output.
16745 If one and only one of the values is -n with n >= 1, the zscale filter
16746 will use a value that maintains the aspect ratio of the input image,
16747 calculated from the other specified dimension. After that it will,
16748 however, make sure that the calculated dimension is divisible by n and
16749 adjust the value if necessary.
16751 If both values are -n with n >= 1, the behavior will be identical to
16752 both values being set to 0 as previously detailed.
16754 See below for the list of accepted constants for use in the dimension
16758 Set the video size. For the syntax of this option, check the
16759 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16762 Set the dither type.
16764 Possible values are:
16769 @item error_diffusion
16775 Set the resize filter type.
16777 Possible values are:
16787 Default is bilinear.
16790 Set the color range.
16792 Possible values are:
16799 Default is same as input.
16802 Set the color primaries.
16804 Possible values are:
16814 Default is same as input.
16817 Set the transfer characteristics.
16819 Possible values are:
16833 Default is same as input.
16836 Set the colorspace matrix.
16838 Possible value are:
16849 Default is same as input.
16852 Set the input color range.
16854 Possible values are:
16861 Default is same as input.
16863 @item primariesin, pin
16864 Set the input color primaries.
16866 Possible values are:
16876 Default is same as input.
16878 @item transferin, tin
16879 Set the input transfer characteristics.
16881 Possible values are:
16892 Default is same as input.
16894 @item matrixin, min
16895 Set the input colorspace matrix.
16897 Possible value are:
16909 Set the output chroma location.
16911 Possible values are:
16922 @item chromalin, cin
16923 Set the input chroma location.
16925 Possible values are:
16937 Set the nominal peak luminance.
16940 The values of the @option{w} and @option{h} options are expressions
16941 containing the following constants:
16946 The input width and height
16950 These are the same as @var{in_w} and @var{in_h}.
16954 The output (scaled) width and height
16958 These are the same as @var{out_w} and @var{out_h}
16961 The same as @var{iw} / @var{ih}
16964 input sample aspect ratio
16967 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16971 horizontal and vertical input chroma subsample values. For example for the
16972 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16976 horizontal and vertical output chroma subsample values. For example for the
16977 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16983 @c man end VIDEO FILTERS
16985 @chapter Video Sources
16986 @c man begin VIDEO SOURCES
16988 Below is a description of the currently available video sources.
16992 Buffer video frames, and make them available to the filter chain.
16994 This source is mainly intended for a programmatic use, in particular
16995 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
16997 It accepts the following parameters:
17002 Specify the size (width and height) of the buffered video frames. For the
17003 syntax of this option, check the
17004 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17007 The input video width.
17010 The input video height.
17013 A string representing the pixel format of the buffered video frames.
17014 It may be a number corresponding to a pixel format, or a pixel format
17018 Specify the timebase assumed by the timestamps of the buffered frames.
17021 Specify the frame rate expected for the video stream.
17023 @item pixel_aspect, sar
17024 The sample (pixel) aspect ratio of the input video.
17027 Specify the optional parameters to be used for the scale filter which
17028 is automatically inserted when an input change is detected in the
17029 input size or format.
17031 @item hw_frames_ctx
17032 When using a hardware pixel format, this should be a reference to an
17033 AVHWFramesContext describing input frames.
17038 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
17041 will instruct the source to accept video frames with size 320x240 and
17042 with format "yuv410p", assuming 1/24 as the timestamps timebase and
17043 square pixels (1:1 sample aspect ratio).
17044 Since the pixel format with name "yuv410p" corresponds to the number 6
17045 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
17046 this example corresponds to:
17048 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
17051 Alternatively, the options can be specified as a flat string, but this
17052 syntax is deprecated:
17054 @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}]
17058 Create a pattern generated by an elementary cellular automaton.
17060 The initial state of the cellular automaton can be defined through the
17061 @option{filename} and @option{pattern} options. If such options are
17062 not specified an initial state is created randomly.
17064 At each new frame a new row in the video is filled with the result of
17065 the cellular automaton next generation. The behavior when the whole
17066 frame is filled is defined by the @option{scroll} option.
17068 This source accepts the following options:
17072 Read the initial cellular automaton state, i.e. the starting row, from
17073 the specified file.
17074 In the file, each non-whitespace character is considered an alive
17075 cell, a newline will terminate the row, and further characters in the
17076 file will be ignored.
17079 Read the initial cellular automaton state, i.e. the starting row, from
17080 the specified string.
17082 Each non-whitespace character in the string is considered an alive
17083 cell, a newline will terminate the row, and further characters in the
17084 string will be ignored.
17087 Set the video rate, that is the number of frames generated per second.
17090 @item random_fill_ratio, ratio
17091 Set the random fill ratio for the initial cellular automaton row. It
17092 is a floating point number value ranging from 0 to 1, defaults to
17095 This option is ignored when a file or a pattern is specified.
17097 @item random_seed, seed
17098 Set the seed for filling randomly the initial row, must be an integer
17099 included between 0 and UINT32_MAX. If not specified, or if explicitly
17100 set to -1, the filter will try to use a good random seed on a best
17104 Set the cellular automaton rule, it is a number ranging from 0 to 255.
17105 Default value is 110.
17108 Set the size of the output video. For the syntax of this option, check the
17109 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17111 If @option{filename} or @option{pattern} is specified, the size is set
17112 by default to the width of the specified initial state row, and the
17113 height is set to @var{width} * PHI.
17115 If @option{size} is set, it must contain the width of the specified
17116 pattern string, and the specified pattern will be centered in the
17119 If a filename or a pattern string is not specified, the size value
17120 defaults to "320x518" (used for a randomly generated initial state).
17123 If set to 1, scroll the output upward when all the rows in the output
17124 have been already filled. If set to 0, the new generated row will be
17125 written over the top row just after the bottom row is filled.
17128 @item start_full, full
17129 If set to 1, completely fill the output with generated rows before
17130 outputting the first frame.
17131 This is the default behavior, for disabling set the value to 0.
17134 If set to 1, stitch the left and right row edges together.
17135 This is the default behavior, for disabling set the value to 0.
17138 @subsection Examples
17142 Read the initial state from @file{pattern}, and specify an output of
17145 cellauto=f=pattern:s=200x400
17149 Generate a random initial row with a width of 200 cells, with a fill
17152 cellauto=ratio=2/3:s=200x200
17156 Create a pattern generated by rule 18 starting by a single alive cell
17157 centered on an initial row with width 100:
17159 cellauto=p=@@:s=100x400:full=0:rule=18
17163 Specify a more elaborated initial pattern:
17165 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
17170 @anchor{coreimagesrc}
17171 @section coreimagesrc
17172 Video source generated on GPU using Apple's CoreImage API on OSX.
17174 This video source is a specialized version of the @ref{coreimage} video filter.
17175 Use a core image generator at the beginning of the applied filterchain to
17176 generate the content.
17178 The coreimagesrc video source accepts the following options:
17180 @item list_generators
17181 List all available generators along with all their respective options as well as
17182 possible minimum and maximum values along with the default values.
17184 list_generators=true
17188 Specify the size of the sourced video. For the syntax of this option, check the
17189 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17190 The default value is @code{320x240}.
17193 Specify the frame rate of the sourced video, as the number of frames
17194 generated per second. It has to be a string in the format
17195 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
17196 number or a valid video frame rate abbreviation. The default value is
17200 Set the sample aspect ratio of the sourced video.
17203 Set the duration of the sourced video. See
17204 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17205 for the accepted syntax.
17207 If not specified, or the expressed duration is negative, the video is
17208 supposed to be generated forever.
17211 Additionally, all options of the @ref{coreimage} video filter are accepted.
17212 A complete filterchain can be used for further processing of the
17213 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
17214 and examples for details.
17216 @subsection Examples
17221 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
17222 given as complete and escaped command-line for Apple's standard bash shell:
17224 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
17226 This example is equivalent to the QRCode example of @ref{coreimage} without the
17227 need for a nullsrc video source.
17231 @section mandelbrot
17233 Generate a Mandelbrot set fractal, and progressively zoom towards the
17234 point specified with @var{start_x} and @var{start_y}.
17236 This source accepts the following options:
17241 Set the terminal pts value. Default value is 400.
17244 Set the terminal scale value.
17245 Must be a floating point value. Default value is 0.3.
17248 Set the inner coloring mode, that is the algorithm used to draw the
17249 Mandelbrot fractal internal region.
17251 It shall assume one of the following values:
17256 Show time until convergence.
17258 Set color based on point closest to the origin of the iterations.
17263 Default value is @var{mincol}.
17266 Set the bailout value. Default value is 10.0.
17269 Set the maximum of iterations performed by the rendering
17270 algorithm. Default value is 7189.
17273 Set outer coloring mode.
17274 It shall assume one of following values:
17276 @item iteration_count
17277 Set iteration cound mode.
17278 @item normalized_iteration_count
17279 set normalized iteration count mode.
17281 Default value is @var{normalized_iteration_count}.
17284 Set frame rate, expressed as number of frames per second. Default
17288 Set frame size. For the syntax of this option, check the "Video
17289 size" section in the ffmpeg-utils manual. Default value is "640x480".
17292 Set the initial scale value. Default value is 3.0.
17295 Set the initial x position. Must be a floating point value between
17296 -100 and 100. Default value is -0.743643887037158704752191506114774.
17299 Set the initial y position. Must be a floating point value between
17300 -100 and 100. Default value is -0.131825904205311970493132056385139.
17305 Generate various test patterns, as generated by the MPlayer test filter.
17307 The size of the generated video is fixed, and is 256x256.
17308 This source is useful in particular for testing encoding features.
17310 This source accepts the following options:
17315 Specify the frame rate of the sourced video, as the number of frames
17316 generated per second. It has to be a string in the format
17317 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
17318 number or a valid video frame rate abbreviation. The default value is
17322 Set the duration of the sourced video. See
17323 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17324 for the accepted syntax.
17326 If not specified, or the expressed duration is negative, the video is
17327 supposed to be generated forever.
17331 Set the number or the name of the test to perform. Supported tests are:
17347 Default value is "all", which will cycle through the list of all tests.
17352 mptestsrc=t=dc_luma
17355 will generate a "dc_luma" test pattern.
17357 @section frei0r_src
17359 Provide a frei0r source.
17361 To enable compilation of this filter you need to install the frei0r
17362 header and configure FFmpeg with @code{--enable-frei0r}.
17364 This source accepts the following parameters:
17369 The size of the video to generate. For the syntax of this option, check the
17370 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17373 The framerate of the generated video. It may be a string of the form
17374 @var{num}/@var{den} or a frame rate abbreviation.
17377 The name to the frei0r source to load. For more information regarding frei0r and
17378 how to set the parameters, read the @ref{frei0r} section in the video filters
17381 @item filter_params
17382 A '|'-separated list of parameters to pass to the frei0r source.
17386 For example, to generate a frei0r partik0l source with size 200x200
17387 and frame rate 10 which is overlaid on the overlay filter main input:
17389 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
17394 Generate a life pattern.
17396 This source is based on a generalization of John Conway's life game.
17398 The sourced input represents a life grid, each pixel represents a cell
17399 which can be in one of two possible states, alive or dead. Every cell
17400 interacts with its eight neighbours, which are the cells that are
17401 horizontally, vertically, or diagonally adjacent.
17403 At each interaction the grid evolves according to the adopted rule,
17404 which specifies the number of neighbor alive cells which will make a
17405 cell stay alive or born. The @option{rule} option allows one to specify
17408 This source accepts the following options:
17412 Set the file from which to read the initial grid state. In the file,
17413 each non-whitespace character is considered an alive cell, and newline
17414 is used to delimit the end of each row.
17416 If this option is not specified, the initial grid is generated
17420 Set the video rate, that is the number of frames generated per second.
17423 @item random_fill_ratio, ratio
17424 Set the random fill ratio for the initial random grid. It is a
17425 floating point number value ranging from 0 to 1, defaults to 1/PHI.
17426 It is ignored when a file is specified.
17428 @item random_seed, seed
17429 Set the seed for filling the initial random grid, must be an integer
17430 included between 0 and UINT32_MAX. If not specified, or if explicitly
17431 set to -1, the filter will try to use a good random seed on a best
17437 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
17438 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
17439 @var{NS} specifies the number of alive neighbor cells which make a
17440 live cell stay alive, and @var{NB} the number of alive neighbor cells
17441 which make a dead cell to become alive (i.e. to "born").
17442 "s" and "b" can be used in place of "S" and "B", respectively.
17444 Alternatively a rule can be specified by an 18-bits integer. The 9
17445 high order bits are used to encode the next cell state if it is alive
17446 for each number of neighbor alive cells, the low order bits specify
17447 the rule for "borning" new cells. Higher order bits encode for an
17448 higher number of neighbor cells.
17449 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
17450 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
17452 Default value is "S23/B3", which is the original Conway's game of life
17453 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
17454 cells, and will born a new cell if there are three alive cells around
17458 Set the size of the output video. For the syntax of this option, check the
17459 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17461 If @option{filename} is specified, the size is set by default to the
17462 same size of the input file. If @option{size} is set, it must contain
17463 the size specified in the input file, and the initial grid defined in
17464 that file is centered in the larger resulting area.
17466 If a filename is not specified, the size value defaults to "320x240"
17467 (used for a randomly generated initial grid).
17470 If set to 1, stitch the left and right grid edges together, and the
17471 top and bottom edges also. Defaults to 1.
17474 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
17475 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
17476 value from 0 to 255.
17479 Set the color of living (or new born) cells.
17482 Set the color of dead cells. If @option{mold} is set, this is the first color
17483 used to represent a dead cell.
17486 Set mold color, for definitely dead and moldy cells.
17488 For the syntax of these 3 color options, check the "Color" section in the
17489 ffmpeg-utils manual.
17492 @subsection Examples
17496 Read a grid from @file{pattern}, and center it on a grid of size
17499 life=f=pattern:s=300x300
17503 Generate a random grid of size 200x200, with a fill ratio of 2/3:
17505 life=ratio=2/3:s=200x200
17509 Specify a custom rule for evolving a randomly generated grid:
17515 Full example with slow death effect (mold) using @command{ffplay}:
17517 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
17524 @anchor{haldclutsrc}
17526 @anchor{rgbtestsrc}
17528 @anchor{smptehdbars}
17531 @anchor{yuvtestsrc}
17532 @section allrgb, allyuv, color, haldclutsrc, nullsrc, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
17534 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
17536 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
17538 The @code{color} source provides an uniformly colored input.
17540 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
17541 @ref{haldclut} filter.
17543 The @code{nullsrc} source returns unprocessed video frames. It is
17544 mainly useful to be employed in analysis / debugging tools, or as the
17545 source for filters which ignore the input data.
17547 The @code{rgbtestsrc} source generates an RGB test pattern useful for
17548 detecting RGB vs BGR issues. You should see a red, green and blue
17549 stripe from top to bottom.
17551 The @code{smptebars} source generates a color bars pattern, based on
17552 the SMPTE Engineering Guideline EG 1-1990.
17554 The @code{smptehdbars} source generates a color bars pattern, based on
17555 the SMPTE RP 219-2002.
17557 The @code{testsrc} source generates a test video pattern, showing a
17558 color pattern, a scrolling gradient and a timestamp. This is mainly
17559 intended for testing purposes.
17561 The @code{testsrc2} source is similar to testsrc, but supports more
17562 pixel formats instead of just @code{rgb24}. This allows using it as an
17563 input for other tests without requiring a format conversion.
17565 The @code{yuvtestsrc} source generates an YUV test pattern. You should
17566 see a y, cb and cr stripe from top to bottom.
17568 The sources accept the following parameters:
17573 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
17574 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
17575 pixels to be used as identity matrix for 3D lookup tables. Each component is
17576 coded on a @code{1/(N*N)} scale.
17579 Specify the color of the source, only available in the @code{color}
17580 source. For the syntax of this option, check the "Color" section in the
17581 ffmpeg-utils manual.
17584 Specify the size of the sourced video. For the syntax of this option, check the
17585 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17586 The default value is @code{320x240}.
17588 This option is not available with the @code{allrgb}, @code{allyuv}, and
17589 @code{haldclutsrc} filters.
17592 Specify the frame rate of the sourced video, as the number of frames
17593 generated per second. It has to be a string in the format
17594 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
17595 number or a valid video frame rate abbreviation. The default value is
17599 Set the duration of the sourced video. See
17600 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17601 for the accepted syntax.
17603 If not specified, or the expressed duration is negative, the video is
17604 supposed to be generated forever.
17607 Set the sample aspect ratio of the sourced video.
17610 Specify the alpha (opacity) of the background, only available in the
17611 @code{testsrc2} source. The value must be between 0 (fully transparent) and
17612 255 (fully opaque, the default).
17615 Set the number of decimals to show in the timestamp, only available in the
17616 @code{testsrc} source.
17618 The displayed timestamp value will correspond to the original
17619 timestamp value multiplied by the power of 10 of the specified
17620 value. Default value is 0.
17623 @subsection Examples
17627 Generate a video with a duration of 5.3 seconds, with size
17628 176x144 and a frame rate of 10 frames per second:
17630 testsrc=duration=5.3:size=qcif:rate=10
17634 The following graph description will generate a red source
17635 with an opacity of 0.2, with size "qcif" and a frame rate of 10
17638 color=c=red@@0.2:s=qcif:r=10
17642 If the input content is to be ignored, @code{nullsrc} can be used. The
17643 following command generates noise in the luminance plane by employing
17644 the @code{geq} filter:
17646 nullsrc=s=256x256, geq=random(1)*255:128:128
17650 @subsection Commands
17652 The @code{color} source supports the following commands:
17656 Set the color of the created image. Accepts the same syntax of the
17657 corresponding @option{color} option.
17662 Generate video using an OpenCL program.
17667 OpenCL program source file.
17670 Kernel name in program.
17673 Size of frames to generate. This must be set.
17676 Pixel format to use for the generated frames. This must be set.
17679 Number of frames generated every second. Default value is '25'.
17683 For details of how the program loading works, see the @ref{program_opencl}
17690 Generate a colour ramp by setting pixel values from the position of the pixel
17691 in the output image. (Note that this will work with all pixel formats, but
17692 the generated output will not be the same.)
17694 __kernel void ramp(__write_only image2d_t dst,
17695 unsigned int index)
17697 int2 loc = (int2)(get_global_id(0), get_global_id(1));
17700 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
17702 write_imagef(dst, loc, val);
17707 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
17709 __kernel void sierpinski_carpet(__write_only image2d_t dst,
17710 unsigned int index)
17712 int2 loc = (int2)(get_global_id(0), get_global_id(1));
17714 float4 value = 0.0f;
17715 int x = loc.x + index;
17716 int y = loc.y + index;
17717 while (x > 0 || y > 0) {
17718 if (x % 3 == 1 && y % 3 == 1) {
17726 write_imagef(dst, loc, value);
17732 @c man end VIDEO SOURCES
17734 @chapter Video Sinks
17735 @c man begin VIDEO SINKS
17737 Below is a description of the currently available video sinks.
17739 @section buffersink
17741 Buffer video frames, and make them available to the end of the filter
17744 This sink is mainly intended for programmatic use, in particular
17745 through the interface defined in @file{libavfilter/buffersink.h}
17746 or the options system.
17748 It accepts a pointer to an AVBufferSinkContext structure, which
17749 defines the incoming buffers' formats, to be passed as the opaque
17750 parameter to @code{avfilter_init_filter} for initialization.
17754 Null video sink: do absolutely nothing with the input video. It is
17755 mainly useful as a template and for use in analysis / debugging
17758 @c man end VIDEO SINKS
17760 @chapter Multimedia Filters
17761 @c man begin MULTIMEDIA FILTERS
17763 Below is a description of the currently available multimedia filters.
17767 Convert input audio to a video output, displaying the audio bit scope.
17769 The filter accepts the following options:
17773 Set frame rate, expressed as number of frames per second. Default
17777 Specify the video size for the output. For the syntax of this option, check the
17778 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17779 Default value is @code{1024x256}.
17782 Specify list of colors separated by space or by '|' which will be used to
17783 draw channels. Unrecognized or missing colors will be replaced
17787 @section ahistogram
17789 Convert input audio to a video output, displaying the volume histogram.
17791 The filter accepts the following options:
17795 Specify how histogram is calculated.
17797 It accepts the following values:
17800 Use single histogram for all channels.
17802 Use separate histogram for each channel.
17804 Default is @code{single}.
17807 Set frame rate, expressed as number of frames per second. Default
17811 Specify the video size for the output. For the syntax of this option, check the
17812 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17813 Default value is @code{hd720}.
17818 It accepts the following values:
17829 reverse logarithmic
17831 Default is @code{log}.
17834 Set amplitude scale.
17836 It accepts the following values:
17843 Default is @code{log}.
17846 Set how much frames to accumulate in histogram.
17847 Defauls is 1. Setting this to -1 accumulates all frames.
17850 Set histogram ratio of window height.
17853 Set sonogram sliding.
17855 It accepts the following values:
17858 replace old rows with new ones.
17860 scroll from top to bottom.
17862 Default is @code{replace}.
17865 @section aphasemeter
17867 Convert input audio to a video output, displaying the audio phase.
17869 The filter accepts the following options:
17873 Set the output frame rate. Default value is @code{25}.
17876 Set the video size for the output. For the syntax of this option, check the
17877 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17878 Default value is @code{800x400}.
17883 Specify the red, green, blue contrast. Default values are @code{2},
17884 @code{7} and @code{1}.
17885 Allowed range is @code{[0, 255]}.
17888 Set color which will be used for drawing median phase. If color is
17889 @code{none} which is default, no median phase value will be drawn.
17892 Enable video output. Default is enabled.
17895 The filter also exports the frame metadata @code{lavfi.aphasemeter.phase} which
17896 represents mean phase of current audio frame. Value is in range @code{[-1, 1]}.
17897 The @code{-1} means left and right channels are completely out of phase and
17898 @code{1} means channels are in phase.
17900 @section avectorscope
17902 Convert input audio to a video output, representing the audio vector
17905 The filter is used to measure the difference between channels of stereo
17906 audio stream. A monoaural signal, consisting of identical left and right
17907 signal, results in straight vertical line. Any stereo separation is visible
17908 as a deviation from this line, creating a Lissajous figure.
17909 If the straight (or deviation from it) but horizontal line appears this
17910 indicates that the left and right channels are out of phase.
17912 The filter accepts the following options:
17916 Set the vectorscope mode.
17918 Available values are:
17921 Lissajous rotated by 45 degrees.
17924 Same as above but not rotated.
17927 Shape resembling half of circle.
17930 Default value is @samp{lissajous}.
17933 Set the video size for the output. For the syntax of this option, check the
17934 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17935 Default value is @code{400x400}.
17938 Set the output frame rate. Default value is @code{25}.
17944 Specify the red, green, blue and alpha contrast. Default values are @code{40},
17945 @code{160}, @code{80} and @code{255}.
17946 Allowed range is @code{[0, 255]}.
17952 Specify the red, green, blue and alpha fade. Default values are @code{15},
17953 @code{10}, @code{5} and @code{5}.
17954 Allowed range is @code{[0, 255]}.
17957 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
17958 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
17961 Set the vectorscope drawing mode.
17963 Available values are:
17966 Draw dot for each sample.
17969 Draw line between previous and current sample.
17972 Default value is @samp{dot}.
17975 Specify amplitude scale of audio samples.
17977 Available values are:
17993 Swap left channel axis with right channel axis.
18003 Mirror only x axis.
18006 Mirror only y axis.
18014 @subsection Examples
18018 Complete example using @command{ffplay}:
18020 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
18021 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
18025 @section bench, abench
18027 Benchmark part of a filtergraph.
18029 The filter accepts the following options:
18033 Start or stop a timer.
18035 Available values are:
18038 Get the current time, set it as frame metadata (using the key
18039 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
18042 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
18043 the input frame metadata to get the time difference. Time difference, average,
18044 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
18045 @code{min}) are then printed. The timestamps are expressed in seconds.
18049 @subsection Examples
18053 Benchmark @ref{selectivecolor} filter:
18055 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
18061 Concatenate audio and video streams, joining them together one after the
18064 The filter works on segments of synchronized video and audio streams. All
18065 segments must have the same number of streams of each type, and that will
18066 also be the number of streams at output.
18068 The filter accepts the following options:
18073 Set the number of segments. Default is 2.
18076 Set the number of output video streams, that is also the number of video
18077 streams in each segment. Default is 1.
18080 Set the number of output audio streams, that is also the number of audio
18081 streams in each segment. Default is 0.
18084 Activate unsafe mode: do not fail if segments have a different format.
18088 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
18089 @var{a} audio outputs.
18091 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
18092 segment, in the same order as the outputs, then the inputs for the second
18095 Related streams do not always have exactly the same duration, for various
18096 reasons including codec frame size or sloppy authoring. For that reason,
18097 related synchronized streams (e.g. a video and its audio track) should be
18098 concatenated at once. The concat filter will use the duration of the longest
18099 stream in each segment (except the last one), and if necessary pad shorter
18100 audio streams with silence.
18102 For this filter to work correctly, all segments must start at timestamp 0.
18104 All corresponding streams must have the same parameters in all segments; the
18105 filtering system will automatically select a common pixel format for video
18106 streams, and a common sample format, sample rate and channel layout for
18107 audio streams, but other settings, such as resolution, must be converted
18108 explicitly by the user.
18110 Different frame rates are acceptable but will result in variable frame rate
18111 at output; be sure to configure the output file to handle it.
18113 @subsection Examples
18117 Concatenate an opening, an episode and an ending, all in bilingual version
18118 (video in stream 0, audio in streams 1 and 2):
18120 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
18121 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
18122 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
18123 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
18127 Concatenate two parts, handling audio and video separately, using the
18128 (a)movie sources, and adjusting the resolution:
18130 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
18131 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
18132 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
18134 Note that a desync will happen at the stitch if the audio and video streams
18135 do not have exactly the same duration in the first file.
18139 @section drawgraph, adrawgraph
18141 Draw a graph using input video or audio metadata.
18143 It accepts the following parameters:
18147 Set 1st frame metadata key from which metadata values will be used to draw a graph.
18150 Set 1st foreground color expression.
18153 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
18156 Set 2nd foreground color expression.
18159 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
18162 Set 3rd foreground color expression.
18165 Set 4th frame metadata key from which metadata values will be used to draw a graph.
18168 Set 4th foreground color expression.
18171 Set minimal value of metadata value.
18174 Set maximal value of metadata value.
18177 Set graph background color. Default is white.
18182 Available values for mode is:
18189 Default is @code{line}.
18194 Available values for slide is:
18197 Draw new frame when right border is reached.
18200 Replace old columns with new ones.
18203 Scroll from right to left.
18206 Scroll from left to right.
18209 Draw single picture.
18212 Default is @code{frame}.
18215 Set size of graph video. For the syntax of this option, check the
18216 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18217 The default value is @code{900x256}.
18219 The foreground color expressions can use the following variables:
18222 Minimal value of metadata value.
18225 Maximal value of metadata value.
18228 Current metadata key value.
18231 The color is defined as 0xAABBGGRR.
18234 Example using metadata from @ref{signalstats} filter:
18236 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
18239 Example using metadata from @ref{ebur128} filter:
18241 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
18247 EBU R128 scanner filter. This filter takes an audio stream as input and outputs
18248 it unchanged. By default, it logs a message at a frequency of 10Hz with the
18249 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
18250 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
18252 The filter also has a video output (see the @var{video} option) with a real
18253 time graph to observe the loudness evolution. The graphic contains the logged
18254 message mentioned above, so it is not printed anymore when this option is set,
18255 unless the verbose logging is set. The main graphing area contains the
18256 short-term loudness (3 seconds of analysis), and the gauge on the right is for
18257 the momentary loudness (400 milliseconds).
18259 More information about the Loudness Recommendation EBU R128 on
18260 @url{http://tech.ebu.ch/loudness}.
18262 The filter accepts the following options:
18267 Activate the video output. The audio stream is passed unchanged whether this
18268 option is set or no. The video stream will be the first output stream if
18269 activated. Default is @code{0}.
18272 Set the video size. This option is for video only. For the syntax of this
18274 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18275 Default and minimum resolution is @code{640x480}.
18278 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
18279 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
18280 other integer value between this range is allowed.
18283 Set metadata injection. If set to @code{1}, the audio input will be segmented
18284 into 100ms output frames, each of them containing various loudness information
18285 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
18287 Default is @code{0}.
18290 Force the frame logging level.
18292 Available values are:
18295 information logging level
18297 verbose logging level
18300 By default, the logging level is set to @var{info}. If the @option{video} or
18301 the @option{metadata} options are set, it switches to @var{verbose}.
18306 Available modes can be cumulated (the option is a @code{flag} type). Possible
18310 Disable any peak mode (default).
18312 Enable sample-peak mode.
18314 Simple peak mode looking for the higher sample value. It logs a message
18315 for sample-peak (identified by @code{SPK}).
18317 Enable true-peak mode.
18319 If enabled, the peak lookup is done on an over-sampled version of the input
18320 stream for better peak accuracy. It logs a message for true-peak.
18321 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
18322 This mode requires a build with @code{libswresample}.
18326 Treat mono input files as "dual mono". If a mono file is intended for playback
18327 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
18328 If set to @code{true}, this option will compensate for this effect.
18329 Multi-channel input files are not affected by this option.
18332 Set a specific pan law to be used for the measurement of dual mono files.
18333 This parameter is optional, and has a default value of -3.01dB.
18336 @subsection Examples
18340 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
18342 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
18346 Run an analysis with @command{ffmpeg}:
18348 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
18352 @section interleave, ainterleave
18354 Temporally interleave frames from several inputs.
18356 @code{interleave} works with video inputs, @code{ainterleave} with audio.
18358 These filters read frames from several inputs and send the oldest
18359 queued frame to the output.
18361 Input streams must have well defined, monotonically increasing frame
18364 In order to submit one frame to output, these filters need to enqueue
18365 at least one frame for each input, so they cannot work in case one
18366 input is not yet terminated and will not receive incoming frames.
18368 For example consider the case when one input is a @code{select} filter
18369 which always drops input frames. The @code{interleave} filter will keep
18370 reading from that input, but it will never be able to send new frames
18371 to output until the input sends an end-of-stream signal.
18373 Also, depending on inputs synchronization, the filters will drop
18374 frames in case one input receives more frames than the other ones, and
18375 the queue is already filled.
18377 These filters accept the following options:
18381 Set the number of different inputs, it is 2 by default.
18384 @subsection Examples
18388 Interleave frames belonging to different streams using @command{ffmpeg}:
18390 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
18394 Add flickering blur effect:
18396 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
18400 @section metadata, ametadata
18402 Manipulate frame metadata.
18404 This filter accepts the following options:
18408 Set mode of operation of the filter.
18410 Can be one of the following:
18414 If both @code{value} and @code{key} is set, select frames
18415 which have such metadata. If only @code{key} is set, select
18416 every frame that has such key in metadata.
18419 Add new metadata @code{key} and @code{value}. If key is already available
18423 Modify value of already present key.
18426 If @code{value} is set, delete only keys that have such value.
18427 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
18431 Print key and its value if metadata was found. If @code{key} is not set print all
18432 metadata values available in frame.
18436 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
18439 Set metadata value which will be used. This option is mandatory for
18440 @code{modify} and @code{add} mode.
18443 Which function to use when comparing metadata value and @code{value}.
18445 Can be one of following:
18449 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
18452 Values are interpreted as strings, returns true if metadata value starts with
18453 the @code{value} option string.
18456 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
18459 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
18462 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
18465 Values are interpreted as floats, returns true if expression from option @code{expr}
18470 Set expression which is used when @code{function} is set to @code{expr}.
18471 The expression is evaluated through the eval API and can contain the following
18476 Float representation of @code{value} from metadata key.
18479 Float representation of @code{value} as supplied by user in @code{value} option.
18483 If specified in @code{print} mode, output is written to the named file. Instead of
18484 plain filename any writable url can be specified. Filename ``-'' is a shorthand
18485 for standard output. If @code{file} option is not set, output is written to the log
18486 with AV_LOG_INFO loglevel.
18490 @subsection Examples
18494 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
18497 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
18500 Print silencedetect output to file @file{metadata.txt}.
18502 silencedetect,ametadata=mode=print:file=metadata.txt
18505 Direct all metadata to a pipe with file descriptor 4.
18507 metadata=mode=print:file='pipe\:4'
18511 @section perms, aperms
18513 Set read/write permissions for the output frames.
18515 These filters are mainly aimed at developers to test direct path in the
18516 following filter in the filtergraph.
18518 The filters accept the following options:
18522 Select the permissions mode.
18524 It accepts the following values:
18527 Do nothing. This is the default.
18529 Set all the output frames read-only.
18531 Set all the output frames directly writable.
18533 Make the frame read-only if writable, and writable if read-only.
18535 Set each output frame read-only or writable randomly.
18539 Set the seed for the @var{random} mode, must be an integer included between
18540 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
18541 @code{-1}, the filter will try to use a good random seed on a best effort
18545 Note: in case of auto-inserted filter between the permission filter and the
18546 following one, the permission might not be received as expected in that
18547 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
18548 perms/aperms filter can avoid this problem.
18550 @section realtime, arealtime
18552 Slow down filtering to match real time approximately.
18554 These filters will pause the filtering for a variable amount of time to
18555 match the output rate with the input timestamps.
18556 They are similar to the @option{re} option to @code{ffmpeg}.
18558 They accept the following options:
18562 Time limit for the pauses. Any pause longer than that will be considered
18563 a timestamp discontinuity and reset the timer. Default is 2 seconds.
18567 @section select, aselect
18569 Select frames to pass in output.
18571 This filter accepts the following options:
18576 Set expression, which is evaluated for each input frame.
18578 If the expression is evaluated to zero, the frame is discarded.
18580 If the evaluation result is negative or NaN, the frame is sent to the
18581 first output; otherwise it is sent to the output with index
18582 @code{ceil(val)-1}, assuming that the input index starts from 0.
18584 For example a value of @code{1.2} corresponds to the output with index
18585 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
18588 Set the number of outputs. The output to which to send the selected
18589 frame is based on the result of the evaluation. Default value is 1.
18592 The expression can contain the following constants:
18596 The (sequential) number of the filtered frame, starting from 0.
18599 The (sequential) number of the selected frame, starting from 0.
18601 @item prev_selected_n
18602 The sequential number of the last selected frame. It's NAN if undefined.
18605 The timebase of the input timestamps.
18608 The PTS (Presentation TimeStamp) of the filtered video frame,
18609 expressed in @var{TB} units. It's NAN if undefined.
18612 The PTS of the filtered video frame,
18613 expressed in seconds. It's NAN if undefined.
18616 The PTS of the previously filtered video frame. It's NAN if undefined.
18618 @item prev_selected_pts
18619 The PTS of the last previously filtered video frame. It's NAN if undefined.
18621 @item prev_selected_t
18622 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
18625 The PTS of the first video frame in the video. It's NAN if undefined.
18628 The time of the first video frame in the video. It's NAN if undefined.
18630 @item pict_type @emph{(video only)}
18631 The type of the filtered frame. It can assume one of the following
18643 @item interlace_type @emph{(video only)}
18644 The frame interlace type. It can assume one of the following values:
18647 The frame is progressive (not interlaced).
18649 The frame is top-field-first.
18651 The frame is bottom-field-first.
18654 @item consumed_sample_n @emph{(audio only)}
18655 the number of selected samples before the current frame
18657 @item samples_n @emph{(audio only)}
18658 the number of samples in the current frame
18660 @item sample_rate @emph{(audio only)}
18661 the input sample rate
18664 This is 1 if the filtered frame is a key-frame, 0 otherwise.
18667 the position in the file of the filtered frame, -1 if the information
18668 is not available (e.g. for synthetic video)
18670 @item scene @emph{(video only)}
18671 value between 0 and 1 to indicate a new scene; a low value reflects a low
18672 probability for the current frame to introduce a new scene, while a higher
18673 value means the current frame is more likely to be one (see the example below)
18675 @item concatdec_select
18676 The concat demuxer can select only part of a concat input file by setting an
18677 inpoint and an outpoint, but the output packets may not be entirely contained
18678 in the selected interval. By using this variable, it is possible to skip frames
18679 generated by the concat demuxer which are not exactly contained in the selected
18682 This works by comparing the frame pts against the @var{lavf.concat.start_time}
18683 and the @var{lavf.concat.duration} packet metadata values which are also
18684 present in the decoded frames.
18686 The @var{concatdec_select} variable is -1 if the frame pts is at least
18687 start_time and either the duration metadata is missing or the frame pts is less
18688 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
18691 That basically means that an input frame is selected if its pts is within the
18692 interval set by the concat demuxer.
18696 The default value of the select expression is "1".
18698 @subsection Examples
18702 Select all frames in input:
18707 The example above is the same as:
18719 Select only I-frames:
18721 select='eq(pict_type\,I)'
18725 Select one frame every 100:
18727 select='not(mod(n\,100))'
18731 Select only frames contained in the 10-20 time interval:
18733 select=between(t\,10\,20)
18737 Select only I-frames contained in the 10-20 time interval:
18739 select=between(t\,10\,20)*eq(pict_type\,I)
18743 Select frames with a minimum distance of 10 seconds:
18745 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
18749 Use aselect to select only audio frames with samples number > 100:
18751 aselect='gt(samples_n\,100)'
18755 Create a mosaic of the first scenes:
18757 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
18760 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
18764 Send even and odd frames to separate outputs, and compose them:
18766 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
18770 Select useful frames from an ffconcat file which is using inpoints and
18771 outpoints but where the source files are not intra frame only.
18773 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
18777 @section sendcmd, asendcmd
18779 Send commands to filters in the filtergraph.
18781 These filters read commands to be sent to other filters in the
18784 @code{sendcmd} must be inserted between two video filters,
18785 @code{asendcmd} must be inserted between two audio filters, but apart
18786 from that they act the same way.
18788 The specification of commands can be provided in the filter arguments
18789 with the @var{commands} option, or in a file specified by the
18790 @var{filename} option.
18792 These filters accept the following options:
18795 Set the commands to be read and sent to the other filters.
18797 Set the filename of the commands to be read and sent to the other
18801 @subsection Commands syntax
18803 A commands description consists of a sequence of interval
18804 specifications, comprising a list of commands to be executed when a
18805 particular event related to that interval occurs. The occurring event
18806 is typically the current frame time entering or leaving a given time
18809 An interval is specified by the following syntax:
18811 @var{START}[-@var{END}] @var{COMMANDS};
18814 The time interval is specified by the @var{START} and @var{END} times.
18815 @var{END} is optional and defaults to the maximum time.
18817 The current frame time is considered within the specified interval if
18818 it is included in the interval [@var{START}, @var{END}), that is when
18819 the time is greater or equal to @var{START} and is lesser than
18822 @var{COMMANDS} consists of a sequence of one or more command
18823 specifications, separated by ",", relating to that interval. The
18824 syntax of a command specification is given by:
18826 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
18829 @var{FLAGS} is optional and specifies the type of events relating to
18830 the time interval which enable sending the specified command, and must
18831 be a non-null sequence of identifier flags separated by "+" or "|" and
18832 enclosed between "[" and "]".
18834 The following flags are recognized:
18837 The command is sent when the current frame timestamp enters the
18838 specified interval. In other words, the command is sent when the
18839 previous frame timestamp was not in the given interval, and the
18843 The command is sent when the current frame timestamp leaves the
18844 specified interval. In other words, the command is sent when the
18845 previous frame timestamp was in the given interval, and the
18849 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
18852 @var{TARGET} specifies the target of the command, usually the name of
18853 the filter class or a specific filter instance name.
18855 @var{COMMAND} specifies the name of the command for the target filter.
18857 @var{ARG} is optional and specifies the optional list of argument for
18858 the given @var{COMMAND}.
18860 Between one interval specification and another, whitespaces, or
18861 sequences of characters starting with @code{#} until the end of line,
18862 are ignored and can be used to annotate comments.
18864 A simplified BNF description of the commands specification syntax
18867 @var{COMMAND_FLAG} ::= "enter" | "leave"
18868 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
18869 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
18870 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
18871 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
18872 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
18875 @subsection Examples
18879 Specify audio tempo change at second 4:
18881 asendcmd=c='4.0 atempo tempo 1.5',atempo
18885 Target a specific filter instance:
18887 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
18891 Specify a list of drawtext and hue commands in a file.
18893 # show text in the interval 5-10
18894 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
18895 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
18897 # desaturate the image in the interval 15-20
18898 15.0-20.0 [enter] hue s 0,
18899 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
18901 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
18903 # apply an exponential saturation fade-out effect, starting from time 25
18904 25 [enter] hue s exp(25-t)
18907 A filtergraph allowing to read and process the above command list
18908 stored in a file @file{test.cmd}, can be specified with:
18910 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
18915 @section setpts, asetpts
18917 Change the PTS (presentation timestamp) of the input frames.
18919 @code{setpts} works on video frames, @code{asetpts} on audio frames.
18921 This filter accepts the following options:
18926 The expression which is evaluated for each frame to construct its timestamp.
18930 The expression is evaluated through the eval API and can contain the following
18935 frame rate, only defined for constant frame-rate video
18938 The presentation timestamp in input
18941 The count of the input frame for video or the number of consumed samples,
18942 not including the current frame for audio, starting from 0.
18944 @item NB_CONSUMED_SAMPLES
18945 The number of consumed samples, not including the current frame (only
18948 @item NB_SAMPLES, S
18949 The number of samples in the current frame (only audio)
18951 @item SAMPLE_RATE, SR
18952 The audio sample rate.
18955 The PTS of the first frame.
18958 the time in seconds of the first frame
18961 State whether the current frame is interlaced.
18964 the time in seconds of the current frame
18967 original position in the file of the frame, or undefined if undefined
18968 for the current frame
18971 The previous input PTS.
18974 previous input time in seconds
18977 The previous output PTS.
18980 previous output time in seconds
18983 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
18987 The wallclock (RTC) time at the start of the movie in microseconds.
18990 The timebase of the input timestamps.
18994 @subsection Examples
18998 Start counting PTS from zero
19000 setpts=PTS-STARTPTS
19004 Apply fast motion effect:
19010 Apply slow motion effect:
19016 Set fixed rate of 25 frames per second:
19022 Set fixed rate 25 fps with some jitter:
19024 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
19028 Apply an offset of 10 seconds to the input PTS:
19034 Generate timestamps from a "live source" and rebase onto the current timebase:
19036 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
19040 Generate timestamps by counting samples:
19049 Force color range for the output video frame.
19051 The @code{setrange} filter marks the color range property for the
19052 output frames. It does not change the input frame, but only sets the
19053 corresponding property, which affects how the frame is treated by
19056 The filter accepts the following options:
19061 Available values are:
19065 Keep the same color range property.
19067 @item unspecified, unknown
19068 Set the color range as unspecified.
19070 @item limited, tv, mpeg
19071 Set the color range as limited.
19073 @item full, pc, jpeg
19074 Set the color range as full.
19078 @section settb, asettb
19080 Set the timebase to use for the output frames timestamps.
19081 It is mainly useful for testing timebase configuration.
19083 It accepts the following parameters:
19088 The expression which is evaluated into the output timebase.
19092 The value for @option{tb} is an arithmetic expression representing a
19093 rational. The expression can contain the constants "AVTB" (the default
19094 timebase), "intb" (the input timebase) and "sr" (the sample rate,
19095 audio only). Default value is "intb".
19097 @subsection Examples
19101 Set the timebase to 1/25:
19107 Set the timebase to 1/10:
19113 Set the timebase to 1001/1000:
19119 Set the timebase to 2*intb:
19125 Set the default timebase value:
19132 Convert input audio to a video output representing frequency spectrum
19133 logarithmically using Brown-Puckette constant Q transform algorithm with
19134 direct frequency domain coefficient calculation (but the transform itself
19135 is not really constant Q, instead the Q factor is actually variable/clamped),
19136 with musical tone scale, from E0 to D#10.
19138 The filter accepts the following options:
19142 Specify the video size for the output. It must be even. For the syntax of this option,
19143 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19144 Default value is @code{1920x1080}.
19147 Set the output frame rate. Default value is @code{25}.
19150 Set the bargraph height. It must be even. Default value is @code{-1} which
19151 computes the bargraph height automatically.
19154 Set the axis height. It must be even. Default value is @code{-1} which computes
19155 the axis height automatically.
19158 Set the sonogram height. It must be even. Default value is @code{-1} which
19159 computes the sonogram height automatically.
19162 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
19163 instead. Default value is @code{1}.
19165 @item sono_v, volume
19166 Specify the sonogram volume expression. It can contain variables:
19169 the @var{bar_v} evaluated expression
19170 @item frequency, freq, f
19171 the frequency where it is evaluated
19172 @item timeclamp, tc
19173 the value of @var{timeclamp} option
19177 @item a_weighting(f)
19178 A-weighting of equal loudness
19179 @item b_weighting(f)
19180 B-weighting of equal loudness
19181 @item c_weighting(f)
19182 C-weighting of equal loudness.
19184 Default value is @code{16}.
19186 @item bar_v, volume2
19187 Specify the bargraph volume expression. It can contain variables:
19190 the @var{sono_v} evaluated expression
19191 @item frequency, freq, f
19192 the frequency where it is evaluated
19193 @item timeclamp, tc
19194 the value of @var{timeclamp} option
19198 @item a_weighting(f)
19199 A-weighting of equal loudness
19200 @item b_weighting(f)
19201 B-weighting of equal loudness
19202 @item c_weighting(f)
19203 C-weighting of equal loudness.
19205 Default value is @code{sono_v}.
19207 @item sono_g, gamma
19208 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
19209 higher gamma makes the spectrum having more range. Default value is @code{3}.
19210 Acceptable range is @code{[1, 7]}.
19212 @item bar_g, gamma2
19213 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
19217 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
19218 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
19220 @item timeclamp, tc
19221 Specify the transform timeclamp. At low frequency, there is trade-off between
19222 accuracy in time domain and frequency domain. If timeclamp is lower,
19223 event in time domain is represented more accurately (such as fast bass drum),
19224 otherwise event in frequency domain is represented more accurately
19225 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
19228 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
19229 limits future samples by applying asymmetric windowing in time domain, useful
19230 when low latency is required. Accepted range is @code{[0, 1]}.
19233 Specify the transform base frequency. Default value is @code{20.01523126408007475},
19234 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
19237 Specify the transform end frequency. Default value is @code{20495.59681441799654},
19238 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
19241 This option is deprecated and ignored.
19244 Specify the transform length in time domain. Use this option to control accuracy
19245 trade-off between time domain and frequency domain at every frequency sample.
19246 It can contain variables:
19248 @item frequency, freq, f
19249 the frequency where it is evaluated
19250 @item timeclamp, tc
19251 the value of @var{timeclamp} option.
19253 Default value is @code{384*tc/(384+tc*f)}.
19256 Specify the transform count for every video frame. Default value is @code{6}.
19257 Acceptable range is @code{[1, 30]}.
19260 Specify the transform count for every single pixel. Default value is @code{0},
19261 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
19264 Specify font file for use with freetype to draw the axis. If not specified,
19265 use embedded font. Note that drawing with font file or embedded font is not
19266 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
19270 Specify fontconfig pattern. This has lower priority than @var{fontfile}.
19271 The : in the pattern may be replaced by | to avoid unnecessary escaping.
19274 Specify font color expression. This is arithmetic expression that should return
19275 integer value 0xRRGGBB. It can contain variables:
19277 @item frequency, freq, f
19278 the frequency where it is evaluated
19279 @item timeclamp, tc
19280 the value of @var{timeclamp} option
19285 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
19286 @item r(x), g(x), b(x)
19287 red, green, and blue value of intensity x.
19289 Default value is @code{st(0, (midi(f)-59.5)/12);
19290 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
19291 r(1-ld(1)) + b(ld(1))}.
19294 Specify image file to draw the axis. This option override @var{fontfile} and
19295 @var{fontcolor} option.
19298 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
19299 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
19300 Default value is @code{1}.
19303 Set colorspace. The accepted values are:
19306 Unspecified (default)
19315 BT.470BG or BT.601-6 625
19318 SMPTE-170M or BT.601-6 525
19324 BT.2020 with non-constant luminance
19329 Set spectrogram color scheme. This is list of floating point values with format
19330 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
19331 The default is @code{1|0.5|0|0|0.5|1}.
19335 @subsection Examples
19339 Playing audio while showing the spectrum:
19341 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
19345 Same as above, but with frame rate 30 fps:
19347 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
19351 Playing at 1280x720:
19353 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
19357 Disable sonogram display:
19363 A1 and its harmonics: A1, A2, (near)E3, A3:
19365 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),
19366 asplit[a][out1]; [a] showcqt [out0]'
19370 Same as above, but with more accuracy in frequency domain:
19372 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),
19373 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
19379 bar_v=10:sono_v=bar_v*a_weighting(f)
19383 Custom gamma, now spectrum is linear to the amplitude.
19389 Custom tlength equation:
19391 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)))'
19395 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
19397 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
19401 Custom font using fontconfig:
19403 font='Courier New,Monospace,mono|bold'
19407 Custom frequency range with custom axis using image file:
19409 axisfile=myaxis.png:basefreq=40:endfreq=10000
19415 Convert input audio to video output representing the audio power spectrum.
19416 Audio amplitude is on Y-axis while frequency is on X-axis.
19418 The filter accepts the following options:
19422 Specify size of video. For the syntax of this option, check the
19423 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19424 Default is @code{1024x512}.
19428 This set how each frequency bin will be represented.
19430 It accepts the following values:
19436 Default is @code{bar}.
19439 Set amplitude scale.
19441 It accepts the following values:
19455 Default is @code{log}.
19458 Set frequency scale.
19460 It accepts the following values:
19469 Reverse logarithmic scale.
19471 Default is @code{lin}.
19476 It accepts the following values:
19492 Default is @code{w2048}
19495 Set windowing function.
19497 It accepts the following values:
19519 Default is @code{hanning}.
19522 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
19523 which means optimal overlap for selected window function will be picked.
19526 Set time averaging. Setting this to 0 will display current maximal peaks.
19527 Default is @code{1}, which means time averaging is disabled.
19530 Specify list of colors separated by space or by '|' which will be used to
19531 draw channel frequencies. Unrecognized or missing colors will be replaced
19535 Set channel display mode.
19537 It accepts the following values:
19542 Default is @code{combined}.
19545 Set minimum amplitude used in @code{log} amplitude scaler.
19549 @anchor{showspectrum}
19550 @section showspectrum
19552 Convert input audio to a video output, representing the audio frequency
19555 The filter accepts the following options:
19559 Specify the video size for the output. For the syntax of this option, check the
19560 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19561 Default value is @code{640x512}.
19564 Specify how the spectrum should slide along the window.
19566 It accepts the following values:
19569 the samples start again on the left when they reach the right
19571 the samples scroll from right to left
19573 frames are only produced when the samples reach the right
19575 the samples scroll from left to right
19578 Default value is @code{replace}.
19581 Specify display mode.
19583 It accepts the following values:
19586 all channels are displayed in the same row
19588 all channels are displayed in separate rows
19591 Default value is @samp{combined}.
19594 Specify display color mode.
19596 It accepts the following values:
19599 each channel is displayed in a separate color
19601 each channel is displayed using the same color scheme
19603 each channel is displayed using the rainbow color scheme
19605 each channel is displayed using the moreland color scheme
19607 each channel is displayed using the nebulae color scheme
19609 each channel is displayed using the fire color scheme
19611 each channel is displayed using the fiery color scheme
19613 each channel is displayed using the fruit color scheme
19615 each channel is displayed using the cool color scheme
19618 Default value is @samp{channel}.
19621 Specify scale used for calculating intensity color values.
19623 It accepts the following values:
19628 square root, default
19639 Default value is @samp{sqrt}.
19642 Set saturation modifier for displayed colors. Negative values provide
19643 alternative color scheme. @code{0} is no saturation at all.
19644 Saturation must be in [-10.0, 10.0] range.
19645 Default value is @code{1}.
19648 Set window function.
19650 It accepts the following values:
19674 Default value is @code{hann}.
19677 Set orientation of time vs frequency axis. Can be @code{vertical} or
19678 @code{horizontal}. Default is @code{vertical}.
19681 Set ratio of overlap window. Default value is @code{0}.
19682 When value is @code{1} overlap is set to recommended size for specific
19683 window function currently used.
19686 Set scale gain for calculating intensity color values.
19687 Default value is @code{1}.
19690 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
19693 Set color rotation, must be in [-1.0, 1.0] range.
19694 Default value is @code{0}.
19697 The usage is very similar to the showwaves filter; see the examples in that
19700 @subsection Examples
19704 Large window with logarithmic color scaling:
19706 showspectrum=s=1280x480:scale=log
19710 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
19712 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
19713 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
19717 @section showspectrumpic
19719 Convert input audio to a single video frame, representing the audio frequency
19722 The filter accepts the following options:
19726 Specify the video size for the output. For the syntax of this option, check the
19727 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19728 Default value is @code{4096x2048}.
19731 Specify display mode.
19733 It accepts the following values:
19736 all channels are displayed in the same row
19738 all channels are displayed in separate rows
19740 Default value is @samp{combined}.
19743 Specify display color mode.
19745 It accepts the following values:
19748 each channel is displayed in a separate color
19750 each channel is displayed using the same color scheme
19752 each channel is displayed using the rainbow color scheme
19754 each channel is displayed using the moreland color scheme
19756 each channel is displayed using the nebulae color scheme
19758 each channel is displayed using the fire color scheme
19760 each channel is displayed using the fiery color scheme
19762 each channel is displayed using the fruit color scheme
19764 each channel is displayed using the cool color scheme
19766 Default value is @samp{intensity}.
19769 Specify scale used for calculating intensity color values.
19771 It accepts the following values:
19776 square root, default
19786 Default value is @samp{log}.
19789 Set saturation modifier for displayed colors. Negative values provide
19790 alternative color scheme. @code{0} is no saturation at all.
19791 Saturation must be in [-10.0, 10.0] range.
19792 Default value is @code{1}.
19795 Set window function.
19797 It accepts the following values:
19820 Default value is @code{hann}.
19823 Set orientation of time vs frequency axis. Can be @code{vertical} or
19824 @code{horizontal}. Default is @code{vertical}.
19827 Set scale gain for calculating intensity color values.
19828 Default value is @code{1}.
19831 Draw time and frequency axes and legends. Default is enabled.
19834 Set color rotation, must be in [-1.0, 1.0] range.
19835 Default value is @code{0}.
19838 @subsection Examples
19842 Extract an audio spectrogram of a whole audio track
19843 in a 1024x1024 picture using @command{ffmpeg}:
19845 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
19849 @section showvolume
19851 Convert input audio volume to a video output.
19853 The filter accepts the following options:
19860 Set border width, allowed range is [0, 5]. Default is 1.
19863 Set channel width, allowed range is [80, 8192]. Default is 400.
19866 Set channel height, allowed range is [1, 900]. Default is 20.
19869 Set fade, allowed range is [0.001, 1]. Default is 0.95.
19872 Set volume color expression.
19874 The expression can use the following variables:
19878 Current max volume of channel in dB.
19884 Current channel number, starting from 0.
19888 If set, displays channel names. Default is enabled.
19891 If set, displays volume values. Default is enabled.
19894 Set orientation, can be @code{horizontal} or @code{vertical},
19895 default is @code{horizontal}.
19898 Set step size, allowed range s [0, 5]. Default is 0, which means
19904 Convert input audio to a video output, representing the samples waves.
19906 The filter accepts the following options:
19910 Specify the video size for the output. For the syntax of this option, check the
19911 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19912 Default value is @code{600x240}.
19917 Available values are:
19920 Draw a point for each sample.
19923 Draw a vertical line for each sample.
19926 Draw a point for each sample and a line between them.
19929 Draw a centered vertical line for each sample.
19932 Default value is @code{point}.
19935 Set the number of samples which are printed on the same column. A
19936 larger value will decrease the frame rate. Must be a positive
19937 integer. This option can be set only if the value for @var{rate}
19938 is not explicitly specified.
19941 Set the (approximate) output frame rate. This is done by setting the
19942 option @var{n}. Default value is "25".
19944 @item split_channels
19945 Set if channels should be drawn separately or overlap. Default value is 0.
19948 Set colors separated by '|' which are going to be used for drawing of each channel.
19951 Set amplitude scale.
19953 Available values are:
19971 @subsection Examples
19975 Output the input file audio and the corresponding video representation
19978 amovie=a.mp3,asplit[out0],showwaves[out1]
19982 Create a synthetic signal and show it with showwaves, forcing a
19983 frame rate of 30 frames per second:
19985 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
19989 @section showwavespic
19991 Convert input audio to a single video frame, representing the samples waves.
19993 The filter accepts the following options:
19997 Specify the video size for the output. For the syntax of this option, check the
19998 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19999 Default value is @code{600x240}.
20001 @item split_channels
20002 Set if channels should be drawn separately or overlap. Default value is 0.
20005 Set colors separated by '|' which are going to be used for drawing of each channel.
20008 Set amplitude scale.
20010 Available values are:
20028 @subsection Examples
20032 Extract a channel split representation of the wave form of a whole audio track
20033 in a 1024x800 picture using @command{ffmpeg}:
20035 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
20039 @section sidedata, asidedata
20041 Delete frame side data, or select frames based on it.
20043 This filter accepts the following options:
20047 Set mode of operation of the filter.
20049 Can be one of the following:
20053 Select every frame with side data of @code{type}.
20056 Delete side data of @code{type}. If @code{type} is not set, delete all side
20062 Set side data type used with all modes. Must be set for @code{select} mode. For
20063 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
20064 in @file{libavutil/frame.h}. For example, to choose
20065 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
20069 @section spectrumsynth
20071 Sythesize audio from 2 input video spectrums, first input stream represents
20072 magnitude across time and second represents phase across time.
20073 The filter will transform from frequency domain as displayed in videos back
20074 to time domain as presented in audio output.
20076 This filter is primarily created for reversing processed @ref{showspectrum}
20077 filter outputs, but can synthesize sound from other spectrograms too.
20078 But in such case results are going to be poor if the phase data is not
20079 available, because in such cases phase data need to be recreated, usually
20080 its just recreated from random noise.
20081 For best results use gray only output (@code{channel} color mode in
20082 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
20083 @code{lin} scale for phase video. To produce phase, for 2nd video, use
20084 @code{data} option. Inputs videos should generally use @code{fullframe}
20085 slide mode as that saves resources needed for decoding video.
20087 The filter accepts the following options:
20091 Specify sample rate of output audio, the sample rate of audio from which
20092 spectrum was generated may differ.
20095 Set number of channels represented in input video spectrums.
20098 Set scale which was used when generating magnitude input spectrum.
20099 Can be @code{lin} or @code{log}. Default is @code{log}.
20102 Set slide which was used when generating inputs spectrums.
20103 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
20104 Default is @code{fullframe}.
20107 Set window function used for resynthesis.
20110 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
20111 which means optimal overlap for selected window function will be picked.
20114 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
20115 Default is @code{vertical}.
20118 @subsection Examples
20122 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
20123 then resynthesize videos back to audio with spectrumsynth:
20125 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
20126 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
20127 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
20131 @section split, asplit
20133 Split input into several identical outputs.
20135 @code{asplit} works with audio input, @code{split} with video.
20137 The filter accepts a single parameter which specifies the number of outputs. If
20138 unspecified, it defaults to 2.
20140 @subsection Examples
20144 Create two separate outputs from the same input:
20146 [in] split [out0][out1]
20150 To create 3 or more outputs, you need to specify the number of
20153 [in] asplit=3 [out0][out1][out2]
20157 Create two separate outputs from the same input, one cropped and
20160 [in] split [splitout1][splitout2];
20161 [splitout1] crop=100:100:0:0 [cropout];
20162 [splitout2] pad=200:200:100:100 [padout];
20166 Create 5 copies of the input audio with @command{ffmpeg}:
20168 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
20174 Receive commands sent through a libzmq client, and forward them to
20175 filters in the filtergraph.
20177 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
20178 must be inserted between two video filters, @code{azmq} between two
20181 To enable these filters you need to install the libzmq library and
20182 headers and configure FFmpeg with @code{--enable-libzmq}.
20184 For more information about libzmq see:
20185 @url{http://www.zeromq.org/}
20187 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
20188 receives messages sent through a network interface defined by the
20189 @option{bind_address} option.
20191 The received message must be in the form:
20193 @var{TARGET} @var{COMMAND} [@var{ARG}]
20196 @var{TARGET} specifies the target of the command, usually the name of
20197 the filter class or a specific filter instance name.
20199 @var{COMMAND} specifies the name of the command for the target filter.
20201 @var{ARG} is optional and specifies the optional argument list for the
20202 given @var{COMMAND}.
20204 Upon reception, the message is processed and the corresponding command
20205 is injected into the filtergraph. Depending on the result, the filter
20206 will send a reply to the client, adopting the format:
20208 @var{ERROR_CODE} @var{ERROR_REASON}
20212 @var{MESSAGE} is optional.
20214 @subsection Examples
20216 Look at @file{tools/zmqsend} for an example of a zmq client which can
20217 be used to send commands processed by these filters.
20219 Consider the following filtergraph generated by @command{ffplay}
20221 ffplay -dumpgraph 1 -f lavfi "
20222 color=s=100x100:c=red [l];
20223 color=s=100x100:c=blue [r];
20224 nullsrc=s=200x100, zmq [bg];
20225 [bg][l] overlay [bg+l];
20226 [bg+l][r] overlay=x=100 "
20229 To change the color of the left side of the video, the following
20230 command can be used:
20232 echo Parsed_color_0 c yellow | tools/zmqsend
20235 To change the right side:
20237 echo Parsed_color_1 c pink | tools/zmqsend
20240 @c man end MULTIMEDIA FILTERS
20242 @chapter Multimedia Sources
20243 @c man begin MULTIMEDIA SOURCES
20245 Below is a description of the currently available multimedia sources.
20249 This is the same as @ref{movie} source, except it selects an audio
20255 Read audio and/or video stream(s) from a movie container.
20257 It accepts the following parameters:
20261 The name of the resource to read (not necessarily a file; it can also be a
20262 device or a stream accessed through some protocol).
20264 @item format_name, f
20265 Specifies the format assumed for the movie to read, and can be either
20266 the name of a container or an input device. If not specified, the
20267 format is guessed from @var{movie_name} or by probing.
20269 @item seek_point, sp
20270 Specifies the seek point in seconds. The frames will be output
20271 starting from this seek point. The parameter is evaluated with
20272 @code{av_strtod}, so the numerical value may be suffixed by an IS
20273 postfix. The default value is "0".
20276 Specifies the streams to read. Several streams can be specified,
20277 separated by "+". The source will then have as many outputs, in the
20278 same order. The syntax is explained in the ``Stream specifiers''
20279 section in the ffmpeg manual. Two special names, "dv" and "da" specify
20280 respectively the default (best suited) video and audio stream. Default
20281 is "dv", or "da" if the filter is called as "amovie".
20283 @item stream_index, si
20284 Specifies the index of the video stream to read. If the value is -1,
20285 the most suitable video stream will be automatically selected. The default
20286 value is "-1". Deprecated. If the filter is called "amovie", it will select
20287 audio instead of video.
20290 Specifies how many times to read the stream in sequence.
20291 If the value is 0, the stream will be looped infinitely.
20292 Default value is "1".
20294 Note that when the movie is looped the source timestamps are not
20295 changed, so it will generate non monotonically increasing timestamps.
20297 @item discontinuity
20298 Specifies the time difference between frames above which the point is
20299 considered a timestamp discontinuity which is removed by adjusting the later
20303 It allows overlaying a second video on top of the main input of
20304 a filtergraph, as shown in this graph:
20306 input -----------> deltapts0 --> overlay --> output
20309 movie --> scale--> deltapts1 -------+
20311 @subsection Examples
20315 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
20316 on top of the input labelled "in":
20318 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
20319 [in] setpts=PTS-STARTPTS [main];
20320 [main][over] overlay=16:16 [out]
20324 Read from a video4linux2 device, and overlay it on top of the input
20327 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
20328 [in] setpts=PTS-STARTPTS [main];
20329 [main][over] overlay=16:16 [out]
20333 Read the first video stream and the audio stream with id 0x81 from
20334 dvd.vob; the video is connected to the pad named "video" and the audio is
20335 connected to the pad named "audio":
20337 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
20341 @subsection Commands
20343 Both movie and amovie support the following commands:
20346 Perform seek using "av_seek_frame".
20347 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
20350 @var{stream_index}: If stream_index is -1, a default
20351 stream is selected, and @var{timestamp} is automatically converted
20352 from AV_TIME_BASE units to the stream specific time_base.
20354 @var{timestamp}: Timestamp in AVStream.time_base units
20355 or, if no stream is specified, in AV_TIME_BASE units.
20357 @var{flags}: Flags which select direction and seeking mode.
20361 Get movie duration in AV_TIME_BASE units.
20365 @c man end MULTIMEDIA SOURCES