1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order. Available values are:
541 Default is @var{4th}.
544 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
551 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
552 each band will be in separate stream:
554 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
558 Same as above, but with higher filter order:
560 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
564 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
566 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
572 Reduce audio bit resolution.
574 This filter is bit crusher with enhanced functionality. A bit crusher
575 is used to audibly reduce number of bits an audio signal is sampled
576 with. This doesn't change the bit depth at all, it just produces the
577 effect. Material reduced in bit depth sounds more harsh and "digital".
578 This filter is able to even round to continuous values instead of discrete
580 Additionally it has a D/C offset which results in different crushing of
581 the lower and the upper half of the signal.
582 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
584 Another feature of this filter is the logarithmic mode.
585 This setting switches from linear distances between bits to logarithmic ones.
586 The result is a much more "natural" sounding crusher which doesn't gate low
587 signals for example. The human ear has a logarithmic perception,
588 so this kind of crushing is much more pleasant.
589 Logarithmic crushing is also able to get anti-aliased.
591 The filter accepts the following options:
607 Can be linear: @code{lin} or logarithmic: @code{log}.
616 Set sample reduction.
619 Enable LFO. By default disabled.
630 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
634 Remove impulsive noise from input audio.
636 Samples detected as impulsive noise are replaced by interpolated samples using
637 autoregressive modelling.
641 Set window size, in milliseconds. Allowed range is from @code{10} to
642 @code{100}. Default value is @code{55} milliseconds.
643 This sets size of window which will be processed at once.
646 Set window overlap, in percentage of window size. Allowed range is from
647 @code{50} to @code{95}. Default value is @code{75} percent.
648 Setting this to a very high value increases impulsive noise removal but makes
649 whole process much slower.
652 Set autoregression order, in percentage of window size. Allowed range is from
653 @code{0} to @code{25}. Default value is @code{2} percent. This option also
654 controls quality of interpolated samples using neighbour good samples.
657 Set threshold value. Allowed range is from @code{1} to @code{100}.
658 Default value is @code{2}.
659 This controls the strength of impulsive noise which is going to be removed.
660 The lower value, the more samples will be detected as impulsive noise.
663 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
664 @code{10}. Default value is @code{2}.
665 If any two samples detected as noise are spaced less than this value then any
666 sample between those two samples will be also detected as noise.
671 It accepts the following values:
674 Select overlap-add method. Even not interpolated samples are slightly
675 changed with this method.
678 Select overlap-save method. Not interpolated samples remain unchanged.
681 Default value is @code{a}.
685 Remove clipped samples from input audio.
687 Samples detected as clipped are replaced by interpolated samples using
688 autoregressive modelling.
692 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
693 Default value is @code{55} milliseconds.
694 This sets size of window which will be processed at once.
697 Set window overlap, in percentage of window size. Allowed range is from @code{50}
698 to @code{95}. Default value is @code{75} percent.
701 Set autoregression order, in percentage of window size. Allowed range is from
702 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
703 quality of interpolated samples using neighbour good samples.
706 Set threshold value. Allowed range is from @code{1} to @code{100}.
707 Default value is @code{10}. Higher values make clip detection less aggressive.
710 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
711 Default value is @code{1000}. Higher values make clip detection less aggressive.
716 It accepts the following values:
719 Select overlap-add method. Even not interpolated samples are slightly changed
723 Select overlap-save method. Not interpolated samples remain unchanged.
726 Default value is @code{a}.
731 Delay one or more audio channels.
733 Samples in delayed channel are filled with silence.
735 The filter accepts the following option:
739 Set list of delays in milliseconds for each channel separated by '|'.
740 Unused delays will be silently ignored. If number of given delays is
741 smaller than number of channels all remaining channels will not be delayed.
742 If you want to delay exact number of samples, append 'S' to number.
743 If you want instead to delay in seconds, append 's' to number.
746 Use last set delay for all remaining channels. By default is disabled.
747 This option if enabled changes how option @code{delays} is interpreted.
754 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
755 the second channel (and any other channels that may be present) unchanged.
761 Delay second channel by 500 samples, the third channel by 700 samples and leave
762 the first channel (and any other channels that may be present) unchanged.
768 Delay all channels by same number of samples:
770 adelay=delays=64S:all=1
775 Remedy denormals in audio by adding extremely low-level noise.
777 This filter shall be placed before any filter that can produce denormals.
779 A description of the accepted parameters follows.
783 Set level of added noise in dB. Default is @code{-351}.
784 Allowed range is from -451 to -90.
787 Set type of added noise.
800 Default is @code{dc}.
805 This filter supports the all above options as @ref{commands}.
807 @section aderivative, aintegral
809 Compute derivative/integral of audio stream.
811 Applying both filters one after another produces original audio.
815 Apply echoing to the input audio.
817 Echoes are reflected sound and can occur naturally amongst mountains
818 (and sometimes large buildings) when talking or shouting; digital echo
819 effects emulate this behaviour and are often used to help fill out the
820 sound of a single instrument or vocal. The time difference between the
821 original signal and the reflection is the @code{delay}, and the
822 loudness of the reflected signal is the @code{decay}.
823 Multiple echoes can have different delays and decays.
825 A description of the accepted parameters follows.
829 Set input gain of reflected signal. Default is @code{0.6}.
832 Set output gain of reflected signal. Default is @code{0.3}.
835 Set list of time intervals in milliseconds between original signal and reflections
836 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
837 Default is @code{1000}.
840 Set list of loudness of reflected signals separated by '|'.
841 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
842 Default is @code{0.5}.
849 Make it sound as if there are twice as many instruments as are actually playing:
851 aecho=0.8:0.88:60:0.4
855 If delay is very short, then it sounds like a (metallic) robot playing music:
861 A longer delay will sound like an open air concert in the mountains:
863 aecho=0.8:0.9:1000:0.3
867 Same as above but with one more mountain:
869 aecho=0.8:0.9:1000|1800:0.3|0.25
874 Audio emphasis filter creates or restores material directly taken from LPs or
875 emphased CDs with different filter curves. E.g. to store music on vinyl the
876 signal has to be altered by a filter first to even out the disadvantages of
877 this recording medium.
878 Once the material is played back the inverse filter has to be applied to
879 restore the distortion of the frequency response.
881 The filter accepts the following options:
891 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
892 use @code{production} mode. Default is @code{reproduction} mode.
895 Set filter type. Selects medium. Can be one of the following:
907 select Compact Disc (CD).
913 select 50µs (FM-KF).
915 select 75µs (FM-KF).
921 This filter supports the all above options as @ref{commands}.
925 Modify an audio signal according to the specified expressions.
927 This filter accepts one or more expressions (one for each channel),
928 which are evaluated and used to modify a corresponding audio signal.
930 It accepts the following parameters:
934 Set the '|'-separated expressions list for each separate channel. If
935 the number of input channels is greater than the number of
936 expressions, the last specified expression is used for the remaining
939 @item channel_layout, c
940 Set output channel layout. If not specified, the channel layout is
941 specified by the number of expressions. If set to @samp{same}, it will
942 use by default the same input channel layout.
945 Each expression in @var{exprs} can contain the following constants and functions:
949 channel number of the current expression
952 number of the evaluated sample, starting from 0
958 time of the evaluated sample expressed in seconds
961 @item nb_out_channels
962 input and output number of channels
965 the value of input channel with number @var{CH}
968 Note: this filter is slow. For faster processing you should use a
977 aeval=val(ch)/2:c=same
981 Invert phase of the second channel:
990 Apply fade-in/out effect to input audio.
992 A description of the accepted parameters follows.
996 Specify the effect type, can be either @code{in} for fade-in, or
997 @code{out} for a fade-out effect. Default is @code{in}.
999 @item start_sample, ss
1000 Specify the number of the start sample for starting to apply the fade
1001 effect. Default is 0.
1003 @item nb_samples, ns
1004 Specify the number of samples for which the fade effect has to last. At
1005 the end of the fade-in effect the output audio will have the same
1006 volume as the input audio, at the end of the fade-out transition
1007 the output audio will be silence. Default is 44100.
1009 @item start_time, st
1010 Specify the start time of the fade effect. Default is 0.
1011 The value must be specified as a time duration; see
1012 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1013 for the accepted syntax.
1014 If set this option is used instead of @var{start_sample}.
1017 Specify the duration of the fade effect. See
1018 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1019 for the accepted syntax.
1020 At the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence.
1023 By default the duration is determined by @var{nb_samples}.
1024 If set this option is used instead of @var{nb_samples}.
1027 Set curve for fade transition.
1029 It accepts the following values:
1032 select triangular, linear slope (default)
1034 select quarter of sine wave
1036 select half of sine wave
1038 select exponential sine wave
1042 select inverted parabola
1056 select inverted quarter of sine wave
1058 select inverted half of sine wave
1060 select double-exponential seat
1062 select double-exponential sigmoid
1064 select logistic sigmoid
1066 select sine cardinal function
1068 select inverted sine cardinal function
1074 @subsection Examples
1078 Fade in first 15 seconds of audio:
1080 afade=t=in:ss=0:d=15
1084 Fade out last 25 seconds of a 900 seconds audio:
1086 afade=t=out:st=875:d=25
1091 Denoise audio samples with FFT.
1093 A description of the accepted parameters follows.
1097 Set the noise reduction in dB, allowed range is 0.01 to 97.
1098 Default value is 12 dB.
1101 Set the noise floor in dB, allowed range is -80 to -20.
1102 Default value is -50 dB.
1107 It accepts the following values:
1116 Select shellac noise.
1119 Select custom noise, defined in @code{bn} option.
1121 Default value is white noise.
1125 Set custom band noise for every one of 15 bands.
1126 Bands are separated by ' ' or '|'.
1129 Set the residual floor in dB, allowed range is -80 to -20.
1130 Default value is -38 dB.
1133 Enable noise tracking. By default is disabled.
1134 With this enabled, noise floor is automatically adjusted.
1137 Enable residual tracking. By default is disabled.
1140 Set the output mode.
1142 It accepts the following values:
1145 Pass input unchanged.
1148 Pass noise filtered out.
1153 Default value is @var{o}.
1157 @subsection Commands
1159 This filter supports the following commands:
1161 @item sample_noise, sn
1162 Start or stop measuring noise profile.
1163 Syntax for the command is : "start" or "stop" string.
1164 After measuring noise profile is stopped it will be
1165 automatically applied in filtering.
1167 @item noise_reduction, nr
1168 Change noise reduction. Argument is single float number.
1169 Syntax for the command is : "@var{noise_reduction}"
1171 @item noise_floor, nf
1172 Change noise floor. Argument is single float number.
1173 Syntax for the command is : "@var{noise_floor}"
1175 @item output_mode, om
1176 Change output mode operation.
1177 Syntax for the command is : "i", "o" or "n" string.
1181 Apply arbitrary expressions to samples in frequency domain.
1185 Set frequency domain real expression for each separate channel separated
1186 by '|'. Default is "re".
1187 If the number of input channels is greater than the number of
1188 expressions, the last specified expression is used for the remaining
1192 Set frequency domain imaginary expression for each separate channel
1193 separated by '|'. Default is "im".
1195 Each expression in @var{real} and @var{imag} can contain the following
1196 constants and functions:
1203 current frequency bin number
1206 number of available bins
1209 channel number of the current expression
1218 current real part of frequency bin of current channel
1221 current imaginary part of frequency bin of current channel
1224 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1227 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1231 Set window size. Allowed range is from 16 to 131072.
1232 Default is @code{4096}
1235 Set window function. Default is @code{hann}.
1238 Set window overlap. If set to 1, the recommended overlap for selected
1239 window function will be picked. Default is @code{0.75}.
1242 @subsection Examples
1246 Leave almost only low frequencies in audio:
1248 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1252 Apply robotize effect:
1254 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1258 Apply whisper effect:
1260 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1267 Apply an arbitrary Finite Impulse Response filter.
1269 This filter is designed for applying long FIR filters,
1270 up to 60 seconds long.
1272 It can be used as component for digital crossover filters,
1273 room equalization, cross talk cancellation, wavefield synthesis,
1274 auralization, ambiophonics, ambisonics and spatialization.
1276 This filter uses the streams higher than first one as FIR coefficients.
1277 If the non-first stream holds a single channel, it will be used
1278 for all input channels in the first stream, otherwise
1279 the number of channels in the non-first stream must be same as
1280 the number of channels in the first stream.
1282 It accepts the following parameters:
1286 Set dry gain. This sets input gain.
1289 Set wet gain. This sets final output gain.
1292 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1295 Enable applying gain measured from power of IR.
1297 Set which approach to use for auto gain measurement.
1301 Do not apply any gain.
1304 select peak gain, very conservative approach. This is default value.
1307 select DC gain, limited application.
1310 select gain to noise approach, this is most popular one.
1314 Set gain to be applied to IR coefficients before filtering.
1315 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1318 Set format of IR stream. Can be @code{mono} or @code{input}.
1319 Default is @code{input}.
1322 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1323 Allowed range is 0.1 to 60 seconds.
1326 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1327 By default it is disabled.
1330 Set for which IR channel to display frequency response. By default is first channel
1331 displayed. This option is used only when @var{response} is enabled.
1334 Set video stream size. This option is used only when @var{response} is enabled.
1337 Set video stream frame rate. This option is used only when @var{response} is enabled.
1340 Set minimal partition size used for convolution. Default is @var{8192}.
1341 Allowed range is from @var{1} to @var{32768}.
1342 Lower values decreases latency at cost of higher CPU usage.
1345 Set maximal partition size used for convolution. Default is @var{8192}.
1346 Allowed range is from @var{8} to @var{32768}.
1347 Lower values may increase CPU usage.
1350 Set number of input impulse responses streams which will be switchable at runtime.
1351 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1354 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1355 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1356 This option can be changed at runtime via @ref{commands}.
1359 @subsection Examples
1363 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1365 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1372 Set output format constraints for the input audio. The framework will
1373 negotiate the most appropriate format to minimize conversions.
1375 It accepts the following parameters:
1378 @item sample_fmts, f
1379 A '|'-separated list of requested sample formats.
1381 @item sample_rates, r
1382 A '|'-separated list of requested sample rates.
1384 @item channel_layouts, cl
1385 A '|'-separated list of requested channel layouts.
1387 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1388 for the required syntax.
1391 If a parameter is omitted, all values are allowed.
1393 Force the output to either unsigned 8-bit or signed 16-bit stereo
1395 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1399 Apply frequency shift to input audio samples.
1401 The filter accepts the following options:
1405 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1406 Default value is 0.0.
1409 @subsection Commands
1411 This filter supports the above option as @ref{commands}.
1415 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1416 processing reduces disturbing noise between useful signals.
1418 Gating is done by detecting the volume below a chosen level @var{threshold}
1419 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1420 floor is set via @var{range}. Because an exact manipulation of the signal
1421 would cause distortion of the waveform the reduction can be levelled over
1422 time. This is done by setting @var{attack} and @var{release}.
1424 @var{attack} determines how long the signal has to fall below the threshold
1425 before any reduction will occur and @var{release} sets the time the signal
1426 has to rise above the threshold to reduce the reduction again.
1427 Shorter signals than the chosen attack time will be left untouched.
1431 Set input level before filtering.
1432 Default is 1. Allowed range is from 0.015625 to 64.
1435 Set the mode of operation. Can be @code{upward} or @code{downward}.
1436 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1437 will be amplified, expanding dynamic range in upward direction.
1438 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1441 Set the level of gain reduction when the signal is below the threshold.
1442 Default is 0.06125. Allowed range is from 0 to 1.
1443 Setting this to 0 disables reduction and then filter behaves like expander.
1446 If a signal rises above this level the gain reduction is released.
1447 Default is 0.125. Allowed range is from 0 to 1.
1450 Set a ratio by which the signal is reduced.
1451 Default is 2. Allowed range is from 1 to 9000.
1454 Amount of milliseconds the signal has to rise above the threshold before gain
1456 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1459 Amount of milliseconds the signal has to fall below the threshold before the
1460 reduction is increased again. Default is 250 milliseconds.
1461 Allowed range is from 0.01 to 9000.
1464 Set amount of amplification of signal after processing.
1465 Default is 1. Allowed range is from 1 to 64.
1468 Curve the sharp knee around the threshold to enter gain reduction more softly.
1469 Default is 2.828427125. Allowed range is from 1 to 8.
1472 Choose if exact signal should be taken for detection or an RMS like one.
1473 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1476 Choose if the average level between all channels or the louder channel affects
1478 Default is @code{average}. Can be @code{average} or @code{maximum}.
1481 @subsection Commands
1483 This filter supports the all above options as @ref{commands}.
1487 Apply an arbitrary Infinite Impulse Response filter.
1489 It accepts the following parameters:
1493 Set B/numerator/zeros/reflection coefficients.
1496 Set A/denominator/poles/ladder coefficients.
1508 Set coefficients format.
1512 lattice-ladder function
1514 analog transfer function
1516 digital transfer function
1518 Z-plane zeros/poles, cartesian (default)
1520 Z-plane zeros/poles, polar radians
1522 Z-plane zeros/poles, polar degrees
1528 Set type of processing.
1540 Set filtering precision.
1544 double-precision floating-point (default)
1546 single-precision floating-point
1554 Normalize filter coefficients, by default is enabled.
1555 Enabling it will normalize magnitude response at DC to 0dB.
1558 How much to use filtered signal in output. Default is 1.
1559 Range is between 0 and 1.
1562 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1563 By default it is disabled.
1566 Set for which IR channel to display frequency response. By default is first channel
1567 displayed. This option is used only when @var{response} is enabled.
1570 Set video stream size. This option is used only when @var{response} is enabled.
1573 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1576 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1577 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1580 Different coefficients and gains can be provided for every channel, in such case
1581 use '|' to separate coefficients or gains. Last provided coefficients will be
1582 used for all remaining channels.
1584 @subsection Examples
1588 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1590 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
1594 Same as above but in @code{zp} format:
1596 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
1600 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1602 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1608 The limiter prevents an input signal from rising over a desired threshold.
1609 This limiter uses lookahead technology to prevent your signal from distorting.
1610 It means that there is a small delay after the signal is processed. Keep in mind
1611 that the delay it produces is the attack time you set.
1613 The filter accepts the following options:
1617 Set input gain. Default is 1.
1620 Set output gain. Default is 1.
1623 Don't let signals above this level pass the limiter. Default is 1.
1626 The limiter will reach its attenuation level in this amount of time in
1627 milliseconds. Default is 5 milliseconds.
1630 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1631 Default is 50 milliseconds.
1634 When gain reduction is always needed ASC takes care of releasing to an
1635 average reduction level rather than reaching a reduction of 0 in the release
1639 Select how much the release time is affected by ASC, 0 means nearly no changes
1640 in release time while 1 produces higher release times.
1643 Auto level output signal. Default is enabled.
1644 This normalizes audio back to 0dB if enabled.
1647 Depending on picked setting it is recommended to upsample input 2x or 4x times
1648 with @ref{aresample} before applying this filter.
1652 Apply a two-pole all-pass filter with central frequency (in Hz)
1653 @var{frequency}, and filter-width @var{width}.
1654 An all-pass filter changes the audio's frequency to phase relationship
1655 without changing its frequency to amplitude relationship.
1657 The filter accepts the following options:
1661 Set frequency in Hz.
1664 Set method to specify band-width of filter.
1679 Specify the band-width of a filter in width_type units.
1682 How much to use filtered signal in output. Default is 1.
1683 Range is between 0 and 1.
1686 Specify which channels to filter, by default all available are filtered.
1689 Normalize biquad coefficients, by default is disabled.
1690 Enabling it will normalize magnitude response at DC to 0dB.
1693 Set the filter order, can be 1 or 2. Default is 2.
1696 Set transform type of IIR filter.
1705 @subsection Commands
1707 This filter supports the following commands:
1710 Change allpass frequency.
1711 Syntax for the command is : "@var{frequency}"
1714 Change allpass width_type.
1715 Syntax for the command is : "@var{width_type}"
1718 Change allpass width.
1719 Syntax for the command is : "@var{width}"
1723 Syntax for the command is : "@var{mix}"
1730 The filter accepts the following options:
1734 Set the number of loops. Setting this value to -1 will result in infinite loops.
1738 Set maximal number of samples. Default is 0.
1741 Set first sample of loop. Default is 0.
1747 Merge two or more audio streams into a single multi-channel stream.
1749 The filter accepts the following options:
1754 Set the number of inputs. Default is 2.
1758 If the channel layouts of the inputs are disjoint, and therefore compatible,
1759 the channel layout of the output will be set accordingly and the channels
1760 will be reordered as necessary. If the channel layouts of the inputs are not
1761 disjoint, the output will have all the channels of the first input then all
1762 the channels of the second input, in that order, and the channel layout of
1763 the output will be the default value corresponding to the total number of
1766 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1767 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1768 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1769 first input, b1 is the first channel of the second input).
1771 On the other hand, if both input are in stereo, the output channels will be
1772 in the default order: a1, a2, b1, b2, and the channel layout will be
1773 arbitrarily set to 4.0, which may or may not be the expected value.
1775 All inputs must have the same sample rate, and format.
1777 If inputs do not have the same duration, the output will stop with the
1780 @subsection Examples
1784 Merge two mono files into a stereo stream:
1786 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1790 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1792 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
1798 Mixes multiple audio inputs into a single output.
1800 Note that this filter only supports float samples (the @var{amerge}
1801 and @var{pan} audio filters support many formats). If the @var{amix}
1802 input has integer samples then @ref{aresample} will be automatically
1803 inserted to perform the conversion to float samples.
1807 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1809 will mix 3 input audio streams to a single output with the same duration as the
1810 first input and a dropout transition time of 3 seconds.
1812 It accepts the following parameters:
1816 The number of inputs. If unspecified, it defaults to 2.
1819 How to determine the end-of-stream.
1823 The duration of the longest input. (default)
1826 The duration of the shortest input.
1829 The duration of the first input.
1833 @item dropout_transition
1834 The transition time, in seconds, for volume renormalization when an input
1835 stream ends. The default value is 2 seconds.
1838 Specify weight of each input audio stream as sequence.
1839 Each weight is separated by space. By default all inputs have same weight.
1842 @subsection Commands
1844 This filter supports the following commands:
1847 Syntax is same as option with same name.
1852 Multiply first audio stream with second audio stream and store result
1853 in output audio stream. Multiplication is done by multiplying each
1854 sample from first stream with sample at same position from second stream.
1856 With this element-wise multiplication one can create amplitude fades and
1857 amplitude modulations.
1859 @section anequalizer
1861 High-order parametric multiband equalizer for each channel.
1863 It accepts the following parameters:
1867 This option string is in format:
1868 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1869 Each equalizer band is separated by '|'.
1873 Set channel number to which equalization will be applied.
1874 If input doesn't have that channel the entry is ignored.
1877 Set central frequency for band.
1878 If input doesn't have that frequency the entry is ignored.
1881 Set band width in Hertz.
1884 Set band gain in dB.
1887 Set filter type for band, optional, can be:
1891 Butterworth, this is default.
1902 With this option activated frequency response of anequalizer is displayed
1906 Set video stream size. Only useful if curves option is activated.
1909 Set max gain that will be displayed. Only useful if curves option is activated.
1910 Setting this to a reasonable value makes it possible to display gain which is derived from
1911 neighbour bands which are too close to each other and thus produce higher gain
1912 when both are activated.
1915 Set frequency scale used to draw frequency response in video output.
1916 Can be linear or logarithmic. Default is logarithmic.
1919 Set color for each channel curve which is going to be displayed in video stream.
1920 This is list of color names separated by space or by '|'.
1921 Unrecognised or missing colors will be replaced by white color.
1924 @subsection Examples
1928 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1929 for first 2 channels using Chebyshev type 1 filter:
1931 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1935 @subsection Commands
1937 This filter supports the following commands:
1940 Alter existing filter parameters.
1941 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1943 @var{fN} is existing filter number, starting from 0, if no such filter is available
1945 @var{freq} set new frequency parameter.
1946 @var{width} set new width parameter in Hertz.
1947 @var{gain} set new gain parameter in dB.
1949 Full filter invocation with asendcmd may look like this:
1950 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1955 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1957 Each sample is adjusted by looking for other samples with similar contexts. This
1958 context similarity is defined by comparing their surrounding patches of size
1959 @option{p}. Patches are searched in an area of @option{r} around the sample.
1961 The filter accepts the following options:
1965 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1968 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1969 Default value is 2 milliseconds.
1972 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1973 Default value is 6 milliseconds.
1976 Set the output mode.
1978 It accepts the following values:
1981 Pass input unchanged.
1984 Pass noise filtered out.
1989 Default value is @var{o}.
1993 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
1996 @subsection Commands
1998 This filter supports the all above options as @ref{commands}.
2001 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2003 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2004 relate to producing the least mean square of the error signal (difference between the desired,
2005 2nd input audio stream and the actual signal, the 1st input audio stream).
2007 A description of the accepted options follows.
2020 Set the filter leakage.
2023 It accepts the following values:
2032 Pass filtered samples.
2035 Pass difference between desired and filtered samples.
2037 Default value is @var{o}.
2041 @subsection Examples
2045 One of many usages of this filter is noise reduction, input audio is filtered
2046 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2048 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2052 @subsection Commands
2054 This filter supports the same commands as options, excluding option @code{order}.
2058 Pass the audio source unchanged to the output.
2062 Pad the end of an audio stream with silence.
2064 This can be used together with @command{ffmpeg} @option{-shortest} to
2065 extend audio streams to the same length as the video stream.
2067 A description of the accepted options follows.
2071 Set silence packet size. Default value is 4096.
2074 Set the number of samples of silence to add to the end. After the
2075 value is reached, the stream is terminated. This option is mutually
2076 exclusive with @option{whole_len}.
2079 Set the minimum total number of samples in the output audio stream. If
2080 the value is longer than the input audio length, silence is added to
2081 the end, until the value is reached. This option is mutually exclusive
2082 with @option{pad_len}.
2085 Specify the duration of samples of silence to add. See
2086 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2087 for the accepted syntax. Used only if set to non-zero value.
2090 Specify the minimum total duration in the output audio stream. See
2091 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2092 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2093 the input audio length, silence is added to the end, until the value is reached.
2094 This option is mutually exclusive with @option{pad_dur}
2097 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2098 nor @option{whole_dur} option is set, the filter will add silence to the end of
2099 the input stream indefinitely.
2101 @subsection Examples
2105 Add 1024 samples of silence to the end of the input:
2111 Make sure the audio output will contain at least 10000 samples, pad
2112 the input with silence if required:
2114 apad=whole_len=10000
2118 Use @command{ffmpeg} to pad the audio input with silence, so that the
2119 video stream will always result the shortest and will be converted
2120 until the end in the output file when using the @option{shortest}
2123 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2128 Add a phasing effect to the input audio.
2130 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2131 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2133 A description of the accepted parameters follows.
2137 Set input gain. Default is 0.4.
2140 Set output gain. Default is 0.74
2143 Set delay in milliseconds. Default is 3.0.
2146 Set decay. Default is 0.4.
2149 Set modulation speed in Hz. Default is 0.5.
2152 Set modulation type. Default is triangular.
2154 It accepts the following values:
2161 @section aphaseshift
2162 Apply phase shift to input audio samples.
2164 The filter accepts the following options:
2168 Specify phase shift. Allowed range is from -1.0 to 1.0.
2169 Default value is 0.0.
2172 @subsection Commands
2174 This filter supports the above option as @ref{commands}.
2178 Audio pulsator is something between an autopanner and a tremolo.
2179 But it can produce funny stereo effects as well. Pulsator changes the volume
2180 of the left and right channel based on a LFO (low frequency oscillator) with
2181 different waveforms and shifted phases.
2182 This filter have the ability to define an offset between left and right
2183 channel. An offset of 0 means that both LFO shapes match each other.
2184 The left and right channel are altered equally - a conventional tremolo.
2185 An offset of 50% means that the shape of the right channel is exactly shifted
2186 in phase (or moved backwards about half of the frequency) - pulsator acts as
2187 an autopanner. At 1 both curves match again. Every setting in between moves the
2188 phase shift gapless between all stages and produces some "bypassing" sounds with
2189 sine and triangle waveforms. The more you set the offset near 1 (starting from
2190 the 0.5) the faster the signal passes from the left to the right speaker.
2192 The filter accepts the following options:
2196 Set input gain. By default it is 1. Range is [0.015625 - 64].
2199 Set output gain. By default it is 1. Range is [0.015625 - 64].
2202 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2203 sawup or sawdown. Default is sine.
2206 Set modulation. Define how much of original signal is affected by the LFO.
2209 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2212 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2215 Set pulse width. Default is 1. Allowed range is [0 - 2].
2218 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2221 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2225 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2229 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2230 if timing is set to hz.
2236 Resample the input audio to the specified parameters, using the
2237 libswresample library. If none are specified then the filter will
2238 automatically convert between its input and output.
2240 This filter is also able to stretch/squeeze the audio data to make it match
2241 the timestamps or to inject silence / cut out audio to make it match the
2242 timestamps, do a combination of both or do neither.
2244 The filter accepts the syntax
2245 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2246 expresses a sample rate and @var{resampler_options} is a list of
2247 @var{key}=@var{value} pairs, separated by ":". See the
2248 @ref{Resampler Options,,"Resampler Options" section in the
2249 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2250 for the complete list of supported options.
2252 @subsection Examples
2256 Resample the input audio to 44100Hz:
2262 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2263 samples per second compensation:
2265 aresample=async=1000
2271 Reverse an audio clip.
2273 Warning: This filter requires memory to buffer the entire clip, so trimming
2276 @subsection Examples
2280 Take the first 5 seconds of a clip, and reverse it.
2282 atrim=end=5,areverse
2288 Reduce noise from speech using Recurrent Neural Networks.
2290 This filter accepts the following options:
2294 Set train model file to load. This option is always required.
2297 @section asetnsamples
2299 Set the number of samples per each output audio frame.
2301 The last output packet may contain a different number of samples, as
2302 the filter will flush all the remaining samples when the input audio
2305 The filter accepts the following options:
2309 @item nb_out_samples, n
2310 Set the number of frames per each output audio frame. The number is
2311 intended as the number of samples @emph{per each channel}.
2312 Default value is 1024.
2315 If set to 1, the filter will pad the last audio frame with zeroes, so
2316 that the last frame will contain the same number of samples as the
2317 previous ones. Default value is 1.
2320 For example, to set the number of per-frame samples to 1234 and
2321 disable padding for the last frame, use:
2323 asetnsamples=n=1234:p=0
2328 Set the sample rate without altering the PCM data.
2329 This will result in a change of speed and pitch.
2331 The filter accepts the following options:
2334 @item sample_rate, r
2335 Set the output sample rate. Default is 44100 Hz.
2340 Show a line containing various information for each input audio frame.
2341 The input audio is not modified.
2343 The shown line contains a sequence of key/value pairs of the form
2344 @var{key}:@var{value}.
2346 The following values are shown in the output:
2350 The (sequential) number of the input frame, starting from 0.
2353 The presentation timestamp of the input frame, in time base units; the time base
2354 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2357 The presentation timestamp of the input frame in seconds.
2360 position of the frame in the input stream, -1 if this information in
2361 unavailable and/or meaningless (for example in case of synthetic audio)
2370 The sample rate for the audio frame.
2373 The number of samples (per channel) in the frame.
2376 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2377 audio, the data is treated as if all the planes were concatenated.
2379 @item plane_checksums
2380 A list of Adler-32 checksums for each data plane.
2384 Apply audio soft clipping.
2386 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2387 along a smooth curve, rather than the abrupt shape of hard-clipping.
2389 This filter accepts the following options:
2393 Set type of soft-clipping.
2395 It accepts the following values:
2409 Set additional parameter which controls sigmoid function.
2412 Set oversampling factor.
2415 @subsection Commands
2417 This filter supports the all above options as @ref{commands}.
2420 Automatic Speech Recognition
2422 This filter uses PocketSphinx for speech recognition. To enable
2423 compilation of this filter, you need to configure FFmpeg with
2424 @code{--enable-pocketsphinx}.
2426 It accepts the following options:
2430 Set sampling rate of input audio. Defaults is @code{16000}.
2431 This need to match speech models, otherwise one will get poor results.
2434 Set dictionary containing acoustic model files.
2437 Set pronunciation dictionary.
2440 Set language model file.
2443 Set language model set.
2446 Set which language model to use.
2449 Set output for log messages.
2452 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2457 Display time domain statistical information about the audio channels.
2458 Statistics are calculated and displayed for each audio channel and,
2459 where applicable, an overall figure is also given.
2461 It accepts the following option:
2464 Short window length in seconds, used for peak and trough RMS measurement.
2465 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2469 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2470 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2473 Available keys for each channel are:
2519 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2520 this @code{lavfi.astats.Overall.Peak_count}.
2522 For description what each key means read below.
2525 Set number of frame after which stats are going to be recalculated.
2526 Default is disabled.
2528 @item measure_perchannel
2529 Select the entries which need to be measured per channel. The metadata keys can
2530 be used as flags, default is @option{all} which measures everything.
2531 @option{none} disables all per channel measurement.
2533 @item measure_overall
2534 Select the entries which need to be measured overall. The metadata keys can
2535 be used as flags, default is @option{all} which measures everything.
2536 @option{none} disables all overall measurement.
2540 A description of each shown parameter follows:
2544 Mean amplitude displacement from zero.
2547 Minimal sample level.
2550 Maximal sample level.
2552 @item Min difference
2553 Minimal difference between two consecutive samples.
2555 @item Max difference
2556 Maximal difference between two consecutive samples.
2558 @item Mean difference
2559 Mean difference between two consecutive samples.
2560 The average of each difference between two consecutive samples.
2562 @item RMS difference
2563 Root Mean Square difference between two consecutive samples.
2567 Standard peak and RMS level measured in dBFS.
2571 Peak and trough values for RMS level measured over a short window.
2574 Standard ratio of peak to RMS level (note: not in dB).
2577 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2578 (i.e. either @var{Min level} or @var{Max level}).
2581 Number of occasions (not the number of samples) that the signal attained either
2582 @var{Min level} or @var{Max level}.
2584 @item Noise floor dB
2585 Minimum local peak measured in dBFS over a short window.
2587 @item Noise floor count
2588 Number of occasions (not the number of samples) that the signal attained
2592 Overall bit depth of audio. Number of bits used for each sample.
2595 Measured dynamic range of audio in dB.
2597 @item Zero crossings
2598 Number of points where the waveform crosses the zero level axis.
2600 @item Zero crossings rate
2601 Rate of Zero crossings and number of audio samples.
2605 Boost subwoofer frequencies.
2607 The filter accepts the following options:
2611 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2612 Default value is 0.7.
2615 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2616 Default value is 0.7.
2619 Set delay line decay gain value. Allowed range is from 0 to 1.
2620 Default value is 0.7.
2623 Set delay line feedback gain value. Allowed range is from 0 to 1.
2624 Default value is 0.9.
2627 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2628 Default value is 100.
2631 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2632 Default value is 0.5.
2635 Set delay. Allowed range is from 1 to 100.
2636 Default value is 20.
2639 @subsection Commands
2641 This filter supports the all above options as @ref{commands}.
2644 Cut super frequencies.
2646 The filter accepts the following options:
2650 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2651 Default value is 20000.
2654 Set filter order. Available values are from 3 to 20.
2655 Default value is 10.
2658 @subsection Commands
2660 This filter supports the all above options as @ref{commands}.
2666 The filter accepts exactly one parameter, the audio tempo. If not
2667 specified then the filter will assume nominal 1.0 tempo. Tempo must
2668 be in the [0.5, 100.0] range.
2670 Note that tempo greater than 2 will skip some samples rather than
2671 blend them in. If for any reason this is a concern it is always
2672 possible to daisy-chain several instances of atempo to achieve the
2673 desired product tempo.
2675 @subsection Examples
2679 Slow down audio to 80% tempo:
2685 To speed up audio to 300% tempo:
2691 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2693 atempo=sqrt(3),atempo=sqrt(3)
2697 @subsection Commands
2699 This filter supports the following commands:
2702 Change filter tempo scale factor.
2703 Syntax for the command is : "@var{tempo}"
2708 Trim the input so that the output contains one continuous subpart of the input.
2710 It accepts the following parameters:
2713 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2714 sample with the timestamp @var{start} will be the first sample in the output.
2717 Specify time of the first audio sample that will be dropped, i.e. the
2718 audio sample immediately preceding the one with the timestamp @var{end} will be
2719 the last sample in the output.
2722 Same as @var{start}, except this option sets the start timestamp in samples
2726 Same as @var{end}, except this option sets the end timestamp in samples instead
2730 The maximum duration of the output in seconds.
2733 The number of the first sample that should be output.
2736 The number of the first sample that should be dropped.
2739 @option{start}, @option{end}, and @option{duration} are expressed as time
2740 duration specifications; see
2741 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2743 Note that the first two sets of the start/end options and the @option{duration}
2744 option look at the frame timestamp, while the _sample options simply count the
2745 samples that pass through the filter. So start/end_pts and start/end_sample will
2746 give different results when the timestamps are wrong, inexact or do not start at
2747 zero. Also note that this filter does not modify the timestamps. If you wish
2748 to have the output timestamps start at zero, insert the asetpts filter after the
2751 If multiple start or end options are set, this filter tries to be greedy and
2752 keep all samples that match at least one of the specified constraints. To keep
2753 only the part that matches all the constraints at once, chain multiple atrim
2756 The defaults are such that all the input is kept. So it is possible to set e.g.
2757 just the end values to keep everything before the specified time.
2762 Drop everything except the second minute of input:
2764 ffmpeg -i INPUT -af atrim=60:120
2768 Keep only the first 1000 samples:
2770 ffmpeg -i INPUT -af atrim=end_sample=1000
2775 @section axcorrelate
2776 Calculate normalized cross-correlation between two input audio streams.
2778 Resulted samples are always between -1 and 1 inclusive.
2779 If result is 1 it means two input samples are highly correlated in that selected segment.
2780 Result 0 means they are not correlated at all.
2781 If result is -1 it means two input samples are out of phase, which means they cancel each
2784 The filter accepts the following options:
2788 Set size of segment over which cross-correlation is calculated.
2789 Default is 256. Allowed range is from 2 to 131072.
2792 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2793 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2794 are always zero and thus need much less calculations to make.
2795 This is generally not true, but is valid for typical audio streams.
2798 @subsection Examples
2802 Calculate correlation between channels in stereo audio stream:
2804 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2810 Apply a two-pole Butterworth band-pass filter with central
2811 frequency @var{frequency}, and (3dB-point) band-width width.
2812 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2813 instead of the default: constant 0dB peak gain.
2814 The filter roll off at 6dB per octave (20dB per decade).
2816 The filter accepts the following options:
2820 Set the filter's central frequency. Default is @code{3000}.
2823 Constant skirt gain if set to 1. Defaults to 0.
2826 Set method to specify band-width of filter.
2841 Specify the band-width of a filter in width_type units.
2844 How much to use filtered signal in output. Default is 1.
2845 Range is between 0 and 1.
2848 Specify which channels to filter, by default all available are filtered.
2851 Normalize biquad coefficients, by default is disabled.
2852 Enabling it will normalize magnitude response at DC to 0dB.
2855 Set transform type of IIR filter.
2864 @subsection Commands
2866 This filter supports the following commands:
2869 Change bandpass frequency.
2870 Syntax for the command is : "@var{frequency}"
2873 Change bandpass width_type.
2874 Syntax for the command is : "@var{width_type}"
2877 Change bandpass width.
2878 Syntax for the command is : "@var{width}"
2881 Change bandpass mix.
2882 Syntax for the command is : "@var{mix}"
2887 Apply a two-pole Butterworth band-reject filter with central
2888 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2889 The filter roll off at 6dB per octave (20dB per decade).
2891 The filter accepts the following options:
2895 Set the filter's central frequency. Default is @code{3000}.
2898 Set method to specify band-width of filter.
2913 Specify the band-width of a filter in width_type units.
2916 How much to use filtered signal in output. Default is 1.
2917 Range is between 0 and 1.
2920 Specify which channels to filter, by default all available are filtered.
2923 Normalize biquad coefficients, by default is disabled.
2924 Enabling it will normalize magnitude response at DC to 0dB.
2927 Set transform type of IIR filter.
2936 @subsection Commands
2938 This filter supports the following commands:
2941 Change bandreject frequency.
2942 Syntax for the command is : "@var{frequency}"
2945 Change bandreject width_type.
2946 Syntax for the command is : "@var{width_type}"
2949 Change bandreject width.
2950 Syntax for the command is : "@var{width}"
2953 Change bandreject mix.
2954 Syntax for the command is : "@var{mix}"
2957 @section bass, lowshelf
2959 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2960 shelving filter with a response similar to that of a standard
2961 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2963 The filter accepts the following options:
2967 Give the gain at 0 Hz. Its useful range is about -20
2968 (for a large cut) to +20 (for a large boost).
2969 Beware of clipping when using a positive gain.
2972 Set the filter's central frequency and so can be used
2973 to extend or reduce the frequency range to be boosted or cut.
2974 The default value is @code{100} Hz.
2977 Set method to specify band-width of filter.
2992 Determine how steep is the filter's shelf transition.
2995 How much to use filtered signal in output. Default is 1.
2996 Range is between 0 and 1.
2999 Specify which channels to filter, by default all available are filtered.
3002 Normalize biquad coefficients, by default is disabled.
3003 Enabling it will normalize magnitude response at DC to 0dB.
3006 Set transform type of IIR filter.
3015 @subsection Commands
3017 This filter supports the following commands:
3020 Change bass frequency.
3021 Syntax for the command is : "@var{frequency}"
3024 Change bass width_type.
3025 Syntax for the command is : "@var{width_type}"
3029 Syntax for the command is : "@var{width}"
3033 Syntax for the command is : "@var{gain}"
3037 Syntax for the command is : "@var{mix}"
3042 Apply a biquad IIR filter with the given coefficients.
3043 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3044 are the numerator and denominator coefficients respectively.
3045 and @var{channels}, @var{c} specify which channels to filter, by default all
3046 available are filtered.
3048 @subsection Commands
3050 This filter supports the following commands:
3058 Change biquad parameter.
3059 Syntax for the command is : "@var{value}"
3062 How much to use filtered signal in output. Default is 1.
3063 Range is between 0 and 1.
3066 Specify which channels to filter, by default all available are filtered.
3069 Normalize biquad coefficients, by default is disabled.
3070 Enabling it will normalize magnitude response at DC to 0dB.
3073 Set transform type of IIR filter.
3083 Bauer stereo to binaural transformation, which improves headphone listening of
3084 stereo audio records.
3086 To enable compilation of this filter you need to configure FFmpeg with
3087 @code{--enable-libbs2b}.
3089 It accepts the following parameters:
3093 Pre-defined crossfeed level.
3097 Default level (fcut=700, feed=50).
3100 Chu Moy circuit (fcut=700, feed=60).
3103 Jan Meier circuit (fcut=650, feed=95).
3108 Cut frequency (in Hz).
3117 Remap input channels to new locations.
3119 It accepts the following parameters:
3122 Map channels from input to output. The argument is a '|'-separated list of
3123 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3124 @var{in_channel} form. @var{in_channel} can be either the name of the input
3125 channel (e.g. FL for front left) or its index in the input channel layout.
3126 @var{out_channel} is the name of the output channel or its index in the output
3127 channel layout. If @var{out_channel} is not given then it is implicitly an
3128 index, starting with zero and increasing by one for each mapping.
3130 @item channel_layout
3131 The channel layout of the output stream.
3134 If no mapping is present, the filter will implicitly map input channels to
3135 output channels, preserving indices.
3137 @subsection Examples
3141 For example, assuming a 5.1+downmix input MOV file,
3143 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3145 will create an output WAV file tagged as stereo from the downmix channels of
3149 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3151 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3155 @section channelsplit
3157 Split each channel from an input audio stream into a separate output stream.
3159 It accepts the following parameters:
3161 @item channel_layout
3162 The channel layout of the input stream. The default is "stereo".
3164 A channel layout describing the channels to be extracted as separate output streams
3165 or "all" to extract each input channel as a separate stream. The default is "all".
3167 Choosing channels not present in channel layout in the input will result in an error.
3170 @subsection Examples
3174 For example, assuming a stereo input MP3 file,
3176 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3178 will create an output Matroska file with two audio streams, one containing only
3179 the left channel and the other the right channel.
3182 Split a 5.1 WAV file into per-channel files:
3184 ffmpeg -i in.wav -filter_complex
3185 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3186 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3187 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3192 Extract only LFE from a 5.1 WAV file:
3194 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3195 -map '[LFE]' lfe.wav
3200 Add a chorus effect to the audio.
3202 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3204 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3205 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3206 The modulation depth defines the range the modulated delay is played before or after
3207 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3208 sound tuned around the original one, like in a chorus where some vocals are slightly
3211 It accepts the following parameters:
3214 Set input gain. Default is 0.4.
3217 Set output gain. Default is 0.4.
3220 Set delays. A typical delay is around 40ms to 60ms.
3232 @subsection Examples
3238 chorus=0.7:0.9:55:0.4:0.25:2
3244 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3248 Fuller sounding chorus with three delays:
3250 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
3255 Compress or expand the audio's dynamic range.
3257 It accepts the following parameters:
3263 A list of times in seconds for each channel over which the instantaneous level
3264 of the input signal is averaged to determine its volume. @var{attacks} refers to
3265 increase of volume and @var{decays} refers to decrease of volume. For most
3266 situations, the attack time (response to the audio getting louder) should be
3267 shorter than the decay time, because the human ear is more sensitive to sudden
3268 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3269 a typical value for decay is 0.8 seconds.
3270 If specified number of attacks & decays is lower than number of channels, the last
3271 set attack/decay will be used for all remaining channels.
3274 A list of points for the transfer function, specified in dB relative to the
3275 maximum possible signal amplitude. Each key points list must be defined using
3276 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3277 @code{x0/y0 x1/y1 x2/y2 ....}
3279 The input values must be in strictly increasing order but the transfer function
3280 does not have to be monotonically rising. The point @code{0/0} is assumed but
3281 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3282 function are @code{-70/-70|-60/-20|1/0}.
3285 Set the curve radius in dB for all joints. It defaults to 0.01.
3288 Set the additional gain in dB to be applied at all points on the transfer
3289 function. This allows for easy adjustment of the overall gain.
3293 Set an initial volume, in dB, to be assumed for each channel when filtering
3294 starts. This permits the user to supply a nominal level initially, so that, for
3295 example, a very large gain is not applied to initial signal levels before the
3296 companding has begun to operate. A typical value for audio which is initially
3297 quiet is -90 dB. It defaults to 0.
3300 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3301 delayed before being fed to the volume adjuster. Specifying a delay
3302 approximately equal to the attack/decay times allows the filter to effectively
3303 operate in predictive rather than reactive mode. It defaults to 0.
3307 @subsection Examples
3311 Make music with both quiet and loud passages suitable for listening to in a
3314 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3317 Another example for audio with whisper and explosion parts:
3319 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3323 A noise gate for when the noise is at a lower level than the signal:
3325 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3329 Here is another noise gate, this time for when the noise is at a higher level
3330 than the signal (making it, in some ways, similar to squelch):
3332 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3336 2:1 compression starting at -6dB:
3338 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3342 2:1 compression starting at -9dB:
3344 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3348 2:1 compression starting at -12dB:
3350 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3354 2:1 compression starting at -18dB:
3356 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3360 3:1 compression starting at -15dB:
3362 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3368 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3374 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
3378 Hard limiter at -6dB:
3380 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3384 Hard limiter at -12dB:
3386 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3390 Hard noise gate at -35 dB:
3392 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3398 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3402 @section compensationdelay
3404 Compensation Delay Line is a metric based delay to compensate differing
3405 positions of microphones or speakers.
3407 For example, you have recorded guitar with two microphones placed in
3408 different locations. Because the front of sound wave has fixed speed in
3409 normal conditions, the phasing of microphones can vary and depends on
3410 their location and interposition. The best sound mix can be achieved when
3411 these microphones are in phase (synchronized). Note that a distance of
3412 ~30 cm between microphones makes one microphone capture the signal in
3413 antiphase to the other microphone. That makes the final mix sound moody.
3414 This filter helps to solve phasing problems by adding different delays
3415 to each microphone track and make them synchronized.
3417 The best result can be reached when you take one track as base and
3418 synchronize other tracks one by one with it.
3419 Remember that synchronization/delay tolerance depends on sample rate, too.
3420 Higher sample rates will give more tolerance.
3422 The filter accepts the following parameters:
3426 Set millimeters distance. This is compensation distance for fine tuning.
3430 Set cm distance. This is compensation distance for tightening distance setup.
3434 Set meters distance. This is compensation distance for hard distance setup.
3438 Set dry amount. Amount of unprocessed (dry) signal.
3442 Set wet amount. Amount of processed (wet) signal.
3446 Set temperature in degrees Celsius. This is the temperature of the environment.
3451 Apply headphone crossfeed filter.
3453 Crossfeed is the process of blending the left and right channels of stereo
3455 It is mainly used to reduce extreme stereo separation of low frequencies.
3457 The intent is to produce more speaker like sound to the listener.
3459 The filter accepts the following options:
3463 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3464 This sets gain of low shelf filter for side part of stereo image.
3465 Default is -6dB. Max allowed is -30db when strength is set to 1.
3468 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3469 This sets cut off frequency of low shelf filter. Default is cut off near
3470 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3473 Set curve slope of low shelf filter. Default is 0.5.
3474 Allowed range is from 0.01 to 1.
3477 Set input gain. Default is 0.9.
3480 Set output gain. Default is 1.
3483 @subsection Commands
3485 This filter supports the all above options as @ref{commands}.
3487 @section crystalizer
3488 Simple algorithm to expand audio dynamic range.
3490 The filter accepts the following options:
3494 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3495 (unchanged sound) to 10.0 (maximum effect).
3498 Enable clipping. By default is enabled.
3501 @subsection Commands
3503 This filter supports the all above options as @ref{commands}.
3506 Apply a DC shift to the audio.
3508 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3509 in the recording chain) from the audio. The effect of a DC offset is reduced
3510 headroom and hence volume. The @ref{astats} filter can be used to determine if
3511 a signal has a DC offset.
3515 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3519 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3520 used to prevent clipping.
3525 Apply de-essing to the audio samples.
3529 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3533 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3537 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3541 Set the output mode.
3543 It accepts the following values:
3546 Pass input unchanged.
3549 Pass ess filtered out.
3554 Default value is @var{o}.
3560 Measure audio dynamic range.
3562 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3563 is found in transition material. And anything less that 8 have very poor dynamics
3564 and is very compressed.
3566 The filter accepts the following options:
3570 Set window length in seconds used to split audio into segments of equal length.
3571 Default is 3 seconds.
3575 Dynamic Audio Normalizer.
3577 This filter applies a certain amount of gain to the input audio in order
3578 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3579 contrast to more "simple" normalization algorithms, the Dynamic Audio
3580 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3581 This allows for applying extra gain to the "quiet" sections of the audio
3582 while avoiding distortions or clipping the "loud" sections. In other words:
3583 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3584 sections, in the sense that the volume of each section is brought to the
3585 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3586 this goal *without* applying "dynamic range compressing". It will retain 100%
3587 of the dynamic range *within* each section of the audio file.
3591 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3592 Default is 500 milliseconds.
3593 The Dynamic Audio Normalizer processes the input audio in small chunks,
3594 referred to as frames. This is required, because a peak magnitude has no
3595 meaning for just a single sample value. Instead, we need to determine the
3596 peak magnitude for a contiguous sequence of sample values. While a "standard"
3597 normalizer would simply use the peak magnitude of the complete file, the
3598 Dynamic Audio Normalizer determines the peak magnitude individually for each
3599 frame. The length of a frame is specified in milliseconds. By default, the
3600 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3601 been found to give good results with most files.
3602 Note that the exact frame length, in number of samples, will be determined
3603 automatically, based on the sampling rate of the individual input audio file.
3606 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3607 number. Default is 31.
3608 Probably the most important parameter of the Dynamic Audio Normalizer is the
3609 @code{window size} of the Gaussian smoothing filter. The filter's window size
3610 is specified in frames, centered around the current frame. For the sake of
3611 simplicity, this must be an odd number. Consequently, the default value of 31
3612 takes into account the current frame, as well as the 15 preceding frames and
3613 the 15 subsequent frames. Using a larger window results in a stronger
3614 smoothing effect and thus in less gain variation, i.e. slower gain
3615 adaptation. Conversely, using a smaller window results in a weaker smoothing
3616 effect and thus in more gain variation, i.e. faster gain adaptation.
3617 In other words, the more you increase this value, the more the Dynamic Audio
3618 Normalizer will behave like a "traditional" normalization filter. On the
3619 contrary, the more you decrease this value, the more the Dynamic Audio
3620 Normalizer will behave like a dynamic range compressor.
3623 Set the target peak value. This specifies the highest permissible magnitude
3624 level for the normalized audio input. This filter will try to approach the
3625 target peak magnitude as closely as possible, but at the same time it also
3626 makes sure that the normalized signal will never exceed the peak magnitude.
3627 A frame's maximum local gain factor is imposed directly by the target peak
3628 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3629 It is not recommended to go above this value.
3632 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3633 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3634 factor for each input frame, i.e. the maximum gain factor that does not
3635 result in clipping or distortion. The maximum gain factor is determined by
3636 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3637 additionally bounds the frame's maximum gain factor by a predetermined
3638 (global) maximum gain factor. This is done in order to avoid excessive gain
3639 factors in "silent" or almost silent frames. By default, the maximum gain
3640 factor is 10.0, For most inputs the default value should be sufficient and
3641 it usually is not recommended to increase this value. Though, for input
3642 with an extremely low overall volume level, it may be necessary to allow even
3643 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3644 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3645 Instead, a "sigmoid" threshold function will be applied. This way, the
3646 gain factors will smoothly approach the threshold value, but never exceed that
3650 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3651 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3652 This means that the maximum local gain factor for each frame is defined
3653 (only) by the frame's highest magnitude sample. This way, the samples can
3654 be amplified as much as possible without exceeding the maximum signal
3655 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3656 Normalizer can also take into account the frame's root mean square,
3657 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3658 determine the power of a time-varying signal. It is therefore considered
3659 that the RMS is a better approximation of the "perceived loudness" than
3660 just looking at the signal's peak magnitude. Consequently, by adjusting all
3661 frames to a constant RMS value, a uniform "perceived loudness" can be
3662 established. If a target RMS value has been specified, a frame's local gain
3663 factor is defined as the factor that would result in exactly that RMS value.
3664 Note, however, that the maximum local gain factor is still restricted by the
3665 frame's highest magnitude sample, in order to prevent clipping.
3668 Enable channels coupling. By default is enabled.
3669 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3670 amount. This means the same gain factor will be applied to all channels, i.e.
3671 the maximum possible gain factor is determined by the "loudest" channel.
3672 However, in some recordings, it may happen that the volume of the different
3673 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3674 In this case, this option can be used to disable the channel coupling. This way,
3675 the gain factor will be determined independently for each channel, depending
3676 only on the individual channel's highest magnitude sample. This allows for
3677 harmonizing the volume of the different channels.
3680 Enable DC bias correction. By default is disabled.
3681 An audio signal (in the time domain) is a sequence of sample values.
3682 In the Dynamic Audio Normalizer these sample values are represented in the
3683 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3684 audio signal, or "waveform", should be centered around the zero point.
3685 That means if we calculate the mean value of all samples in a file, or in a
3686 single frame, then the result should be 0.0 or at least very close to that
3687 value. If, however, there is a significant deviation of the mean value from
3688 0.0, in either positive or negative direction, this is referred to as a
3689 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3690 Audio Normalizer provides optional DC bias correction.
3691 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3692 the mean value, or "DC correction" offset, of each input frame and subtract
3693 that value from all of the frame's sample values which ensures those samples
3694 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3695 boundaries, the DC correction offset values will be interpolated smoothly
3696 between neighbouring frames.
3698 @item altboundary, b
3699 Enable alternative boundary mode. By default is disabled.
3700 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3701 around each frame. This includes the preceding frames as well as the
3702 subsequent frames. However, for the "boundary" frames, located at the very
3703 beginning and at the very end of the audio file, not all neighbouring
3704 frames are available. In particular, for the first few frames in the audio
3705 file, the preceding frames are not known. And, similarly, for the last few
3706 frames in the audio file, the subsequent frames are not known. Thus, the
3707 question arises which gain factors should be assumed for the missing frames
3708 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3709 to deal with this situation. The default boundary mode assumes a gain factor
3710 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3711 "fade out" at the beginning and at the end of the input, respectively.
3714 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3715 By default, the Dynamic Audio Normalizer does not apply "traditional"
3716 compression. This means that signal peaks will not be pruned and thus the
3717 full dynamic range will be retained within each local neighbourhood. However,
3718 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3719 normalization algorithm with a more "traditional" compression.
3720 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3721 (thresholding) function. If (and only if) the compression feature is enabled,
3722 all input frames will be processed by a soft knee thresholding function prior
3723 to the actual normalization process. Put simply, the thresholding function is
3724 going to prune all samples whose magnitude exceeds a certain threshold value.
3725 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3726 value. Instead, the threshold value will be adjusted for each individual
3728 In general, smaller parameters result in stronger compression, and vice versa.
3729 Values below 3.0 are not recommended, because audible distortion may appear.
3732 Set the target threshold value. This specifies the lowest permissible
3733 magnitude level for the audio input which will be normalized.
3734 If input frame volume is above this value frame will be normalized.
3735 Otherwise frame may not be normalized at all. The default value is set
3736 to 0, which means all input frames will be normalized.
3737 This option is mostly useful if digital noise is not wanted to be amplified.
3740 @subsection Commands
3742 This filter supports the all above options as @ref{commands}.
3746 Make audio easier to listen to on headphones.
3748 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3749 so that when listened to on headphones the stereo image is moved from
3750 inside your head (standard for headphones) to outside and in front of
3751 the listener (standard for speakers).
3757 Apply a two-pole peaking equalisation (EQ) filter. With this
3758 filter, the signal-level at and around a selected frequency can
3759 be increased or decreased, whilst (unlike bandpass and bandreject
3760 filters) that at all other frequencies is unchanged.
3762 In order to produce complex equalisation curves, this filter can
3763 be given several times, each with a different central frequency.
3765 The filter accepts the following options:
3769 Set the filter's central frequency in Hz.
3772 Set method to specify band-width of filter.
3787 Specify the band-width of a filter in width_type units.
3790 Set the required gain or attenuation in dB.
3791 Beware of clipping when using a positive gain.
3794 How much to use filtered signal in output. Default is 1.
3795 Range is between 0 and 1.
3798 Specify which channels to filter, by default all available are filtered.
3801 Normalize biquad coefficients, by default is disabled.
3802 Enabling it will normalize magnitude response at DC to 0dB.
3805 Set transform type of IIR filter.
3814 @subsection Examples
3817 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3819 equalizer=f=1000:t=h:width=200:g=-10
3823 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3825 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3829 @subsection Commands
3831 This filter supports the following commands:
3834 Change equalizer frequency.
3835 Syntax for the command is : "@var{frequency}"
3838 Change equalizer width_type.
3839 Syntax for the command is : "@var{width_type}"
3842 Change equalizer width.
3843 Syntax for the command is : "@var{width}"
3846 Change equalizer gain.
3847 Syntax for the command is : "@var{gain}"
3850 Change equalizer mix.
3851 Syntax for the command is : "@var{mix}"
3854 @section extrastereo
3856 Linearly increases the difference between left and right channels which
3857 adds some sort of "live" effect to playback.
3859 The filter accepts the following options:
3863 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3864 (average of both channels), with 1.0 sound will be unchanged, with
3865 -1.0 left and right channels will be swapped.
3868 Enable clipping. By default is enabled.
3871 @subsection Commands
3873 This filter supports the all above options as @ref{commands}.
3875 @section firequalizer
3876 Apply FIR Equalization using arbitrary frequency response.
3878 The filter accepts the following option:
3882 Set gain curve equation (in dB). The expression can contain variables:
3885 the evaluated frequency
3889 channel number, set to 0 when multichannels evaluation is disabled
3891 channel id, see libavutil/channel_layout.h, set to the first channel id when
3892 multichannels evaluation is disabled
3896 channel_layout, see libavutil/channel_layout.h
3901 @item gain_interpolate(f)
3902 interpolate gain on frequency f based on gain_entry
3903 @item cubic_interpolate(f)
3904 same as gain_interpolate, but smoother
3906 This option is also available as command. Default is @code{gain_interpolate(f)}.
3909 Set gain entry for gain_interpolate function. The expression can
3913 store gain entry at frequency f with value g
3915 This option is also available as command.
3918 Set filter delay in seconds. Higher value means more accurate.
3919 Default is @code{0.01}.
3922 Set filter accuracy in Hz. Lower value means more accurate.
3923 Default is @code{5}.
3926 Set window function. Acceptable values are:
3929 rectangular window, useful when gain curve is already smooth
3931 hann window (default)
3937 3-terms continuous 1st derivative nuttall window
3939 minimum 3-terms discontinuous nuttall window
3941 4-terms continuous 1st derivative nuttall window
3943 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3945 blackman-harris window
3951 If enabled, use fixed number of audio samples. This improves speed when
3952 filtering with large delay. Default is disabled.
3955 Enable multichannels evaluation on gain. Default is disabled.
3958 Enable zero phase mode by subtracting timestamp to compensate delay.
3959 Default is disabled.
3962 Set scale used by gain. Acceptable values are:
3965 linear frequency, linear gain
3967 linear frequency, logarithmic (in dB) gain (default)
3969 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3971 logarithmic frequency, logarithmic gain
3975 Set file for dumping, suitable for gnuplot.
3978 Set scale for dumpfile. Acceptable values are same with scale option.
3982 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3983 Default is disabled.
3986 Enable minimum phase impulse response. Default is disabled.
3989 @subsection Examples
3994 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3997 lowpass at 1000 Hz with gain_entry:
3999 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4002 custom equalization:
4004 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4007 higher delay with zero phase to compensate delay:
4009 firequalizer=delay=0.1:fixed=on:zero_phase=on
4012 lowpass on left channel, highpass on right channel:
4014 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4015 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4020 Apply a flanging effect to the audio.
4022 The filter accepts the following options:
4026 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4029 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4032 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4036 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4037 Default value is 71.
4040 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4043 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4044 Default value is @var{sinusoidal}.
4047 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4048 Default value is 25.
4051 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4052 Default is @var{linear}.
4056 Apply Haas effect to audio.
4058 Note that this makes most sense to apply on mono signals.
4059 With this filter applied to mono signals it give some directionality and
4060 stretches its stereo image.
4062 The filter accepts the following options:
4066 Set input level. By default is @var{1}, or 0dB
4069 Set output level. By default is @var{1}, or 0dB.
4072 Set gain applied to side part of signal. By default is @var{1}.
4075 Set kind of middle source. Can be one of the following:
4085 Pick middle part signal of stereo image.
4088 Pick side part signal of stereo image.
4092 Change middle phase. By default is disabled.
4095 Set left channel delay. By default is @var{2.05} milliseconds.
4098 Set left channel balance. By default is @var{-1}.
4101 Set left channel gain. By default is @var{1}.
4104 Change left phase. By default is disabled.
4107 Set right channel delay. By defaults is @var{2.12} milliseconds.
4110 Set right channel balance. By default is @var{1}.
4113 Set right channel gain. By default is @var{1}.
4116 Change right phase. By default is enabled.
4121 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4122 embedded HDCD codes is expanded into a 20-bit PCM stream.
4124 The filter supports the Peak Extend and Low-level Gain Adjustment features
4125 of HDCD, and detects the Transient Filter flag.
4128 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4131 When using the filter with wav, note the default encoding for wav is 16-bit,
4132 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4133 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4135 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4136 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4139 The filter accepts the following options:
4142 @item disable_autoconvert
4143 Disable any automatic format conversion or resampling in the filter graph.
4145 @item process_stereo
4146 Process the stereo channels together. If target_gain does not match between
4147 channels, consider it invalid and use the last valid target_gain.
4150 Set the code detect timer period in ms.
4153 Always extend peaks above -3dBFS even if PE isn't signaled.
4156 Replace audio with a solid tone and adjust the amplitude to signal some
4157 specific aspect of the decoding process. The output file can be loaded in
4158 an audio editor alongside the original to aid analysis.
4160 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4167 Gain adjustment level at each sample
4169 Samples where peak extend occurs
4171 Samples where the code detect timer is active
4173 Samples where the target gain does not match between channels
4179 Apply head-related transfer functions (HRTFs) to create virtual
4180 loudspeakers around the user for binaural listening via headphones.
4181 The HRIRs are provided via additional streams, for each channel
4182 one stereo input stream is needed.
4184 The filter accepts the following options:
4188 Set mapping of input streams for convolution.
4189 The argument is a '|'-separated list of channel names in order as they
4190 are given as additional stream inputs for filter.
4191 This also specify number of input streams. Number of input streams
4192 must be not less than number of channels in first stream plus one.
4195 Set gain applied to audio. Value is in dB. Default is 0.
4198 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4199 processing audio in time domain which is slow.
4200 @var{freq} is processing audio in frequency domain which is fast.
4201 Default is @var{freq}.
4204 Set custom gain for LFE channels. Value is in dB. Default is 0.
4207 Set size of frame in number of samples which will be processed at once.
4208 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4211 Set format of hrir stream.
4212 Default value is @var{stereo}. Alternative value is @var{multich}.
4213 If value is set to @var{stereo}, number of additional streams should
4214 be greater or equal to number of input channels in first input stream.
4215 Also each additional stream should have stereo number of channels.
4216 If value is set to @var{multich}, number of additional streams should
4217 be exactly one. Also number of input channels of additional stream
4218 should be equal or greater than twice number of channels of first input
4222 @subsection Examples
4226 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4227 each amovie filter use stereo file with IR coefficients as input.
4228 The files give coefficients for each position of virtual loudspeaker:
4231 -filter_complex "amovie=azi_270_ele_0_DFC.wav[sr];amovie=azi_90_ele_0_DFC.wav[sl];amovie=azi_225_ele_0_DFC.wav[br];amovie=azi_135_ele_0_DFC.wav[bl];amovie=azi_0_ele_0_DFC.wav,asplit[fc][lfe];amovie=azi_35_ele_0_DFC.wav[fl];amovie=azi_325_ele_0_DFC.wav[fr];[0:a][fl][fr][fc][lfe][bl][br][sl][sr]headphone=FL|FR|FC|LFE|BL|BR|SL|SR"
4236 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4237 but now in @var{multich} @var{hrir} format.
4239 ffmpeg -i input.wav -filter_complex "amovie=minp.wav[hrirs];[0:a][hrirs]headphone=map=FL|FR|FC|LFE|BL|BR|SL|SR:hrir=multich"
4246 Apply a high-pass filter with 3dB point frequency.
4247 The filter can be either single-pole, or double-pole (the default).
4248 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4250 The filter accepts the following options:
4254 Set frequency in Hz. Default is 3000.
4257 Set number of poles. Default is 2.
4260 Set method to specify band-width of filter.
4275 Specify the band-width of a filter in width_type units.
4276 Applies only to double-pole filter.
4277 The default is 0.707q and gives a Butterworth response.
4280 How much to use filtered signal in output. Default is 1.
4281 Range is between 0 and 1.
4284 Specify which channels to filter, by default all available are filtered.
4287 Normalize biquad coefficients, by default is disabled.
4288 Enabling it will normalize magnitude response at DC to 0dB.
4291 Set transform type of IIR filter.
4300 @subsection Commands
4302 This filter supports the following commands:
4305 Change highpass frequency.
4306 Syntax for the command is : "@var{frequency}"
4309 Change highpass width_type.
4310 Syntax for the command is : "@var{width_type}"
4313 Change highpass width.
4314 Syntax for the command is : "@var{width}"
4317 Change highpass mix.
4318 Syntax for the command is : "@var{mix}"
4323 Join multiple input streams into one multi-channel stream.
4325 It accepts the following parameters:
4329 The number of input streams. It defaults to 2.
4331 @item channel_layout
4332 The desired output channel layout. It defaults to stereo.
4335 Map channels from inputs to output. The argument is a '|'-separated list of
4336 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4337 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4338 can be either the name of the input channel (e.g. FL for front left) or its
4339 index in the specified input stream. @var{out_channel} is the name of the output
4343 The filter will attempt to guess the mappings when they are not specified
4344 explicitly. It does so by first trying to find an unused matching input channel
4345 and if that fails it picks the first unused input channel.
4347 Join 3 inputs (with properly set channel layouts):
4349 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4352 Build a 5.1 output from 6 single-channel streams:
4354 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4355 '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'
4361 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4363 To enable compilation of this filter you need to configure FFmpeg with
4364 @code{--enable-ladspa}.
4368 Specifies the name of LADSPA plugin library to load. If the environment
4369 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4370 each one of the directories specified by the colon separated list in
4371 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4372 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4373 @file{/usr/lib/ladspa/}.
4376 Specifies the plugin within the library. Some libraries contain only
4377 one plugin, but others contain many of them. If this is not set filter
4378 will list all available plugins within the specified library.
4381 Set the '|' separated list of controls which are zero or more floating point
4382 values that determine the behavior of the loaded plugin (for example delay,
4384 Controls need to be defined using the following syntax:
4385 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4386 @var{valuei} is the value set on the @var{i}-th control.
4387 Alternatively they can be also defined using the following syntax:
4388 @var{value0}|@var{value1}|@var{value2}|..., where
4389 @var{valuei} is the value set on the @var{i}-th control.
4390 If @option{controls} is set to @code{help}, all available controls and
4391 their valid ranges are printed.
4393 @item sample_rate, s
4394 Specify the sample rate, default to 44100. Only used if plugin have
4398 Set the number of samples per channel per each output frame, default
4399 is 1024. Only used if plugin have zero inputs.
4402 Set the minimum duration of the sourced audio. See
4403 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4404 for the accepted syntax.
4405 Note that the resulting duration may be greater than the specified duration,
4406 as the generated audio is always cut at the end of a complete frame.
4407 If not specified, or the expressed duration is negative, the audio is
4408 supposed to be generated forever.
4409 Only used if plugin have zero inputs.
4412 Enable latency compensation, by default is disabled.
4413 Only used if plugin have inputs.
4416 @subsection Examples
4420 List all available plugins within amp (LADSPA example plugin) library:
4426 List all available controls and their valid ranges for @code{vcf_notch}
4427 plugin from @code{VCF} library:
4429 ladspa=f=vcf:p=vcf_notch:c=help
4433 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4436 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4440 Add reverberation to the audio using TAP-plugins
4441 (Tom's Audio Processing plugins):
4443 ladspa=file=tap_reverb:tap_reverb
4447 Generate white noise, with 0.2 amplitude:
4449 ladspa=file=cmt:noise_source_white:c=c0=.2
4453 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4454 @code{C* Audio Plugin Suite} (CAPS) library:
4456 ladspa=file=caps:Click:c=c1=20'
4460 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4462 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4466 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4467 @code{SWH Plugins} collection:
4469 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4473 Attenuate low frequencies using Multiband EQ from Steve Harris
4474 @code{SWH Plugins} collection:
4476 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4480 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4483 ladspa=caps:Narrower
4487 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4489 ladspa=caps:White:.2
4493 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4495 ladspa=caps:Fractal:c=c1=1
4499 Dynamic volume normalization using @code{VLevel} plugin:
4501 ladspa=vlevel-ladspa:vlevel_mono
4505 @subsection Commands
4507 This filter supports the following commands:
4510 Modify the @var{N}-th control value.
4512 If the specified value is not valid, it is ignored and prior one is kept.
4517 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4518 Support for both single pass (livestreams, files) and double pass (files) modes.
4519 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4520 detect true peaks, the audio stream will be upsampled to 192 kHz.
4521 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4523 The filter accepts the following options:
4527 Set integrated loudness target.
4528 Range is -70.0 - -5.0. Default value is -24.0.
4531 Set loudness range target.
4532 Range is 1.0 - 20.0. Default value is 7.0.
4535 Set maximum true peak.
4536 Range is -9.0 - +0.0. Default value is -2.0.
4538 @item measured_I, measured_i
4539 Measured IL of input file.
4540 Range is -99.0 - +0.0.
4542 @item measured_LRA, measured_lra
4543 Measured LRA of input file.
4544 Range is 0.0 - 99.0.
4546 @item measured_TP, measured_tp
4547 Measured true peak of input file.
4548 Range is -99.0 - +99.0.
4550 @item measured_thresh
4551 Measured threshold of input file.
4552 Range is -99.0 - +0.0.
4555 Set offset gain. Gain is applied before the true-peak limiter.
4556 Range is -99.0 - +99.0. Default is +0.0.
4559 Normalize by linearly scaling the source audio.
4560 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4561 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4562 be lower than source LRA and the change in integrated loudness shouldn't
4563 result in a true peak which exceeds the target TP. If any of these
4564 conditions aren't met, normalization mode will revert to @var{dynamic}.
4565 Options are @code{true} or @code{false}. Default is @code{true}.
4568 Treat mono input files as "dual-mono". If a mono file is intended for playback
4569 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4570 If set to @code{true}, this option will compensate for this effect.
4571 Multi-channel input files are not affected by this option.
4572 Options are true or false. Default is false.
4575 Set print format for stats. Options are summary, json, or none.
4576 Default value is none.
4581 Apply a low-pass filter with 3dB point frequency.
4582 The filter can be either single-pole or double-pole (the default).
4583 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4585 The filter accepts the following options:
4589 Set frequency in Hz. Default is 500.
4592 Set number of poles. Default is 2.
4595 Set method to specify band-width of filter.
4610 Specify the band-width of a filter in width_type units.
4611 Applies only to double-pole filter.
4612 The default is 0.707q and gives a Butterworth response.
4615 How much to use filtered signal in output. Default is 1.
4616 Range is between 0 and 1.
4619 Specify which channels to filter, by default all available are filtered.
4622 Normalize biquad coefficients, by default is disabled.
4623 Enabling it will normalize magnitude response at DC to 0dB.
4626 Set transform type of IIR filter.
4635 @subsection Examples
4638 Lowpass only LFE channel, it LFE is not present it does nothing:
4644 @subsection Commands
4646 This filter supports the following commands:
4649 Change lowpass frequency.
4650 Syntax for the command is : "@var{frequency}"
4653 Change lowpass width_type.
4654 Syntax for the command is : "@var{width_type}"
4657 Change lowpass width.
4658 Syntax for the command is : "@var{width}"
4662 Syntax for the command is : "@var{mix}"
4667 Load a LV2 (LADSPA Version 2) plugin.
4669 To enable compilation of this filter you need to configure FFmpeg with
4670 @code{--enable-lv2}.
4674 Specifies the plugin URI. You may need to escape ':'.
4677 Set the '|' separated list of controls which are zero or more floating point
4678 values that determine the behavior of the loaded plugin (for example delay,
4680 If @option{controls} is set to @code{help}, all available controls and
4681 their valid ranges are printed.
4683 @item sample_rate, s
4684 Specify the sample rate, default to 44100. Only used if plugin have
4688 Set the number of samples per channel per each output frame, default
4689 is 1024. Only used if plugin have zero inputs.
4692 Set the minimum duration of the sourced audio. See
4693 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4694 for the accepted syntax.
4695 Note that the resulting duration may be greater than the specified duration,
4696 as the generated audio is always cut at the end of a complete frame.
4697 If not specified, or the expressed duration is negative, the audio is
4698 supposed to be generated forever.
4699 Only used if plugin have zero inputs.
4702 @subsection Examples
4706 Apply bass enhancer plugin from Calf:
4708 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4712 Apply vinyl plugin from Calf:
4714 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4718 Apply bit crusher plugin from ArtyFX:
4720 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4725 Multiband Compress or expand the audio's dynamic range.
4727 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4728 This is akin to the crossover of a loudspeaker, and results in flat frequency
4729 response when absent compander action.
4731 It accepts the following parameters:
4735 This option syntax is:
4736 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4737 For explanation of each item refer to compand filter documentation.
4743 Mix channels with specific gain levels. The filter accepts the output
4744 channel layout followed by a set of channels definitions.
4746 This filter is also designed to efficiently remap the channels of an audio
4749 The filter accepts parameters of the form:
4750 "@var{l}|@var{outdef}|@var{outdef}|..."
4754 output channel layout or number of channels
4757 output channel specification, of the form:
4758 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4761 output channel to define, either a channel name (FL, FR, etc.) or a channel
4762 number (c0, c1, etc.)
4765 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4768 input channel to use, see out_name for details; it is not possible to mix
4769 named and numbered input channels
4772 If the `=' in a channel specification is replaced by `<', then the gains for
4773 that specification will be renormalized so that the total is 1, thus
4774 avoiding clipping noise.
4776 @subsection Mixing examples
4778 For example, if you want to down-mix from stereo to mono, but with a bigger
4779 factor for the left channel:
4781 pan=1c|c0=0.9*c0+0.1*c1
4784 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4785 7-channels surround:
4787 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4790 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4791 that should be preferred (see "-ac" option) unless you have very specific
4794 @subsection Remapping examples
4796 The channel remapping will be effective if, and only if:
4799 @item gain coefficients are zeroes or ones,
4800 @item only one input per channel output,
4803 If all these conditions are satisfied, the filter will notify the user ("Pure
4804 channel mapping detected"), and use an optimized and lossless method to do the
4807 For example, if you have a 5.1 source and want a stereo audio stream by
4808 dropping the extra channels:
4810 pan="stereo| c0=FL | c1=FR"
4813 Given the same source, you can also switch front left and front right channels
4814 and keep the input channel layout:
4816 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4819 If the input is a stereo audio stream, you can mute the front left channel (and
4820 still keep the stereo channel layout) with:
4825 Still with a stereo audio stream input, you can copy the right channel in both
4826 front left and right:
4828 pan="stereo| c0=FR | c1=FR"
4833 ReplayGain scanner filter. This filter takes an audio stream as an input and
4834 outputs it unchanged.
4835 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4839 Convert the audio sample format, sample rate and channel layout. It is
4840 not meant to be used directly.
4843 Apply time-stretching and pitch-shifting with librubberband.
4845 To enable compilation of this filter, you need to configure FFmpeg with
4846 @code{--enable-librubberband}.
4848 The filter accepts the following options:
4852 Set tempo scale factor.
4855 Set pitch scale factor.
4858 Set transients detector.
4859 Possible values are:
4868 Possible values are:
4877 Possible values are:
4884 Set processing window size.
4885 Possible values are:
4894 Possible values are:
4901 Enable formant preservation when shift pitching.
4902 Possible values are:
4910 Possible values are:
4919 Possible values are:
4926 @subsection Commands
4928 This filter supports the following commands:
4931 Change filter tempo scale factor.
4932 Syntax for the command is : "@var{tempo}"
4935 Change filter pitch scale factor.
4936 Syntax for the command is : "@var{pitch}"
4939 @section sidechaincompress
4941 This filter acts like normal compressor but has the ability to compress
4942 detected signal using second input signal.
4943 It needs two input streams and returns one output stream.
4944 First input stream will be processed depending on second stream signal.
4945 The filtered signal then can be filtered with other filters in later stages of
4946 processing. See @ref{pan} and @ref{amerge} filter.
4948 The filter accepts the following options:
4952 Set input gain. Default is 1. Range is between 0.015625 and 64.
4955 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4956 Default is @code{downward}.
4959 If a signal of second stream raises above this level it will affect the gain
4960 reduction of first stream.
4961 By default is 0.125. Range is between 0.00097563 and 1.
4964 Set a ratio about which the signal is reduced. 1:2 means that if the level
4965 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4966 Default is 2. Range is between 1 and 20.
4969 Amount of milliseconds the signal has to rise above the threshold before gain
4970 reduction starts. Default is 20. Range is between 0.01 and 2000.
4973 Amount of milliseconds the signal has to fall below the threshold before
4974 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4977 Set the amount by how much signal will be amplified after processing.
4978 Default is 1. Range is from 1 to 64.
4981 Curve the sharp knee around the threshold to enter gain reduction more softly.
4982 Default is 2.82843. Range is between 1 and 8.
4985 Choose if the @code{average} level between all channels of side-chain stream
4986 or the louder(@code{maximum}) channel of side-chain stream affects the
4987 reduction. Default is @code{average}.
4990 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4991 of @code{rms}. Default is @code{rms} which is mainly smoother.
4994 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4997 How much to use compressed signal in output. Default is 1.
4998 Range is between 0 and 1.
5001 @subsection Commands
5003 This filter supports the all above options as @ref{commands}.
5005 @subsection Examples
5009 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5010 depending on the signal of 2nd input and later compressed signal to be
5011 merged with 2nd input:
5013 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5017 @section sidechaingate
5019 A sidechain gate acts like a normal (wideband) gate but has the ability to
5020 filter the detected signal before sending it to the gain reduction stage.
5021 Normally a gate uses the full range signal to detect a level above the
5023 For example: If you cut all lower frequencies from your sidechain signal
5024 the gate will decrease the volume of your track only if not enough highs
5025 appear. With this technique you are able to reduce the resonation of a
5026 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5028 It needs two input streams and returns one output stream.
5029 First input stream will be processed depending on second stream signal.
5031 The filter accepts the following options:
5035 Set input level before filtering.
5036 Default is 1. Allowed range is from 0.015625 to 64.
5039 Set the mode of operation. Can be @code{upward} or @code{downward}.
5040 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5041 will be amplified, expanding dynamic range in upward direction.
5042 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5045 Set the level of gain reduction when the signal is below the threshold.
5046 Default is 0.06125. Allowed range is from 0 to 1.
5047 Setting this to 0 disables reduction and then filter behaves like expander.
5050 If a signal rises above this level the gain reduction is released.
5051 Default is 0.125. Allowed range is from 0 to 1.
5054 Set a ratio about which the signal is reduced.
5055 Default is 2. Allowed range is from 1 to 9000.
5058 Amount of milliseconds the signal has to rise above the threshold before gain
5060 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5063 Amount of milliseconds the signal has to fall below the threshold before the
5064 reduction is increased again. Default is 250 milliseconds.
5065 Allowed range is from 0.01 to 9000.
5068 Set amount of amplification of signal after processing.
5069 Default is 1. Allowed range is from 1 to 64.
5072 Curve the sharp knee around the threshold to enter gain reduction more softly.
5073 Default is 2.828427125. Allowed range is from 1 to 8.
5076 Choose if exact signal should be taken for detection or an RMS like one.
5077 Default is rms. Can be peak or rms.
5080 Choose if the average level between all channels or the louder channel affects
5082 Default is average. Can be average or maximum.
5085 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5088 @subsection Commands
5090 This filter supports the all above options as @ref{commands}.
5092 @section silencedetect
5094 Detect silence in an audio stream.
5096 This filter logs a message when it detects that the input audio volume is less
5097 or equal to a noise tolerance value for a duration greater or equal to the
5098 minimum detected noise duration.
5100 The printed times and duration are expressed in seconds. The
5101 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5102 is set on the first frame whose timestamp equals or exceeds the detection
5103 duration and it contains the timestamp of the first frame of the silence.
5105 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5106 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5107 keys are set on the first frame after the silence. If @option{mono} is
5108 enabled, and each channel is evaluated separately, the @code{.X}
5109 suffixed keys are used, and @code{X} corresponds to the channel number.
5111 The filter accepts the following options:
5115 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5116 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5119 Set silence duration until notification (default is 2 seconds). See
5120 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5121 for the accepted syntax.
5124 Process each channel separately, instead of combined. By default is disabled.
5127 @subsection Examples
5131 Detect 5 seconds of silence with -50dB noise tolerance:
5133 silencedetect=n=-50dB:d=5
5137 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5138 tolerance in @file{silence.mp3}:
5140 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5144 @section silenceremove
5146 Remove silence from the beginning, middle or end of the audio.
5148 The filter accepts the following options:
5152 This value is used to indicate if audio should be trimmed at beginning of
5153 the audio. A value of zero indicates no silence should be trimmed from the
5154 beginning. When specifying a non-zero value, it trims audio up until it
5155 finds non-silence. Normally, when trimming silence from beginning of audio
5156 the @var{start_periods} will be @code{1} but it can be increased to higher
5157 values to trim all audio up to specific count of non-silence periods.
5158 Default value is @code{0}.
5160 @item start_duration
5161 Specify the amount of time that non-silence must be detected before it stops
5162 trimming audio. By increasing the duration, bursts of noises can be treated
5163 as silence and trimmed off. Default value is @code{0}.
5165 @item start_threshold
5166 This indicates what sample value should be treated as silence. For digital
5167 audio, a value of @code{0} may be fine but for audio recorded from analog,
5168 you may wish to increase the value to account for background noise.
5169 Can be specified in dB (in case "dB" is appended to the specified value)
5170 or amplitude ratio. Default value is @code{0}.
5173 Specify max duration of silence at beginning that will be kept after
5174 trimming. Default is 0, which is equal to trimming all samples detected
5178 Specify mode of detection of silence end in start of multi-channel audio.
5179 Can be @var{any} or @var{all}. Default is @var{any}.
5180 With @var{any}, any sample that is detected as non-silence will cause
5181 stopped trimming of silence.
5182 With @var{all}, only if all channels are detected as non-silence will cause
5183 stopped trimming of silence.
5186 Set the count for trimming silence from the end of audio.
5187 To remove silence from the middle of a file, specify a @var{stop_periods}
5188 that is negative. This value is then treated as a positive value and is
5189 used to indicate the effect should restart processing as specified by
5190 @var{start_periods}, making it suitable for removing periods of silence
5191 in the middle of the audio.
5192 Default value is @code{0}.
5195 Specify a duration of silence that must exist before audio is not copied any
5196 more. By specifying a higher duration, silence that is wanted can be left in
5198 Default value is @code{0}.
5200 @item stop_threshold
5201 This is the same as @option{start_threshold} but for trimming silence from
5203 Can be specified in dB (in case "dB" is appended to the specified value)
5204 or amplitude ratio. Default value is @code{0}.
5207 Specify max duration of silence at end that will be kept after
5208 trimming. Default is 0, which is equal to trimming all samples detected
5212 Specify mode of detection of silence start in end of multi-channel audio.
5213 Can be @var{any} or @var{all}. Default is @var{any}.
5214 With @var{any}, any sample that is detected as non-silence will cause
5215 stopped trimming of silence.
5216 With @var{all}, only if all channels are detected as non-silence will cause
5217 stopped trimming of silence.
5220 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5221 and works better with digital silence which is exactly 0.
5222 Default value is @code{rms}.
5225 Set duration in number of seconds used to calculate size of window in number
5226 of samples for detecting silence.
5227 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5230 @subsection Examples
5234 The following example shows how this filter can be used to start a recording
5235 that does not contain the delay at the start which usually occurs between
5236 pressing the record button and the start of the performance:
5238 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5242 Trim all silence encountered from beginning to end where there is more than 1
5243 second of silence in audio:
5245 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5249 Trim all digital silence samples, using peak detection, from beginning to end
5250 where there is more than 0 samples of digital silence in audio and digital
5251 silence is detected in all channels at same positions in stream:
5253 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5259 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5260 loudspeakers around the user for binaural listening via headphones (audio
5261 formats up to 9 channels supported).
5262 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5263 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5264 Austrian Academy of Sciences.
5266 To enable compilation of this filter you need to configure FFmpeg with
5267 @code{--enable-libmysofa}.
5269 The filter accepts the following options:
5273 Set the SOFA file used for rendering.
5276 Set gain applied to audio. Value is in dB. Default is 0.
5279 Set rotation of virtual loudspeakers in deg. Default is 0.
5282 Set elevation of virtual speakers in deg. Default is 0.
5285 Set distance in meters between loudspeakers and the listener with near-field
5286 HRTFs. Default is 1.
5289 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5290 processing audio in time domain which is slow.
5291 @var{freq} is processing audio in frequency domain which is fast.
5292 Default is @var{freq}.
5295 Set custom positions of virtual loudspeakers. Syntax for this option is:
5296 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5297 Each virtual loudspeaker is described with short channel name following with
5298 azimuth and elevation in degrees.
5299 Each virtual loudspeaker description is separated by '|'.
5300 For example to override front left and front right channel positions use:
5301 'speakers=FL 45 15|FR 345 15'.
5302 Descriptions with unrecognised channel names are ignored.
5305 Set custom gain for LFE channels. Value is in dB. Default is 0.
5308 Set custom frame size in number of samples. Default is 1024.
5309 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5310 is set to @var{freq}.
5313 Should all IRs be normalized upon importing SOFA file.
5314 By default is enabled.
5317 Should nearest IRs be interpolated with neighbor IRs if exact position
5318 does not match. By default is disabled.
5321 Minphase all IRs upon loading of SOFA file. By default is disabled.
5324 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5327 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5330 @subsection Examples
5334 Using ClubFritz6 sofa file:
5336 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5340 Using ClubFritz12 sofa file and bigger radius with small rotation:
5342 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5346 Similar as above but with custom speaker positions for front left, front right, back left and back right
5347 and also with custom gain:
5349 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5356 This filter expands or compresses each half-cycle of audio samples
5357 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5358 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5360 The filter accepts the following options:
5364 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5365 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5368 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5369 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5370 would be such that local peak value reaches target peak value but never to surpass it and that
5371 ratio between new and previous peak value does not surpass this option value.
5373 @item compression, c
5374 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5375 This option controls maximum local half-cycle of samples compression. This option is used
5376 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5377 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5378 that peak's half-cycle will be compressed by current compression factor.
5381 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5382 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5383 Any half-cycle samples with their local peak value below or same as this option value will be
5384 compressed by current compression factor, otherwise, if greater than threshold value they will be
5385 expanded with expansion factor so that it could reach peak target value but never surpass it.
5388 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5389 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5390 each new half-cycle until it reaches @option{expansion} value.
5391 Setting this options too high may lead to distortions.
5394 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5395 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5396 each new half-cycle until it reaches @option{compression} value.
5399 Specify which channels to filter, by default all available channels are filtered.
5402 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5403 option. When enabled any half-cycle of samples with their local peak value below or same as
5404 @option{threshold} option will be expanded otherwise it will be compressed.
5407 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5408 When disabled each filtered channel gain calculation is independent, otherwise when this option
5409 is enabled the minimum of all possible gains for each filtered channel is used.
5412 @subsection Commands
5414 This filter supports the all above options as @ref{commands}.
5416 @section stereotools
5418 This filter has some handy utilities to manage stereo signals, for converting
5419 M/S stereo recordings to L/R signal while having control over the parameters
5420 or spreading the stereo image of master track.
5422 The filter accepts the following options:
5426 Set input level before filtering for both channels. Defaults is 1.
5427 Allowed range is from 0.015625 to 64.
5430 Set output level after filtering for both channels. Defaults is 1.
5431 Allowed range is from 0.015625 to 64.
5434 Set input balance between both channels. Default is 0.
5435 Allowed range is from -1 to 1.
5438 Set output balance between both channels. Default is 0.
5439 Allowed range is from -1 to 1.
5442 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5443 clipping. Disabled by default.
5446 Mute the left channel. Disabled by default.
5449 Mute the right channel. Disabled by default.
5452 Change the phase of the left channel. Disabled by default.
5455 Change the phase of the right channel. Disabled by default.
5458 Set stereo mode. Available values are:
5462 Left/Right to Left/Right, this is default.
5465 Left/Right to Mid/Side.
5468 Mid/Side to Left/Right.
5471 Left/Right to Left/Left.
5474 Left/Right to Right/Right.
5477 Left/Right to Left + Right.
5480 Left/Right to Right/Left.
5483 Mid/Side to Left/Left.
5486 Mid/Side to Right/Right.
5490 Set level of side signal. Default is 1.
5491 Allowed range is from 0.015625 to 64.
5494 Set balance of side signal. Default is 0.
5495 Allowed range is from -1 to 1.
5498 Set level of the middle signal. Default is 1.
5499 Allowed range is from 0.015625 to 64.
5502 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5505 Set stereo base between mono and inversed channels. Default is 0.
5506 Allowed range is from -1 to 1.
5509 Set delay in milliseconds how much to delay left from right channel and
5510 vice versa. Default is 0. Allowed range is from -20 to 20.
5513 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5516 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5518 @item bmode_in, bmode_out
5519 Set balance mode for balance_in/balance_out option.
5521 Can be one of the following:
5525 Classic balance mode. Attenuate one channel at time.
5526 Gain is raised up to 1.
5529 Similar as classic mode above but gain is raised up to 2.
5532 Equal power distribution, from -6dB to +6dB range.
5536 @subsection Examples
5540 Apply karaoke like effect:
5542 stereotools=mlev=0.015625
5546 Convert M/S signal to L/R:
5548 "stereotools=mode=ms>lr"
5552 @section stereowiden
5554 This filter enhance the stereo effect by suppressing signal common to both
5555 channels and by delaying the signal of left into right and vice versa,
5556 thereby widening the stereo effect.
5558 The filter accepts the following options:
5562 Time in milliseconds of the delay of left signal into right and vice versa.
5563 Default is 20 milliseconds.
5566 Amount of gain in delayed signal into right and vice versa. Gives a delay
5567 effect of left signal in right output and vice versa which gives widening
5568 effect. Default is 0.3.
5571 Cross feed of left into right with inverted phase. This helps in suppressing
5572 the mono. If the value is 1 it will cancel all the signal common to both
5573 channels. Default is 0.3.
5576 Set level of input signal of original channel. Default is 0.8.
5579 @subsection Commands
5581 This filter supports the all above options except @code{delay} as @ref{commands}.
5583 @section superequalizer
5584 Apply 18 band equalizer.
5586 The filter accepts the following options:
5593 Set 131Hz band gain.
5595 Set 185Hz band gain.
5597 Set 262Hz band gain.
5599 Set 370Hz band gain.
5601 Set 523Hz band gain.
5603 Set 740Hz band gain.
5605 Set 1047Hz band gain.
5607 Set 1480Hz band gain.
5609 Set 2093Hz band gain.
5611 Set 2960Hz band gain.
5613 Set 4186Hz band gain.
5615 Set 5920Hz band gain.
5617 Set 8372Hz band gain.
5619 Set 11840Hz band gain.
5621 Set 16744Hz band gain.
5623 Set 20000Hz band gain.
5627 Apply audio surround upmix filter.
5629 This filter allows to produce multichannel output from audio stream.
5631 The filter accepts the following options:
5635 Set output channel layout. By default, this is @var{5.1}.
5637 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5638 for the required syntax.
5641 Set input channel layout. By default, this is @var{stereo}.
5643 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5644 for the required syntax.
5647 Set input volume level. By default, this is @var{1}.
5650 Set output volume level. By default, this is @var{1}.
5653 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5656 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5659 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5662 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5663 In @var{add} mode, LFE channel is created from input audio and added to output.
5664 In @var{sub} mode, LFE channel is created from input audio and added to output but
5665 also all non-LFE output channels are subtracted with output LFE channel.
5668 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5669 Default is @var{90}.
5672 Set front center input volume. By default, this is @var{1}.
5675 Set front center output volume. By default, this is @var{1}.
5678 Set front left input volume. By default, this is @var{1}.
5681 Set front left output volume. By default, this is @var{1}.
5684 Set front right input volume. By default, this is @var{1}.
5687 Set front right output volume. By default, this is @var{1}.
5690 Set side left input volume. By default, this is @var{1}.
5693 Set side left output volume. By default, this is @var{1}.
5696 Set side right input volume. By default, this is @var{1}.
5699 Set side right output volume. By default, this is @var{1}.
5702 Set back left input volume. By default, this is @var{1}.
5705 Set back left output volume. By default, this is @var{1}.
5708 Set back right input volume. By default, this is @var{1}.
5711 Set back right output volume. By default, this is @var{1}.
5714 Set back center input volume. By default, this is @var{1}.
5717 Set back center output volume. By default, this is @var{1}.
5720 Set LFE input volume. By default, this is @var{1}.
5723 Set LFE output volume. By default, this is @var{1}.
5726 Set spread usage of stereo image across X axis for all channels.
5729 Set spread usage of stereo image across Y axis for all channels.
5731 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5732 Set spread usage of stereo image across X axis for each channel.
5734 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5735 Set spread usage of stereo image across Y axis for each channel.
5738 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5741 Set window function.
5743 It accepts the following values:
5766 Default is @code{hann}.
5769 Set window overlap. If set to 1, the recommended overlap for selected
5770 window function will be picked. Default is @code{0.5}.
5773 @section treble, highshelf
5775 Boost or cut treble (upper) frequencies of the audio using a two-pole
5776 shelving filter with a response similar to that of a standard
5777 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5779 The filter accepts the following options:
5783 Give the gain at whichever is the lower of ~22 kHz and the
5784 Nyquist frequency. Its useful range is about -20 (for a large cut)
5785 to +20 (for a large boost). Beware of clipping when using a positive gain.
5788 Set the filter's central frequency and so can be used
5789 to extend or reduce the frequency range to be boosted or cut.
5790 The default value is @code{3000} Hz.
5793 Set method to specify band-width of filter.
5808 Determine how steep is the filter's shelf transition.
5811 How much to use filtered signal in output. Default is 1.
5812 Range is between 0 and 1.
5815 Specify which channels to filter, by default all available are filtered.
5818 Normalize biquad coefficients, by default is disabled.
5819 Enabling it will normalize magnitude response at DC to 0dB.
5822 Set transform type of IIR filter.
5831 @subsection Commands
5833 This filter supports the following commands:
5836 Change treble frequency.
5837 Syntax for the command is : "@var{frequency}"
5840 Change treble width_type.
5841 Syntax for the command is : "@var{width_type}"
5844 Change treble width.
5845 Syntax for the command is : "@var{width}"
5849 Syntax for the command is : "@var{gain}"
5853 Syntax for the command is : "@var{mix}"
5858 Sinusoidal amplitude modulation.
5860 The filter accepts the following options:
5864 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5865 (20 Hz or lower) will result in a tremolo effect.
5866 This filter may also be used as a ring modulator by specifying
5867 a modulation frequency higher than 20 Hz.
5868 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5871 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5872 Default value is 0.5.
5877 Sinusoidal phase modulation.
5879 The filter accepts the following options:
5883 Modulation frequency in Hertz.
5884 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5887 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5888 Default value is 0.5.
5893 Adjust the input audio volume.
5895 It accepts the following parameters:
5899 Set audio volume expression.
5901 Output values are clipped to the maximum value.
5903 The output audio volume is given by the relation:
5905 @var{output_volume} = @var{volume} * @var{input_volume}
5908 The default value for @var{volume} is "1.0".
5911 This parameter represents the mathematical precision.
5913 It determines which input sample formats will be allowed, which affects the
5914 precision of the volume scaling.
5918 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5920 32-bit floating-point; this limits input sample format to FLT. (default)
5922 64-bit floating-point; this limits input sample format to DBL.
5926 Choose the behaviour on encountering ReplayGain side data in input frames.
5930 Remove ReplayGain side data, ignoring its contents (the default).
5933 Ignore ReplayGain side data, but leave it in the frame.
5936 Prefer the track gain, if present.
5939 Prefer the album gain, if present.
5942 @item replaygain_preamp
5943 Pre-amplification gain in dB to apply to the selected replaygain gain.
5945 Default value for @var{replaygain_preamp} is 0.0.
5947 @item replaygain_noclip
5948 Prevent clipping by limiting the gain applied.
5950 Default value for @var{replaygain_noclip} is 1.
5953 Set when the volume expression is evaluated.
5955 It accepts the following values:
5958 only evaluate expression once during the filter initialization, or
5959 when the @samp{volume} command is sent
5962 evaluate expression for each incoming frame
5965 Default value is @samp{once}.
5968 The volume expression can contain the following parameters.
5972 frame number (starting at zero)
5975 @item nb_consumed_samples
5976 number of samples consumed by the filter
5978 number of samples in the current frame
5980 original frame position in the file
5986 PTS at start of stream
5988 time at start of stream
5994 last set volume value
5997 Note that when @option{eval} is set to @samp{once} only the
5998 @var{sample_rate} and @var{tb} variables are available, all other
5999 variables will evaluate to NAN.
6001 @subsection Commands
6003 This filter supports the following commands:
6006 Modify the volume expression.
6007 The command accepts the same syntax of the corresponding option.
6009 If the specified expression is not valid, it is kept at its current
6013 @subsection Examples
6017 Halve the input audio volume:
6021 volume=volume=-6.0206dB
6024 In all the above example the named key for @option{volume} can be
6025 omitted, for example like in:
6031 Increase input audio power by 6 decibels using fixed-point precision:
6033 volume=volume=6dB:precision=fixed
6037 Fade volume after time 10 with an annihilation period of 5 seconds:
6039 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6043 @section volumedetect
6045 Detect the volume of the input video.
6047 The filter has no parameters. The input is not modified. Statistics about
6048 the volume will be printed in the log when the input stream end is reached.
6050 In particular it will show the mean volume (root mean square), maximum
6051 volume (on a per-sample basis), and the beginning of a histogram of the
6052 registered volume values (from the maximum value to a cumulated 1/1000 of
6055 All volumes are in decibels relative to the maximum PCM value.
6057 @subsection Examples
6059 Here is an excerpt of the output:
6061 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6062 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6063 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6064 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6065 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6066 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6067 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6068 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6069 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6075 The mean square energy is approximately -27 dB, or 10^-2.7.
6077 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6079 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6082 In other words, raising the volume by +4 dB does not cause any clipping,
6083 raising it by +5 dB causes clipping for 6 samples, etc.
6085 @c man end AUDIO FILTERS
6087 @chapter Audio Sources
6088 @c man begin AUDIO SOURCES
6090 Below is a description of the currently available audio sources.
6094 Buffer audio frames, and make them available to the filter chain.
6096 This source is mainly intended for a programmatic use, in particular
6097 through the interface defined in @file{libavfilter/buffersrc.h}.
6099 It accepts the following parameters:
6103 The timebase which will be used for timestamps of submitted frames. It must be
6104 either a floating-point number or in @var{numerator}/@var{denominator} form.
6107 The sample rate of the incoming audio buffers.
6110 The sample format of the incoming audio buffers.
6111 Either a sample format name or its corresponding integer representation from
6112 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6114 @item channel_layout
6115 The channel layout of the incoming audio buffers.
6116 Either a channel layout name from channel_layout_map in
6117 @file{libavutil/channel_layout.c} or its corresponding integer representation
6118 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6121 The number of channels of the incoming audio buffers.
6122 If both @var{channels} and @var{channel_layout} are specified, then they
6127 @subsection Examples
6130 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6133 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6134 Since the sample format with name "s16p" corresponds to the number
6135 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6138 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6143 Generate an audio signal specified by an expression.
6145 This source accepts in input one or more expressions (one for each
6146 channel), which are evaluated and used to generate a corresponding
6149 This source accepts the following options:
6153 Set the '|'-separated expressions list for each separate channel. In case the
6154 @option{channel_layout} option is not specified, the selected channel layout
6155 depends on the number of provided expressions. Otherwise the last
6156 specified expression is applied to the remaining output channels.
6158 @item channel_layout, c
6159 Set the channel layout. The number of channels in the specified layout
6160 must be equal to the number of specified expressions.
6163 Set the minimum duration of the sourced audio. See
6164 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6165 for the accepted syntax.
6166 Note that the resulting duration may be greater than the specified
6167 duration, as the generated audio is always cut at the end of a
6170 If not specified, or the expressed duration is negative, the audio is
6171 supposed to be generated forever.
6174 Set the number of samples per channel per each output frame,
6177 @item sample_rate, s
6178 Specify the sample rate, default to 44100.
6181 Each expression in @var{exprs} can contain the following constants:
6185 number of the evaluated sample, starting from 0
6188 time of the evaluated sample expressed in seconds, starting from 0
6195 @subsection Examples
6205 Generate a sin signal with frequency of 440 Hz, set sample rate to
6208 aevalsrc="sin(440*2*PI*t):s=8000"
6212 Generate a two channels signal, specify the channel layout (Front
6213 Center + Back Center) explicitly:
6215 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6219 Generate white noise:
6221 aevalsrc="-2+random(0)"
6225 Generate an amplitude modulated signal:
6227 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6231 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6233 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6240 Generate a FIR coefficients using frequency sampling method.
6242 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6244 The filter accepts the following options:
6248 Set number of filter coefficents in output audio stream.
6249 Default value is 1025.
6252 Set frequency points from where magnitude and phase are set.
6253 This must be in non decreasing order, and first element must be 0, while last element
6254 must be 1. Elements are separated by white spaces.
6257 Set magnitude value for every frequency point set by @option{frequency}.
6258 Number of values must be same as number of frequency points.
6259 Values are separated by white spaces.
6262 Set phase value for every frequency point set by @option{frequency}.
6263 Number of values must be same as number of frequency points.
6264 Values are separated by white spaces.
6266 @item sample_rate, r
6267 Set sample rate, default is 44100.
6270 Set number of samples per each frame. Default is 1024.
6273 Set window function. Default is blackman.
6278 The null audio source, return unprocessed audio frames. It is mainly useful
6279 as a template and to be employed in analysis / debugging tools, or as
6280 the source for filters which ignore the input data (for example the sox
6283 This source accepts the following options:
6287 @item channel_layout, cl
6289 Specifies the channel layout, and can be either an integer or a string
6290 representing a channel layout. The default value of @var{channel_layout}
6293 Check the channel_layout_map definition in
6294 @file{libavutil/channel_layout.c} for the mapping between strings and
6295 channel layout values.
6297 @item sample_rate, r
6298 Specifies the sample rate, and defaults to 44100.
6301 Set the number of samples per requested frames.
6304 Set the duration of the sourced audio. See
6305 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6306 for the accepted syntax.
6308 If not specified, or the expressed duration is negative, the audio is
6309 supposed to be generated forever.
6312 @subsection Examples
6316 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6318 anullsrc=r=48000:cl=4
6322 Do the same operation with a more obvious syntax:
6324 anullsrc=r=48000:cl=mono
6328 All the parameters need to be explicitly defined.
6332 Synthesize a voice utterance using the libflite library.
6334 To enable compilation of this filter you need to configure FFmpeg with
6335 @code{--enable-libflite}.
6337 Note that versions of the flite library prior to 2.0 are not thread-safe.
6339 The filter accepts the following options:
6344 If set to 1, list the names of the available voices and exit
6345 immediately. Default value is 0.
6348 Set the maximum number of samples per frame. Default value is 512.
6351 Set the filename containing the text to speak.
6354 Set the text to speak.
6357 Set the voice to use for the speech synthesis. Default value is
6358 @code{kal}. See also the @var{list_voices} option.
6361 @subsection Examples
6365 Read from file @file{speech.txt}, and synthesize the text using the
6366 standard flite voice:
6368 flite=textfile=speech.txt
6372 Read the specified text selecting the @code{slt} voice:
6374 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6378 Input text to ffmpeg:
6380 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6384 Make @file{ffplay} speak the specified text, using @code{flite} and
6385 the @code{lavfi} device:
6387 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6391 For more information about libflite, check:
6392 @url{http://www.festvox.org/flite/}
6396 Generate a noise audio signal.
6398 The filter accepts the following options:
6401 @item sample_rate, r
6402 Specify the sample rate. Default value is 48000 Hz.
6405 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6409 Specify the duration of the generated audio stream. Not specifying this option
6410 results in noise with an infinite length.
6412 @item color, colour, c
6413 Specify the color of noise. Available noise colors are white, pink, brown,
6414 blue, violet and velvet. Default color is white.
6417 Specify a value used to seed the PRNG.
6420 Set the number of samples per each output frame, default is 1024.
6423 @subsection Examples
6428 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6430 anoisesrc=d=60:c=pink:r=44100:a=0.5
6436 Generate odd-tap Hilbert transform FIR coefficients.
6438 The resulting stream can be used with @ref{afir} filter for phase-shifting
6439 the signal by 90 degrees.
6441 This is used in many matrix coding schemes and for analytic signal generation.
6442 The process is often written as a multiplication by i (or j), the imaginary unit.
6444 The filter accepts the following options:
6448 @item sample_rate, s
6449 Set sample rate, default is 44100.
6452 Set length of FIR filter, default is 22051.
6455 Set number of samples per each frame.
6458 Set window function to be used when generating FIR coefficients.
6463 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6465 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6467 The filter accepts the following options:
6470 @item sample_rate, r
6471 Set sample rate, default is 44100.
6474 Set number of samples per each frame. Default is 1024.
6477 Set high-pass frequency. Default is 0.
6480 Set low-pass frequency. Default is 0.
6481 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6482 is higher than 0 then filter will create band-pass filter coefficients,
6483 otherwise band-reject filter coefficients.
6486 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6489 Set Kaiser window beta.
6492 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6495 Enable rounding, by default is disabled.
6498 Set number of taps for high-pass filter.
6501 Set number of taps for low-pass filter.
6506 Generate an audio signal made of a sine wave with amplitude 1/8.
6508 The audio signal is bit-exact.
6510 The filter accepts the following options:
6515 Set the carrier frequency. Default is 440 Hz.
6517 @item beep_factor, b
6518 Enable a periodic beep every second with frequency @var{beep_factor} times
6519 the carrier frequency. Default is 0, meaning the beep is disabled.
6521 @item sample_rate, r
6522 Specify the sample rate, default is 44100.
6525 Specify the duration of the generated audio stream.
6527 @item samples_per_frame
6528 Set the number of samples per output frame.
6530 The expression can contain the following constants:
6534 The (sequential) number of the output audio frame, starting from 0.
6537 The PTS (Presentation TimeStamp) of the output audio frame,
6538 expressed in @var{TB} units.
6541 The PTS of the output audio frame, expressed in seconds.
6544 The timebase of the output audio frames.
6547 Default is @code{1024}.
6550 @subsection Examples
6555 Generate a simple 440 Hz sine wave:
6561 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6565 sine=frequency=220:beep_factor=4:duration=5
6569 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6572 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6576 @c man end AUDIO SOURCES
6578 @chapter Audio Sinks
6579 @c man begin AUDIO SINKS
6581 Below is a description of the currently available audio sinks.
6583 @section abuffersink
6585 Buffer audio frames, and make them available to the end of filter chain.
6587 This sink is mainly intended for programmatic use, in particular
6588 through the interface defined in @file{libavfilter/buffersink.h}
6589 or the options system.
6591 It accepts a pointer to an AVABufferSinkContext structure, which
6592 defines the incoming buffers' formats, to be passed as the opaque
6593 parameter to @code{avfilter_init_filter} for initialization.
6596 Null audio sink; do absolutely nothing with the input audio. It is
6597 mainly useful as a template and for use in analysis / debugging
6600 @c man end AUDIO SINKS
6602 @chapter Video Filters
6603 @c man begin VIDEO FILTERS
6605 When you configure your FFmpeg build, you can disable any of the
6606 existing filters using @code{--disable-filters}.
6607 The configure output will show the video filters included in your
6610 Below is a description of the currently available video filters.
6614 Mark a region of interest in a video frame.
6616 The frame data is passed through unchanged, but metadata is attached
6617 to the frame indicating regions of interest which can affect the
6618 behaviour of later encoding. Multiple regions can be marked by
6619 applying the filter multiple times.
6623 Region distance in pixels from the left edge of the frame.
6625 Region distance in pixels from the top edge of the frame.
6627 Region width in pixels.
6629 Region height in pixels.
6631 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6632 and may contain the following variables:
6635 Width of the input frame.
6637 Height of the input frame.
6641 Quantisation offset to apply within the region.
6643 This must be a real value in the range -1 to +1. A value of zero
6644 indicates no quality change. A negative value asks for better quality
6645 (less quantisation), while a positive value asks for worse quality
6646 (greater quantisation).
6648 The range is calibrated so that the extreme values indicate the
6649 largest possible offset - if the rest of the frame is encoded with the
6650 worst possible quality, an offset of -1 indicates that this region
6651 should be encoded with the best possible quality anyway. Intermediate
6652 values are then interpolated in some codec-dependent way.
6654 For example, in 10-bit H.264 the quantisation parameter varies between
6655 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6656 this region should be encoded with a QP around one-tenth of the full
6657 range better than the rest of the frame. So, if most of the frame
6658 were to be encoded with a QP of around 30, this region would get a QP
6659 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6660 An extreme value of -1 would indicate that this region should be
6661 encoded with the best possible quality regardless of the treatment of
6662 the rest of the frame - that is, should be encoded at a QP of -12.
6664 If set to true, remove any existing regions of interest marked on the
6665 frame before adding the new one.
6668 @subsection Examples
6672 Mark the centre quarter of the frame as interesting.
6674 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6677 Mark the 100-pixel-wide region on the left edge of the frame as very
6678 uninteresting (to be encoded at much lower quality than the rest of
6681 addroi=0:0:100:ih:+1/5
6685 @section alphaextract
6687 Extract the alpha component from the input as a grayscale video. This
6688 is especially useful with the @var{alphamerge} filter.
6692 Add or replace the alpha component of the primary input with the
6693 grayscale value of a second input. This is intended for use with
6694 @var{alphaextract} to allow the transmission or storage of frame
6695 sequences that have alpha in a format that doesn't support an alpha
6698 For example, to reconstruct full frames from a normal YUV-encoded video
6699 and a separate video created with @var{alphaextract}, you might use:
6701 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6706 Amplify differences between current pixel and pixels of adjacent frames in
6707 same pixel location.
6709 This filter accepts the following options:
6713 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6714 For example radius of 3 will instruct filter to calculate average of 7 frames.
6717 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6720 Set threshold for difference amplification. Any difference greater or equal to
6721 this value will not alter source pixel. Default is 10.
6722 Allowed range is from 0 to 65535.
6725 Set tolerance for difference amplification. Any difference lower to
6726 this value will not alter source pixel. Default is 0.
6727 Allowed range is from 0 to 65535.
6730 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6731 This option controls maximum possible value that will decrease source pixel value.
6734 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6735 This option controls maximum possible value that will increase source pixel value.
6738 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6741 @subsection Commands
6743 This filter supports the following @ref{commands} that corresponds to option of same name:
6755 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6756 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6757 Substation Alpha) subtitles files.
6759 This filter accepts the following option in addition to the common options from
6760 the @ref{subtitles} filter:
6764 Set the shaping engine
6766 Available values are:
6769 The default libass shaping engine, which is the best available.
6771 Fast, font-agnostic shaper that can do only substitutions
6773 Slower shaper using OpenType for substitutions and positioning
6776 The default is @code{auto}.
6780 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6782 The filter accepts the following options:
6786 Set threshold A for 1st plane. Default is 0.02.
6787 Valid range is 0 to 0.3.
6790 Set threshold B for 1st plane. Default is 0.04.
6791 Valid range is 0 to 5.
6794 Set threshold A for 2nd plane. Default is 0.02.
6795 Valid range is 0 to 0.3.
6798 Set threshold B for 2nd plane. Default is 0.04.
6799 Valid range is 0 to 5.
6802 Set threshold A for 3rd plane. Default is 0.02.
6803 Valid range is 0 to 0.3.
6806 Set threshold B for 3rd plane. Default is 0.04.
6807 Valid range is 0 to 5.
6809 Threshold A is designed to react on abrupt changes in the input signal and
6810 threshold B is designed to react on continuous changes in the input signal.
6813 Set number of frames filter will use for averaging. Default is 9. Must be odd
6814 number in range [5, 129].
6817 Set what planes of frame filter will use for averaging. Default is all.
6820 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
6821 Alternatively can be set to @code{s} serial.
6823 Parallel can be faster then serial, while other way around is never true.
6824 Parallel will abort early on first change being greater then thresholds, while serial
6825 will continue processing other side of frames if they are equal or below thresholds.
6828 @subsection Commands
6829 This filter supports same @ref{commands} as options except option @code{s}.
6830 The command accepts the same syntax of the corresponding option.
6834 Apply average blur filter.
6836 The filter accepts the following options:
6840 Set horizontal radius size.
6843 Set which planes to filter. By default all planes are filtered.
6846 Set vertical radius size, if zero it will be same as @code{sizeX}.
6847 Default is @code{0}.
6850 @subsection Commands
6851 This filter supports same commands as options.
6852 The command accepts the same syntax of the corresponding option.
6854 If the specified expression is not valid, it is kept at its current
6859 Compute the bounding box for the non-black pixels in the input frame
6862 This filter computes the bounding box containing all the pixels with a
6863 luminance value greater than the minimum allowed value.
6864 The parameters describing the bounding box are printed on the filter
6867 The filter accepts the following option:
6871 Set the minimal luminance value. Default is @code{16}.
6875 Apply bilateral filter, spatial smoothing while preserving edges.
6877 The filter accepts the following options:
6880 Set sigma of gaussian function to calculate spatial weight.
6881 Allowed range is 0 to 512. Default is 0.1.
6884 Set sigma of gaussian function to calculate range weight.
6885 Allowed range is 0 to 1. Default is 0.1.
6888 Set planes to filter. Default is first only.
6891 @section bitplanenoise
6893 Show and measure bit plane noise.
6895 The filter accepts the following options:
6899 Set which plane to analyze. Default is @code{1}.
6902 Filter out noisy pixels from @code{bitplane} set above.
6903 Default is disabled.
6906 @section blackdetect
6908 Detect video intervals that are (almost) completely black. Can be
6909 useful to detect chapter transitions, commercials, or invalid
6912 The filter outputs its detection analysis to both the log as well as
6913 frame metadata. If a black segment of at least the specified minimum
6914 duration is found, a line with the start and end timestamps as well
6915 as duration is printed to the log with level @code{info}. In addition,
6916 a log line with level @code{debug} is printed per frame showing the
6917 black amount detected for that frame.
6919 The filter also attaches metadata to the first frame of a black
6920 segment with key @code{lavfi.black_start} and to the first frame
6921 after the black segment ends with key @code{lavfi.black_end}. The
6922 value is the frame's timestamp. This metadata is added regardless
6923 of the minimum duration specified.
6925 The filter accepts the following options:
6928 @item black_min_duration, d
6929 Set the minimum detected black duration expressed in seconds. It must
6930 be a non-negative floating point number.
6932 Default value is 2.0.
6934 @item picture_black_ratio_th, pic_th
6935 Set the threshold for considering a picture "black".
6936 Express the minimum value for the ratio:
6938 @var{nb_black_pixels} / @var{nb_pixels}
6941 for which a picture is considered black.
6942 Default value is 0.98.
6944 @item pixel_black_th, pix_th
6945 Set the threshold for considering a pixel "black".
6947 The threshold expresses the maximum pixel luminance value for which a
6948 pixel is considered "black". The provided value is scaled according to
6949 the following equation:
6951 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6954 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6955 the input video format, the range is [0-255] for YUV full-range
6956 formats and [16-235] for YUV non full-range formats.
6958 Default value is 0.10.
6961 The following example sets the maximum pixel threshold to the minimum
6962 value, and detects only black intervals of 2 or more seconds:
6964 blackdetect=d=2:pix_th=0.00
6969 Detect frames that are (almost) completely black. Can be useful to
6970 detect chapter transitions or commercials. Output lines consist of
6971 the frame number of the detected frame, the percentage of blackness,
6972 the position in the file if known or -1 and the timestamp in seconds.
6974 In order to display the output lines, you need to set the loglevel at
6975 least to the AV_LOG_INFO value.
6977 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6978 The value represents the percentage of pixels in the picture that
6979 are below the threshold value.
6981 It accepts the following parameters:
6986 The percentage of the pixels that have to be below the threshold; it defaults to
6989 @item threshold, thresh
6990 The threshold below which a pixel value is considered black; it defaults to
6998 Blend two video frames into each other.
7000 The @code{blend} filter takes two input streams and outputs one
7001 stream, the first input is the "top" layer and second input is
7002 "bottom" layer. By default, the output terminates when the longest input terminates.
7004 The @code{tblend} (time blend) filter takes two consecutive frames
7005 from one single stream, and outputs the result obtained by blending
7006 the new frame on top of the old frame.
7008 A description of the accepted options follows.
7016 Set blend mode for specific pixel component or all pixel components in case
7017 of @var{all_mode}. Default value is @code{normal}.
7019 Available values for component modes are:
7061 Set blend opacity for specific pixel component or all pixel components in case
7062 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7069 Set blend expression for specific pixel component or all pixel components in case
7070 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7072 The expressions can use the following variables:
7076 The sequential number of the filtered frame, starting from @code{0}.
7080 the coordinates of the current sample
7084 the width and height of currently filtered plane
7088 Width and height scale for the plane being filtered. It is the
7089 ratio between the dimensions of the current plane to the luma plane,
7090 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7091 the luma plane and @code{0.5,0.5} for the chroma planes.
7094 Time of the current frame, expressed in seconds.
7097 Value of pixel component at current location for first video frame (top layer).
7100 Value of pixel component at current location for second video frame (bottom layer).
7104 The @code{blend} filter also supports the @ref{framesync} options.
7106 @subsection Examples
7110 Apply transition from bottom layer to top layer in first 10 seconds:
7112 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7116 Apply linear horizontal transition from top layer to bottom layer:
7118 blend=all_expr='A*(X/W)+B*(1-X/W)'
7122 Apply 1x1 checkerboard effect:
7124 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7128 Apply uncover left effect:
7130 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7134 Apply uncover down effect:
7136 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7140 Apply uncover up-left effect:
7142 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7146 Split diagonally video and shows top and bottom layer on each side:
7148 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7152 Display differences between the current and the previous frame:
7154 tblend=all_mode=grainextract
7160 Denoise frames using Block-Matching 3D algorithm.
7162 The filter accepts the following options.
7166 Set denoising strength. Default value is 1.
7167 Allowed range is from 0 to 999.9.
7168 The denoising algorithm is very sensitive to sigma, so adjust it
7169 according to the source.
7172 Set local patch size. This sets dimensions in 2D.
7175 Set sliding step for processing blocks. Default value is 4.
7176 Allowed range is from 1 to 64.
7177 Smaller values allows processing more reference blocks and is slower.
7180 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7181 When set to 1, no block matching is done. Larger values allows more blocks
7183 Allowed range is from 1 to 256.
7186 Set radius for search block matching. Default is 9.
7187 Allowed range is from 1 to INT32_MAX.
7190 Set step between two search locations for block matching. Default is 1.
7191 Allowed range is from 1 to 64. Smaller is slower.
7194 Set threshold of mean square error for block matching. Valid range is 0 to
7198 Set thresholding parameter for hard thresholding in 3D transformed domain.
7199 Larger values results in stronger hard-thresholding filtering in frequency
7203 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7204 Default is @code{basic}.
7207 If enabled, filter will use 2nd stream for block matching.
7208 Default is disabled for @code{basic} value of @var{estim} option,
7209 and always enabled if value of @var{estim} is @code{final}.
7212 Set planes to filter. Default is all available except alpha.
7215 @subsection Examples
7219 Basic filtering with bm3d:
7221 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7225 Same as above, but filtering only luma:
7227 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7231 Same as above, but with both estimation modes:
7233 split[a][b],[a]bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic[a],[b][a]bm3d=sigma=3:block=4:bstep=2:group=16:estim=final:ref=1
7237 Same as above, but prefilter with @ref{nlmeans} filter instead:
7239 split[a][b],[a]nlmeans=s=3:r=7:p=3[a],[b][a]bm3d=sigma=3:block=4:bstep=2:group=16:estim=final:ref=1
7245 Apply a boxblur algorithm to the input video.
7247 It accepts the following parameters:
7251 @item luma_radius, lr
7252 @item luma_power, lp
7253 @item chroma_radius, cr
7254 @item chroma_power, cp
7255 @item alpha_radius, ar
7256 @item alpha_power, ap
7260 A description of the accepted options follows.
7263 @item luma_radius, lr
7264 @item chroma_radius, cr
7265 @item alpha_radius, ar
7266 Set an expression for the box radius in pixels used for blurring the
7267 corresponding input plane.
7269 The radius value must be a non-negative number, and must not be
7270 greater than the value of the expression @code{min(w,h)/2} for the
7271 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7274 Default value for @option{luma_radius} is "2". If not specified,
7275 @option{chroma_radius} and @option{alpha_radius} default to the
7276 corresponding value set for @option{luma_radius}.
7278 The expressions can contain the following constants:
7282 The input width and height in pixels.
7286 The input chroma image width and height in pixels.
7290 The horizontal and vertical chroma subsample values. For example, for the
7291 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7294 @item luma_power, lp
7295 @item chroma_power, cp
7296 @item alpha_power, ap
7297 Specify how many times the boxblur filter is applied to the
7298 corresponding plane.
7300 Default value for @option{luma_power} is 2. If not specified,
7301 @option{chroma_power} and @option{alpha_power} default to the
7302 corresponding value set for @option{luma_power}.
7304 A value of 0 will disable the effect.
7307 @subsection Examples
7311 Apply a boxblur filter with the luma, chroma, and alpha radii
7314 boxblur=luma_radius=2:luma_power=1
7319 Set the luma radius to 2, and alpha and chroma radius to 0:
7321 boxblur=2:1:cr=0:ar=0
7325 Set the luma and chroma radii to a fraction of the video dimension:
7327 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7333 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7334 Deinterlacing Filter").
7336 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7337 interpolation algorithms.
7338 It accepts the following parameters:
7342 The interlacing mode to adopt. It accepts one of the following values:
7346 Output one frame for each frame.
7348 Output one frame for each field.
7351 The default value is @code{send_field}.
7354 The picture field parity assumed for the input interlaced video. It accepts one
7355 of the following values:
7359 Assume the top field is first.
7361 Assume the bottom field is first.
7363 Enable automatic detection of field parity.
7366 The default value is @code{auto}.
7367 If the interlacing is unknown or the decoder does not export this information,
7368 top field first will be assumed.
7371 Specify which frames to deinterlace. Accepts one of the following
7376 Deinterlace all frames.
7378 Only deinterlace frames marked as interlaced.
7381 The default value is @code{all}.
7386 Apply Contrast Adaptive Sharpen filter to video stream.
7388 The filter accepts the following options:
7392 Set the sharpening strength. Default value is 0.
7395 Set planes to filter. Default value is to filter all
7396 planes except alpha plane.
7400 Remove all color information for all colors except for certain one.
7402 The filter accepts the following options:
7406 The color which will not be replaced with neutral chroma.
7409 Similarity percentage with the above color.
7410 0.01 matches only the exact key color, while 1.0 matches everything.
7414 0.0 makes pixels either fully gray, or not gray at all.
7415 Higher values result in more preserved color.
7418 Signals that the color passed is already in YUV instead of RGB.
7420 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7421 This can be used to pass exact YUV values as hexadecimal numbers.
7424 @subsection Commands
7425 This filter supports same @ref{commands} as options.
7426 The command accepts the same syntax of the corresponding option.
7428 If the specified expression is not valid, it is kept at its current
7432 YUV colorspace color/chroma keying.
7434 The filter accepts the following options:
7438 The color which will be replaced with transparency.
7441 Similarity percentage with the key color.
7443 0.01 matches only the exact key color, while 1.0 matches everything.
7448 0.0 makes pixels either fully transparent, or not transparent at all.
7450 Higher values result in semi-transparent pixels, with a higher transparency
7451 the more similar the pixels color is to the key color.
7454 Signals that the color passed is already in YUV instead of RGB.
7456 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7457 This can be used to pass exact YUV values as hexadecimal numbers.
7460 @subsection Commands
7461 This filter supports same @ref{commands} as options.
7462 The command accepts the same syntax of the corresponding option.
7464 If the specified expression is not valid, it is kept at its current
7467 @subsection Examples
7471 Make every green pixel in the input image transparent:
7473 ffmpeg -i input.png -vf chromakey=green out.png
7477 Overlay a greenscreen-video on top of a static black background.
7479 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
7484 Reduce chrominance noise.
7486 The filter accepts the following options:
7490 Set threshold for averaging chrominance values.
7491 Sum of absolute difference of U and V pixel components or current
7492 pixel and neighbour pixels lower than this threshold will be used in
7493 averaging. Luma component is left unchanged and is copied to output.
7494 Default value is 30. Allowed range is from 1 to 200.
7497 Set horizontal radius of rectangle used for averaging.
7498 Allowed range is from 1 to 100. Default value is 5.
7501 Set vertical radius of rectangle used for averaging.
7502 Allowed range is from 1 to 100. Default value is 5.
7505 Set horizontal step when averaging. Default value is 1.
7506 Allowed range is from 1 to 50.
7507 Mostly useful to speed-up filtering.
7510 Set vertical step when averaging. Default value is 1.
7511 Allowed range is from 1 to 50.
7512 Mostly useful to speed-up filtering.
7515 @subsection Commands
7516 This filter supports same @ref{commands} as options.
7517 The command accepts the same syntax of the corresponding option.
7519 @section chromashift
7520 Shift chroma pixels horizontally and/or vertically.
7522 The filter accepts the following options:
7525 Set amount to shift chroma-blue horizontally.
7527 Set amount to shift chroma-blue vertically.
7529 Set amount to shift chroma-red horizontally.
7531 Set amount to shift chroma-red vertically.
7533 Set edge mode, can be @var{smear}, default, or @var{warp}.
7536 @subsection Commands
7538 This filter supports the all above options as @ref{commands}.
7542 Display CIE color diagram with pixels overlaid onto it.
7544 The filter accepts the following options:
7559 @item uhdtv, rec2020
7573 Set what gamuts to draw.
7575 See @code{system} option for available values.
7578 Set ciescope size, by default set to 512.
7581 Set intensity used to map input pixel values to CIE diagram.
7584 Set contrast used to draw tongue colors that are out of active color system gamut.
7587 Correct gamma displayed on scope, by default enabled.
7590 Show white point on CIE diagram, by default disabled.
7593 Set input gamma. Used only with XYZ input color space.
7598 Visualize information exported by some codecs.
7600 Some codecs can export information through frames using side-data or other
7601 means. For example, some MPEG based codecs export motion vectors through the
7602 @var{export_mvs} flag in the codec @option{flags2} option.
7604 The filter accepts the following option:
7608 Set motion vectors to visualize.
7610 Available flags for @var{mv} are:
7614 forward predicted MVs of P-frames
7616 forward predicted MVs of B-frames
7618 backward predicted MVs of B-frames
7622 Display quantization parameters using the chroma planes.
7625 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7627 Available flags for @var{mv_type} are:
7631 forward predicted MVs
7633 backward predicted MVs
7636 @item frame_type, ft
7637 Set frame type to visualize motion vectors of.
7639 Available flags for @var{frame_type} are:
7643 intra-coded frames (I-frames)
7645 predicted frames (P-frames)
7647 bi-directionally predicted frames (B-frames)
7651 @subsection Examples
7655 Visualize forward predicted MVs of all frames using @command{ffplay}:
7657 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7661 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7663 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7667 @section colorbalance
7668 Modify intensity of primary colors (red, green and blue) of input frames.
7670 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7671 regions for the red-cyan, green-magenta or blue-yellow balance.
7673 A positive adjustment value shifts the balance towards the primary color, a negative
7674 value towards the complementary color.
7676 The filter accepts the following options:
7682 Adjust red, green and blue shadows (darkest pixels).
7687 Adjust red, green and blue midtones (medium pixels).
7692 Adjust red, green and blue highlights (brightest pixels).
7694 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7697 Preserve lightness when changing color balance. Default is disabled.
7700 @subsection Examples
7704 Add red color cast to shadows:
7710 @subsection Commands
7712 This filter supports the all above options as @ref{commands}.
7714 @section colorchannelmixer
7716 Adjust video input frames by re-mixing color channels.
7718 This filter modifies a color channel by adding the values associated to
7719 the other channels of the same pixels. For example if the value to
7720 modify is red, the output value will be:
7722 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7725 The filter accepts the following options:
7732 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7733 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7739 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7740 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7746 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7747 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7753 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7754 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7756 Allowed ranges for options are @code{[-2.0, 2.0]}.
7759 @subsection Examples
7763 Convert source to grayscale:
7765 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7768 Simulate sepia tones:
7770 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7774 @subsection Commands
7776 This filter supports the all above options as @ref{commands}.
7779 RGB colorspace color keying.
7781 The filter accepts the following options:
7785 The color which will be replaced with transparency.
7788 Similarity percentage with the key color.
7790 0.01 matches only the exact key color, while 1.0 matches everything.
7795 0.0 makes pixels either fully transparent, or not transparent at all.
7797 Higher values result in semi-transparent pixels, with a higher transparency
7798 the more similar the pixels color is to the key color.
7801 @subsection Examples
7805 Make every green pixel in the input image transparent:
7807 ffmpeg -i input.png -vf colorkey=green out.png
7811 Overlay a greenscreen-video on top of a static background image.
7813 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
7817 @subsection Commands
7818 This filter supports same @ref{commands} as options.
7819 The command accepts the same syntax of the corresponding option.
7821 If the specified expression is not valid, it is kept at its current
7825 Remove all color information for all RGB colors except for certain one.
7827 The filter accepts the following options:
7831 The color which will not be replaced with neutral gray.
7834 Similarity percentage with the above color.
7835 0.01 matches only the exact key color, while 1.0 matches everything.
7838 Blend percentage. 0.0 makes pixels fully gray.
7839 Higher values result in more preserved color.
7842 @subsection Commands
7843 This filter supports same @ref{commands} as options.
7844 The command accepts the same syntax of the corresponding option.
7846 If the specified expression is not valid, it is kept at its current
7849 @section colorlevels
7851 Adjust video input frames using levels.
7853 The filter accepts the following options:
7860 Adjust red, green, blue and alpha input black point.
7861 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7867 Adjust red, green, blue and alpha input white point.
7868 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7870 Input levels are used to lighten highlights (bright tones), darken shadows
7871 (dark tones), change the balance of bright and dark tones.
7877 Adjust red, green, blue and alpha output black point.
7878 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7884 Adjust red, green, blue and alpha output white point.
7885 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7887 Output levels allows manual selection of a constrained output level range.
7890 @subsection Examples
7894 Make video output darker:
7896 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7902 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7906 Make video output lighter:
7908 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7912 Increase brightness:
7914 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7918 @subsection Commands
7920 This filter supports the all above options as @ref{commands}.
7922 @section colormatrix
7924 Convert color matrix.
7926 The filter accepts the following options:
7931 Specify the source and destination color matrix. Both values must be
7934 The accepted values are:
7962 For example to convert from BT.601 to SMPTE-240M, use the command:
7964 colormatrix=bt601:smpte240m
7969 Convert colorspace, transfer characteristics or color primaries.
7970 Input video needs to have an even size.
7972 The filter accepts the following options:
7977 Specify all color properties at once.
7979 The accepted values are:
8009 Specify output colorspace.
8011 The accepted values are:
8020 BT.470BG or BT.601-6 625
8023 SMPTE-170M or BT.601-6 525
8032 BT.2020 with non-constant luminance
8038 Specify output transfer characteristics.
8040 The accepted values are:
8052 Constant gamma of 2.2
8055 Constant gamma of 2.8
8058 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8076 BT.2020 for 10-bits content
8079 BT.2020 for 12-bits content
8085 Specify output color primaries.
8087 The accepted values are:
8096 BT.470BG or BT.601-6 625
8099 SMPTE-170M or BT.601-6 525
8123 Specify output color range.
8125 The accepted values are:
8128 TV (restricted) range
8131 MPEG (restricted) range
8142 Specify output color format.
8144 The accepted values are:
8147 YUV 4:2:0 planar 8-bits
8150 YUV 4:2:0 planar 10-bits
8153 YUV 4:2:0 planar 12-bits
8156 YUV 4:2:2 planar 8-bits
8159 YUV 4:2:2 planar 10-bits
8162 YUV 4:2:2 planar 12-bits
8165 YUV 4:4:4 planar 8-bits
8168 YUV 4:4:4 planar 10-bits
8171 YUV 4:4:4 planar 12-bits
8176 Do a fast conversion, which skips gamma/primary correction. This will take
8177 significantly less CPU, but will be mathematically incorrect. To get output
8178 compatible with that produced by the colormatrix filter, use fast=1.
8181 Specify dithering mode.
8183 The accepted values are:
8189 Floyd-Steinberg dithering
8193 Whitepoint adaptation mode.
8195 The accepted values are:
8198 Bradford whitepoint adaptation
8201 von Kries whitepoint adaptation
8204 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8208 Override all input properties at once. Same accepted values as @ref{all}.
8211 Override input colorspace. Same accepted values as @ref{space}.
8214 Override input color primaries. Same accepted values as @ref{primaries}.
8217 Override input transfer characteristics. Same accepted values as @ref{trc}.
8220 Override input color range. Same accepted values as @ref{range}.
8224 The filter converts the transfer characteristics, color space and color
8225 primaries to the specified user values. The output value, if not specified,
8226 is set to a default value based on the "all" property. If that property is
8227 also not specified, the filter will log an error. The output color range and
8228 format default to the same value as the input color range and format. The
8229 input transfer characteristics, color space, color primaries and color range
8230 should be set on the input data. If any of these are missing, the filter will
8231 log an error and no conversion will take place.
8233 For example to convert the input to SMPTE-240M, use the command:
8235 colorspace=smpte240m
8238 @section convolution
8240 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8242 The filter accepts the following options:
8249 Set matrix for each plane.
8250 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8251 and from 1 to 49 odd number of signed integers in @var{row} mode.
8257 Set multiplier for calculated value for each plane.
8258 If unset or 0, it will be sum of all matrix elements.
8264 Set bias for each plane. This value is added to the result of the multiplication.
8265 Useful for making the overall image brighter or darker. Default is 0.0.
8271 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8272 Default is @var{square}.
8275 @subsection Examples
8281 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"
8287 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"
8293 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"
8299 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"
8303 Apply laplacian edge detector which includes diagonals:
8305 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"
8311 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"
8317 Apply 2D convolution of video stream in frequency domain using second stream
8320 The filter accepts the following options:
8324 Set which planes to process.
8327 Set which impulse video frames will be processed, can be @var{first}
8328 or @var{all}. Default is @var{all}.
8331 The @code{convolve} filter also supports the @ref{framesync} options.
8335 Copy the input video source unchanged to the output. This is mainly useful for
8340 Video filtering on GPU using Apple's CoreImage API on OSX.
8342 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8343 processed by video hardware. However, software-based OpenGL implementations
8344 exist which means there is no guarantee for hardware processing. It depends on
8347 There are many filters and image generators provided by Apple that come with a
8348 large variety of options. The filter has to be referenced by its name along
8351 The coreimage filter accepts the following options:
8354 List all available filters and generators along with all their respective
8355 options as well as possible minimum and maximum values along with the default
8362 Specify all filters by their respective name and options.
8363 Use @var{list_filters} to determine all valid filter names and options.
8364 Numerical options are specified by a float value and are automatically clamped
8365 to their respective value range. Vector and color options have to be specified
8366 by a list of space separated float values. Character escaping has to be done.
8367 A special option name @code{default} is available to use default options for a
8370 It is required to specify either @code{default} or at least one of the filter options.
8371 All omitted options are used with their default values.
8372 The syntax of the filter string is as follows:
8374 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8378 Specify a rectangle where the output of the filter chain is copied into the
8379 input image. It is given by a list of space separated float values:
8381 output_rect=x\ y\ width\ height
8383 If not given, the output rectangle equals the dimensions of the input image.
8384 The output rectangle is automatically cropped at the borders of the input
8385 image. Negative values are valid for each component.
8387 output_rect=25\ 25\ 100\ 100
8391 Several filters can be chained for successive processing without GPU-HOST
8392 transfers allowing for fast processing of complex filter chains.
8393 Currently, only filters with zero (generators) or exactly one (filters) input
8394 image and one output image are supported. Also, transition filters are not yet
8397 Some filters generate output images with additional padding depending on the
8398 respective filter kernel. The padding is automatically removed to ensure the
8399 filter output has the same size as the input image.
8401 For image generators, the size of the output image is determined by the
8402 previous output image of the filter chain or the input image of the whole
8403 filterchain, respectively. The generators do not use the pixel information of
8404 this image to generate their output. However, the generated output is
8405 blended onto this image, resulting in partial or complete coverage of the
8408 The @ref{coreimagesrc} video source can be used for generating input images
8409 which are directly fed into the filter chain. By using it, providing input
8410 images by another video source or an input video is not required.
8412 @subsection Examples
8417 List all filters available:
8419 coreimage=list_filters=true
8423 Use the CIBoxBlur filter with default options to blur an image:
8425 coreimage=filter=CIBoxBlur@@default
8429 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8430 its center at 100x100 and a radius of 50 pixels:
8432 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8436 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8437 given as complete and escaped command-line for Apple's standard bash shell:
8439 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8445 Cover a rectangular object
8447 It accepts the following options:
8451 Filepath of the optional cover image, needs to be in yuv420.
8456 It accepts the following values:
8459 cover it by the supplied image
8461 cover it by interpolating the surrounding pixels
8464 Default value is @var{blur}.
8467 @subsection Examples
8471 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8473 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8479 Crop the input video to given dimensions.
8481 It accepts the following parameters:
8485 The width of the output video. It defaults to @code{iw}.
8486 This expression is evaluated only once during the filter
8487 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8490 The height of the output video. It defaults to @code{ih}.
8491 This expression is evaluated only once during the filter
8492 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8495 The horizontal position, in the input video, of the left edge of the output
8496 video. It defaults to @code{(in_w-out_w)/2}.
8497 This expression is evaluated per-frame.
8500 The vertical position, in the input video, of the top edge of the output video.
8501 It defaults to @code{(in_h-out_h)/2}.
8502 This expression is evaluated per-frame.
8505 If set to 1 will force the output display aspect ratio
8506 to be the same of the input, by changing the output sample aspect
8507 ratio. It defaults to 0.
8510 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8511 width/height/x/y as specified and will not be rounded to nearest smaller value.
8515 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8516 expressions containing the following constants:
8521 The computed values for @var{x} and @var{y}. They are evaluated for
8526 The input width and height.
8530 These are the same as @var{in_w} and @var{in_h}.
8534 The output (cropped) width and height.
8538 These are the same as @var{out_w} and @var{out_h}.
8541 same as @var{iw} / @var{ih}
8544 input sample aspect ratio
8547 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8551 horizontal and vertical chroma subsample values. For example for the
8552 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8555 The number of the input frame, starting from 0.
8558 the position in the file of the input frame, NAN if unknown
8561 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8565 The expression for @var{out_w} may depend on the value of @var{out_h},
8566 and the expression for @var{out_h} may depend on @var{out_w}, but they
8567 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8568 evaluated after @var{out_w} and @var{out_h}.
8570 The @var{x} and @var{y} parameters specify the expressions for the
8571 position of the top-left corner of the output (non-cropped) area. They
8572 are evaluated for each frame. If the evaluated value is not valid, it
8573 is approximated to the nearest valid value.
8575 The expression for @var{x} may depend on @var{y}, and the expression
8576 for @var{y} may depend on @var{x}.
8578 @subsection Examples
8582 Crop area with size 100x100 at position (12,34).
8587 Using named options, the example above becomes:
8589 crop=w=100:h=100:x=12:y=34
8593 Crop the central input area with size 100x100:
8599 Crop the central input area with size 2/3 of the input video:
8601 crop=2/3*in_w:2/3*in_h
8605 Crop the input video central square:
8612 Delimit the rectangle with the top-left corner placed at position
8613 100:100 and the right-bottom corner corresponding to the right-bottom
8614 corner of the input image.
8616 crop=in_w-100:in_h-100:100:100
8620 Crop 10 pixels from the left and right borders, and 20 pixels from
8621 the top and bottom borders
8623 crop=in_w-2*10:in_h-2*20
8627 Keep only the bottom right quarter of the input image:
8629 crop=in_w/2:in_h/2:in_w/2:in_h/2
8633 Crop height for getting Greek harmony:
8635 crop=in_w:1/PHI*in_w
8639 Apply trembling effect:
8641 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)
8645 Apply erratic camera effect depending on timestamp:
8647 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)"
8651 Set x depending on the value of y:
8653 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8657 @subsection Commands
8659 This filter supports the following commands:
8665 Set width/height of the output video and the horizontal/vertical position
8667 The command accepts the same syntax of the corresponding option.
8669 If the specified expression is not valid, it is kept at its current
8675 Auto-detect the crop size.
8677 It calculates the necessary cropping parameters and prints the
8678 recommended parameters via the logging system. The detected dimensions
8679 correspond to the non-black area of the input video.
8681 It accepts the following parameters:
8686 Set higher black value threshold, which can be optionally specified
8687 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8688 value greater to the set value is considered non-black. It defaults to 24.
8689 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8690 on the bitdepth of the pixel format.
8693 The value which the width/height should be divisible by. It defaults to
8694 16. The offset is automatically adjusted to center the video. Use 2 to
8695 get only even dimensions (needed for 4:2:2 video). 16 is best when
8696 encoding to most video codecs.
8698 @item reset_count, reset
8699 Set the counter that determines after how many frames cropdetect will
8700 reset the previously detected largest video area and start over to
8701 detect the current optimal crop area. Default value is 0.
8703 This can be useful when channel logos distort the video area. 0
8704 indicates 'never reset', and returns the largest area encountered during
8711 Delay video filtering until a given wallclock timestamp. The filter first
8712 passes on @option{preroll} amount of frames, then it buffers at most
8713 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8714 it forwards the buffered frames and also any subsequent frames coming in its
8717 The filter can be used synchronize the output of multiple ffmpeg processes for
8718 realtime output devices like decklink. By putting the delay in the filtering
8719 chain and pre-buffering frames the process can pass on data to output almost
8720 immediately after the target wallclock timestamp is reached.
8722 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8728 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8731 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8734 The maximum duration of content to buffer before waiting for the cue expressed
8735 in seconds. Default is 0.
8742 Apply color adjustments using curves.
8744 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8745 component (red, green and blue) has its values defined by @var{N} key points
8746 tied from each other using a smooth curve. The x-axis represents the pixel
8747 values from the input frame, and the y-axis the new pixel values to be set for
8750 By default, a component curve is defined by the two points @var{(0;0)} and
8751 @var{(1;1)}. This creates a straight line where each original pixel value is
8752 "adjusted" to its own value, which means no change to the image.
8754 The filter allows you to redefine these two points and add some more. A new
8755 curve (using a natural cubic spline interpolation) will be define to pass
8756 smoothly through all these new coordinates. The new defined points needs to be
8757 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8758 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8759 the vector spaces, the values will be clipped accordingly.
8761 The filter accepts the following options:
8765 Select one of the available color presets. This option can be used in addition
8766 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8767 options takes priority on the preset values.
8768 Available presets are:
8771 @item color_negative
8774 @item increase_contrast
8776 @item linear_contrast
8777 @item medium_contrast
8779 @item strong_contrast
8782 Default is @code{none}.
8784 Set the master key points. These points will define a second pass mapping. It
8785 is sometimes called a "luminance" or "value" mapping. It can be used with
8786 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8787 post-processing LUT.
8789 Set the key points for the red component.
8791 Set the key points for the green component.
8793 Set the key points for the blue component.
8795 Set the key points for all components (not including master).
8796 Can be used in addition to the other key points component
8797 options. In this case, the unset component(s) will fallback on this
8798 @option{all} setting.
8800 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
8802 Save Gnuplot script of the curves in specified file.
8805 To avoid some filtergraph syntax conflicts, each key points list need to be
8806 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
8808 @subsection Examples
8812 Increase slightly the middle level of blue:
8814 curves=blue='0/0 0.5/0.58 1/1'
8820 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'
8822 Here we obtain the following coordinates for each components:
8825 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8827 @code{(0;0) (0.50;0.48) (1;1)}
8829 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8833 The previous example can also be achieved with the associated built-in preset:
8835 curves=preset=vintage
8845 Use a Photoshop preset and redefine the points of the green component:
8847 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8851 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8852 and @command{gnuplot}:
8854 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8855 gnuplot -p /tmp/curves.plt
8861 Video data analysis filter.
8863 This filter shows hexadecimal pixel values of part of video.
8865 The filter accepts the following options:
8869 Set output video size.
8872 Set x offset from where to pick pixels.
8875 Set y offset from where to pick pixels.
8878 Set scope mode, can be one of the following:
8881 Draw hexadecimal pixel values with white color on black background.
8884 Draw hexadecimal pixel values with input video pixel color on black
8888 Draw hexadecimal pixel values on color background picked from input video,
8889 the text color is picked in such way so its always visible.
8893 Draw rows and columns numbers on left and top of video.
8896 Set background opacity.
8899 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
8903 Apply Directional blur filter.
8905 The filter accepts the following options:
8909 Set angle of directional blur. Default is @code{45}.
8912 Set radius of directional blur. Default is @code{5}.
8915 Set which planes to filter. By default all planes are filtered.
8918 @subsection Commands
8919 This filter supports same @ref{commands} as options.
8920 The command accepts the same syntax of the corresponding option.
8922 If the specified expression is not valid, it is kept at its current
8927 Denoise frames using 2D DCT (frequency domain filtering).
8929 This filter is not designed for real time.
8931 The filter accepts the following options:
8935 Set the noise sigma constant.
8937 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8938 coefficient (absolute value) below this threshold with be dropped.
8940 If you need a more advanced filtering, see @option{expr}.
8942 Default is @code{0}.
8945 Set number overlapping pixels for each block. Since the filter can be slow, you
8946 may want to reduce this value, at the cost of a less effective filter and the
8947 risk of various artefacts.
8949 If the overlapping value doesn't permit processing the whole input width or
8950 height, a warning will be displayed and according borders won't be denoised.
8952 Default value is @var{blocksize}-1, which is the best possible setting.
8955 Set the coefficient factor expression.
8957 For each coefficient of a DCT block, this expression will be evaluated as a
8958 multiplier value for the coefficient.
8960 If this is option is set, the @option{sigma} option will be ignored.
8962 The absolute value of the coefficient can be accessed through the @var{c}
8966 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8967 @var{blocksize}, which is the width and height of the processed blocks.
8969 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8970 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8971 on the speed processing. Also, a larger block size does not necessarily means a
8975 @subsection Examples
8977 Apply a denoise with a @option{sigma} of @code{4.5}:
8982 The same operation can be achieved using the expression system:
8984 dctdnoiz=e='gte(c, 4.5*3)'
8987 Violent denoise using a block size of @code{16x16}:
8994 Remove banding artifacts from input video.
8995 It works by replacing banded pixels with average value of referenced pixels.
8997 The filter accepts the following options:
9004 Set banding detection threshold for each plane. Default is 0.02.
9005 Valid range is 0.00003 to 0.5.
9006 If difference between current pixel and reference pixel is less than threshold,
9007 it will be considered as banded.
9010 Banding detection range in pixels. Default is 16. If positive, random number
9011 in range 0 to set value will be used. If negative, exact absolute value
9013 The range defines square of four pixels around current pixel.
9016 Set direction in radians from which four pixel will be compared. If positive,
9017 random direction from 0 to set direction will be picked. If negative, exact of
9018 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9019 will pick only pixels on same row and -PI/2 will pick only pixels on same
9023 If enabled, current pixel is compared with average value of all four
9024 surrounding pixels. The default is enabled. If disabled current pixel is
9025 compared with all four surrounding pixels. The pixel is considered banded
9026 if only all four differences with surrounding pixels are less than threshold.
9029 If enabled, current pixel is changed if and only if all pixel components are banded,
9030 e.g. banding detection threshold is triggered for all color components.
9031 The default is disabled.
9036 Remove blocking artifacts from input video.
9038 The filter accepts the following options:
9042 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9043 This controls what kind of deblocking is applied.
9046 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9052 Set blocking detection thresholds. Allowed range is 0 to 1.
9053 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9054 Using higher threshold gives more deblocking strength.
9055 Setting @var{alpha} controls threshold detection at exact edge of block.
9056 Remaining options controls threshold detection near the edge. Each one for
9057 below/above or left/right. Setting any of those to @var{0} disables
9061 Set planes to filter. Default is to filter all available planes.
9064 @subsection Examples
9068 Deblock using weak filter and block size of 4 pixels.
9070 deblock=filter=weak:block=4
9074 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9075 deblocking more edges.
9077 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9081 Similar as above, but filter only first plane.
9083 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9087 Similar as above, but filter only second and third plane.
9089 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9096 Drop duplicated frames at regular intervals.
9098 The filter accepts the following options:
9102 Set the number of frames from which one will be dropped. Setting this to
9103 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9104 Default is @code{5}.
9107 Set the threshold for duplicate detection. If the difference metric for a frame
9108 is less than or equal to this value, then it is declared as duplicate. Default
9112 Set scene change threshold. Default is @code{15}.
9116 Set the size of the x and y-axis blocks used during metric calculations.
9117 Larger blocks give better noise suppression, but also give worse detection of
9118 small movements. Must be a power of two. Default is @code{32}.
9121 Mark main input as a pre-processed input and activate clean source input
9122 stream. This allows the input to be pre-processed with various filters to help
9123 the metrics calculation while keeping the frame selection lossless. When set to
9124 @code{1}, the first stream is for the pre-processed input, and the second
9125 stream is the clean source from where the kept frames are chosen. Default is
9129 Set whether or not chroma is considered in the metric calculations. Default is
9135 Apply 2D deconvolution of video stream in frequency domain using second stream
9138 The filter accepts the following options:
9142 Set which planes to process.
9145 Set which impulse video frames will be processed, can be @var{first}
9146 or @var{all}. Default is @var{all}.
9149 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9150 and height are not same and not power of 2 or if stream prior to convolving
9154 The @code{deconvolve} filter also supports the @ref{framesync} options.
9158 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9160 It accepts the following options:
9164 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9165 @var{rainbows} for cross-color reduction.
9168 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9171 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9174 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9177 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9182 Apply deflate effect to the video.
9184 This filter replaces the pixel by the local(3x3) average by taking into account
9185 only values lower than the pixel.
9187 It accepts the following options:
9194 Limit the maximum change for each plane, default is 65535.
9195 If 0, plane will remain unchanged.
9198 @subsection Commands
9200 This filter supports the all above options as @ref{commands}.
9204 Remove temporal frame luminance variations.
9206 It accepts the following options:
9210 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9213 Set averaging mode to smooth temporal luminance variations.
9215 Available values are:
9240 Do not actually modify frame. Useful when one only wants metadata.
9245 Remove judder produced by partially interlaced telecined content.
9247 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9248 source was partially telecined content then the output of @code{pullup,dejudder}
9249 will have a variable frame rate. May change the recorded frame rate of the
9250 container. Aside from that change, this filter will not affect constant frame
9253 The option available in this filter is:
9257 Specify the length of the window over which the judder repeats.
9259 Accepts any integer greater than 1. Useful values are:
9263 If the original was telecined from 24 to 30 fps (Film to NTSC).
9266 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9269 If a mixture of the two.
9272 The default is @samp{4}.
9277 Suppress a TV station logo by a simple interpolation of the surrounding
9278 pixels. Just set a rectangle covering the logo and watch it disappear
9279 (and sometimes something even uglier appear - your mileage may vary).
9281 It accepts the following parameters:
9286 Specify the top left corner coordinates of the logo. They must be
9291 Specify the width and height of the logo to clear. They must be
9295 Specify the thickness of the fuzzy edge of the rectangle (added to
9296 @var{w} and @var{h}). The default value is 1. This option is
9297 deprecated, setting higher values should no longer be necessary and
9301 When set to 1, a green rectangle is drawn on the screen to simplify
9302 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9303 The default value is 0.
9305 The rectangle is drawn on the outermost pixels which will be (partly)
9306 replaced with interpolated values. The values of the next pixels
9307 immediately outside this rectangle in each direction will be used to
9308 compute the interpolated pixel values inside the rectangle.
9312 @subsection Examples
9316 Set a rectangle covering the area with top left corner coordinates 0,0
9317 and size 100x77, and a band of size 10:
9319 delogo=x=0:y=0:w=100:h=77:band=10
9327 Remove the rain in the input image/video by applying the derain methods based on
9328 convolutional neural networks. Supported models:
9332 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9333 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9336 Training as well as model generation scripts are provided in
9337 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9339 Native model files (.model) can be generated from TensorFlow model
9340 files (.pb) by using tools/python/convert.py
9342 The filter accepts the following options:
9346 Specify which filter to use. This option accepts the following values:
9350 Derain filter. To conduct derain filter, you need to use a derain model.
9353 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9355 Default value is @samp{derain}.
9358 Specify which DNN backend to use for model loading and execution. This option accepts
9359 the following values:
9363 Native implementation of DNN loading and execution.
9366 TensorFlow backend. To enable this backend you
9367 need to install the TensorFlow for C library (see
9368 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9369 @code{--enable-libtensorflow}
9371 Default value is @samp{native}.
9374 Set path to model file specifying network architecture and its parameters.
9375 Note that different backends use different file formats. TensorFlow and native
9376 backend can load files for only its format.
9379 It can also be finished with @ref{dnn_processing} filter.
9383 Attempt to fix small changes in horizontal and/or vertical shift. This
9384 filter helps remove camera shake from hand-holding a camera, bumping a
9385 tripod, moving on a vehicle, etc.
9387 The filter accepts the following options:
9395 Specify a rectangular area where to limit the search for motion
9397 If desired the search for motion vectors can be limited to a
9398 rectangular area of the frame defined by its top left corner, width
9399 and height. These parameters have the same meaning as the drawbox
9400 filter which can be used to visualise the position of the bounding
9403 This is useful when simultaneous movement of subjects within the frame
9404 might be confused for camera motion by the motion vector search.
9406 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9407 then the full frame is used. This allows later options to be set
9408 without specifying the bounding box for the motion vector search.
9410 Default - search the whole frame.
9414 Specify the maximum extent of movement in x and y directions in the
9415 range 0-64 pixels. Default 16.
9418 Specify how to generate pixels to fill blanks at the edge of the
9419 frame. Available values are:
9422 Fill zeroes at blank locations
9424 Original image at blank locations
9426 Extruded edge value at blank locations
9428 Mirrored edge at blank locations
9430 Default value is @samp{mirror}.
9433 Specify the blocksize to use for motion search. Range 4-128 pixels,
9437 Specify the contrast threshold for blocks. Only blocks with more than
9438 the specified contrast (difference between darkest and lightest
9439 pixels) will be considered. Range 1-255, default 125.
9442 Specify the search strategy. Available values are:
9445 Set exhaustive search
9447 Set less exhaustive search.
9449 Default value is @samp{exhaustive}.
9452 If set then a detailed log of the motion search is written to the
9459 Remove unwanted contamination of foreground colors, caused by reflected color of
9460 greenscreen or bluescreen.
9462 This filter accepts the following options:
9466 Set what type of despill to use.
9469 Set how spillmap will be generated.
9472 Set how much to get rid of still remaining spill.
9475 Controls amount of red in spill area.
9478 Controls amount of green in spill area.
9479 Should be -1 for greenscreen.
9482 Controls amount of blue in spill area.
9483 Should be -1 for bluescreen.
9486 Controls brightness of spill area, preserving colors.
9489 Modify alpha from generated spillmap.
9492 @subsection Commands
9494 This filter supports the all above options as @ref{commands}.
9498 Apply an exact inverse of the telecine operation. It requires a predefined
9499 pattern specified using the pattern option which must be the same as that passed
9500 to the telecine filter.
9502 This filter accepts the following options:
9511 The default value is @code{top}.
9515 A string of numbers representing the pulldown pattern you wish to apply.
9516 The default value is @code{23}.
9519 A number representing position of the first frame with respect to the telecine
9520 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9525 Apply dilation effect to the video.
9527 This filter replaces the pixel by the local(3x3) maximum.
9529 It accepts the following options:
9536 Limit the maximum change for each plane, default is 65535.
9537 If 0, plane will remain unchanged.
9540 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9543 Flags to local 3x3 coordinates maps like this:
9550 @subsection Commands
9552 This filter supports the all above options as @ref{commands}.
9556 Displace pixels as indicated by second and third input stream.
9558 It takes three input streams and outputs one stream, the first input is the
9559 source, and second and third input are displacement maps.
9561 The second input specifies how much to displace pixels along the
9562 x-axis, while the third input specifies how much to displace pixels
9564 If one of displacement map streams terminates, last frame from that
9565 displacement map will be used.
9567 Note that once generated, displacements maps can be reused over and over again.
9569 A description of the accepted options follows.
9573 Set displace behavior for pixels that are out of range.
9575 Available values are:
9578 Missing pixels are replaced by black pixels.
9581 Adjacent pixels will spread out to replace missing pixels.
9584 Out of range pixels are wrapped so they point to pixels of other side.
9587 Out of range pixels will be replaced with mirrored pixels.
9589 Default is @samp{smear}.
9593 @subsection Examples
9597 Add ripple effect to rgb input of video size hd720:
9599 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
9603 Add wave effect to rgb input of video size hd720:
9605 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
9609 @anchor{dnn_processing}
9610 @section dnn_processing
9612 Do image processing with deep neural networks. It works together with another filter
9613 which converts the pixel format of the Frame to what the dnn network requires.
9615 The filter accepts the following options:
9619 Specify which DNN backend to use for model loading and execution. This option accepts
9620 the following values:
9624 Native implementation of DNN loading and execution.
9627 TensorFlow backend. To enable this backend you
9628 need to install the TensorFlow for C library (see
9629 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9630 @code{--enable-libtensorflow}
9633 OpenVINO backend. To enable this backend you
9634 need to build and install the OpenVINO for C library (see
9635 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9636 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9637 be needed if the header files and libraries are not installed into system path)
9641 Default value is @samp{native}.
9644 Set path to model file specifying network architecture and its parameters.
9645 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9646 backend can load files for only its format.
9648 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9651 Set the input name of the dnn network.
9654 Set the output name of the dnn network.
9658 @subsection Examples
9662 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9664 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9668 Halve the pixel value of the frame with format gray32f:
9670 ffmpeg -i input.jpg -vf format=grayf32,dnn_processing=model=halve_gray_float.model:input=dnn_in:output=dnn_out:dnn_backend=native -y out.native.png
9674 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9676 ./ffmpeg -i 480p.jpg -vf format=yuv420p,scale=w=iw*2:h=ih*2,dnn_processing=dnn_backend=tensorflow:model=srcnn.pb:input=x:output=y -y srcnn.jpg
9680 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9682 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9689 Draw a colored box on the input image.
9691 It accepts the following parameters:
9696 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9700 The expressions which specify the width and height of the box; if 0 they are interpreted as
9701 the input width and height. It defaults to 0.
9704 Specify the color of the box to write. For the general syntax of this option,
9705 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9706 value @code{invert} is used, the box edge color is the same as the
9707 video with inverted luma.
9710 The expression which sets the thickness of the box edge.
9711 A value of @code{fill} will create a filled box. Default value is @code{3}.
9713 See below for the list of accepted constants.
9716 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
9717 will overwrite the video's color and alpha pixels.
9718 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
9721 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9722 following constants:
9726 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9730 horizontal and vertical chroma subsample values. For example for the
9731 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9735 The input width and height.
9738 The input sample aspect ratio.
9742 The x and y offset coordinates where the box is drawn.
9746 The width and height of the drawn box.
9749 The thickness of the drawn box.
9751 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9752 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9756 @subsection Examples
9760 Draw a black box around the edge of the input image:
9766 Draw a box with color red and an opacity of 50%:
9768 drawbox=10:20:200:60:red@@0.5
9771 The previous example can be specified as:
9773 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
9777 Fill the box with pink color:
9779 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
9783 Draw a 2-pixel red 2.40:1 mask:
9785 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
9789 @subsection Commands
9790 This filter supports same commands as options.
9791 The command accepts the same syntax of the corresponding option.
9793 If the specified expression is not valid, it is kept at its current
9798 Draw a graph using input video metadata.
9800 It accepts the following parameters:
9804 Set 1st frame metadata key from which metadata values will be used to draw a graph.
9807 Set 1st foreground color expression.
9810 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
9813 Set 2nd foreground color expression.
9816 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
9819 Set 3rd foreground color expression.
9822 Set 4th frame metadata key from which metadata values will be used to draw a graph.
9825 Set 4th foreground color expression.
9828 Set minimal value of metadata value.
9831 Set maximal value of metadata value.
9834 Set graph background color. Default is white.
9839 Available values for mode is:
9846 Default is @code{line}.
9851 Available values for slide is:
9854 Draw new frame when right border is reached.
9857 Replace old columns with new ones.
9860 Scroll from right to left.
9863 Scroll from left to right.
9866 Draw single picture.
9869 Default is @code{frame}.
9872 Set size of graph video. For the syntax of this option, check the
9873 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
9874 The default value is @code{900x256}.
9877 Set the output frame rate. Default value is @code{25}.
9879 The foreground color expressions can use the following variables:
9882 Minimal value of metadata value.
9885 Maximal value of metadata value.
9888 Current metadata key value.
9891 The color is defined as 0xAABBGGRR.
9894 Example using metadata from @ref{signalstats} filter:
9896 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
9899 Example using metadata from @ref{ebur128} filter:
9901 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
9906 Draw a grid on the input image.
9908 It accepts the following parameters:
9913 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
9917 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
9918 input width and height, respectively, minus @code{thickness}, so image gets
9919 framed. Default to 0.
9922 Specify the color of the grid. For the general syntax of this option,
9923 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9924 value @code{invert} is used, the grid color is the same as the
9925 video with inverted luma.
9928 The expression which sets the thickness of the grid line. Default value is @code{1}.
9930 See below for the list of accepted constants.
9933 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
9934 will overwrite the video's color and alpha pixels.
9935 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
9938 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9939 following constants:
9943 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9947 horizontal and vertical chroma subsample values. For example for the
9948 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9952 The input grid cell width and height.
9955 The input sample aspect ratio.
9959 The x and y coordinates of some point of grid intersection (meant to configure offset).
9963 The width and height of the drawn cell.
9966 The thickness of the drawn cell.
9968 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9969 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9973 @subsection Examples
9977 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
9979 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
9983 Draw a white 3x3 grid with an opacity of 50%:
9985 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
9989 @subsection Commands
9990 This filter supports same commands as options.
9991 The command accepts the same syntax of the corresponding option.
9993 If the specified expression is not valid, it is kept at its current
9999 Draw a text string or text from a specified file on top of a video, using the
10000 libfreetype library.
10002 To enable compilation of this filter, you need to configure FFmpeg with
10003 @code{--enable-libfreetype}.
10004 To enable default font fallback and the @var{font} option you need to
10005 configure FFmpeg with @code{--enable-libfontconfig}.
10006 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10007 @code{--enable-libfribidi}.
10011 It accepts the following parameters:
10016 Used to draw a box around text using the background color.
10017 The value must be either 1 (enable) or 0 (disable).
10018 The default value of @var{box} is 0.
10021 Set the width of the border to be drawn around the box using @var{boxcolor}.
10022 The default value of @var{boxborderw} is 0.
10025 The color to be used for drawing box around text. For the syntax of this
10026 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10028 The default value of @var{boxcolor} is "white".
10031 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10032 The default value of @var{line_spacing} is 0.
10035 Set the width of the border to be drawn around the text using @var{bordercolor}.
10036 The default value of @var{borderw} is 0.
10039 Set the color to be used for drawing border around text. For the syntax of this
10040 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10042 The default value of @var{bordercolor} is "black".
10045 Select how the @var{text} is expanded. Can be either @code{none},
10046 @code{strftime} (deprecated) or
10047 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10051 Set a start time for the count. Value is in microseconds. Only applied
10052 in the deprecated strftime expansion mode. To emulate in normal expansion
10053 mode use the @code{pts} function, supplying the start time (in seconds)
10054 as the second argument.
10057 If true, check and fix text coords to avoid clipping.
10060 The color to be used for drawing fonts. For the syntax of this option, check
10061 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10063 The default value of @var{fontcolor} is "black".
10065 @item fontcolor_expr
10066 String which is expanded the same way as @var{text} to obtain dynamic
10067 @var{fontcolor} value. By default this option has empty value and is not
10068 processed. When this option is set, it overrides @var{fontcolor} option.
10071 The font family to be used for drawing text. By default Sans.
10074 The font file to be used for drawing text. The path must be included.
10075 This parameter is mandatory if the fontconfig support is disabled.
10078 Draw the text applying alpha blending. The value can
10079 be a number between 0.0 and 1.0.
10080 The expression accepts the same variables @var{x, y} as well.
10081 The default value is 1.
10082 Please see @var{fontcolor_expr}.
10085 The font size to be used for drawing text.
10086 The default value of @var{fontsize} is 16.
10089 If set to 1, attempt to shape the text (for example, reverse the order of
10090 right-to-left text and join Arabic characters) before drawing it.
10091 Otherwise, just draw the text exactly as given.
10092 By default 1 (if supported).
10094 @item ft_load_flags
10095 The flags to be used for loading the fonts.
10097 The flags map the corresponding flags supported by libfreetype, and are
10098 a combination of the following values:
10105 @item vertical_layout
10106 @item force_autohint
10109 @item ignore_global_advance_width
10111 @item ignore_transform
10113 @item linear_design
10117 Default value is "default".
10119 For more information consult the documentation for the FT_LOAD_*
10123 The color to be used for drawing a shadow behind the drawn text. For the
10124 syntax of this option, check the @ref{color syntax,,"Color" section in the
10125 ffmpeg-utils manual,ffmpeg-utils}.
10127 The default value of @var{shadowcolor} is "black".
10131 The x and y offsets for the text shadow position with respect to the
10132 position of the text. They can be either positive or negative
10133 values. The default value for both is "0".
10136 The starting frame number for the n/frame_num variable. The default value
10140 The size in number of spaces to use for rendering the tab.
10141 Default value is 4.
10144 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10145 format. It can be used with or without text parameter. @var{timecode_rate}
10146 option must be specified.
10148 @item timecode_rate, rate, r
10149 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10150 integer. Minimum value is "1".
10151 Drop-frame timecode is supported for frame rates 30 & 60.
10154 If set to 1, the output of the timecode option will wrap around at 24 hours.
10155 Default is 0 (disabled).
10158 The text string to be drawn. The text must be a sequence of UTF-8
10159 encoded characters.
10160 This parameter is mandatory if no file is specified with the parameter
10164 A text file containing text to be drawn. The text must be a sequence
10165 of UTF-8 encoded characters.
10167 This parameter is mandatory if no text string is specified with the
10168 parameter @var{text}.
10170 If both @var{text} and @var{textfile} are specified, an error is thrown.
10173 If set to 1, the @var{textfile} will be reloaded before each frame.
10174 Be sure to update it atomically, or it may be read partially, or even fail.
10178 The expressions which specify the offsets where text will be drawn
10179 within the video frame. They are relative to the top/left border of the
10182 The default value of @var{x} and @var{y} is "0".
10184 See below for the list of accepted constants and functions.
10187 The parameters for @var{x} and @var{y} are expressions containing the
10188 following constants and functions:
10192 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10196 horizontal and vertical chroma subsample values. For example for the
10197 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10200 the height of each text line
10208 @item max_glyph_a, ascent
10209 the maximum distance from the baseline to the highest/upper grid
10210 coordinate used to place a glyph outline point, for all the rendered
10212 It is a positive value, due to the grid's orientation with the Y axis
10215 @item max_glyph_d, descent
10216 the maximum distance from the baseline to the lowest grid coordinate
10217 used to place a glyph outline point, for all the rendered glyphs.
10218 This is a negative value, due to the grid's orientation, with the Y axis
10222 maximum glyph height, that is the maximum height for all the glyphs
10223 contained in the rendered text, it is equivalent to @var{ascent} -
10227 maximum glyph width, that is the maximum width for all the glyphs
10228 contained in the rendered text
10231 the number of input frame, starting from 0
10233 @item rand(min, max)
10234 return a random number included between @var{min} and @var{max}
10237 The input sample aspect ratio.
10240 timestamp expressed in seconds, NAN if the input timestamp is unknown
10243 the height of the rendered text
10246 the width of the rendered text
10250 the x and y offset coordinates where the text is drawn.
10252 These parameters allow the @var{x} and @var{y} expressions to refer
10253 to each other, so you can for example specify @code{y=x/dar}.
10256 A one character description of the current frame's picture type.
10259 The current packet's position in the input file or stream
10260 (in bytes, from the start of the input). A value of -1 indicates
10261 this info is not available.
10264 The current packet's duration, in seconds.
10267 The current packet's size (in bytes).
10270 @anchor{drawtext_expansion}
10271 @subsection Text expansion
10273 If @option{expansion} is set to @code{strftime},
10274 the filter recognizes strftime() sequences in the provided text and
10275 expands them accordingly. Check the documentation of strftime(). This
10276 feature is deprecated.
10278 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10280 If @option{expansion} is set to @code{normal} (which is the default),
10281 the following expansion mechanism is used.
10283 The backslash character @samp{\}, followed by any character, always expands to
10284 the second character.
10286 Sequences of the form @code{%@{...@}} are expanded. The text between the
10287 braces is a function name, possibly followed by arguments separated by ':'.
10288 If the arguments contain special characters or delimiters (':' or '@}'),
10289 they should be escaped.
10291 Note that they probably must also be escaped as the value for the
10292 @option{text} option in the filter argument string and as the filter
10293 argument in the filtergraph description, and possibly also for the shell,
10294 that makes up to four levels of escaping; using a text file avoids these
10297 The following functions are available:
10302 The expression evaluation result.
10304 It must take one argument specifying the expression to be evaluated,
10305 which accepts the same constants and functions as the @var{x} and
10306 @var{y} values. Note that not all constants should be used, for
10307 example the text size is not known when evaluating the expression, so
10308 the constants @var{text_w} and @var{text_h} will have an undefined
10311 @item expr_int_format, eif
10312 Evaluate the expression's value and output as formatted integer.
10314 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10315 The second argument specifies the output format. Allowed values are @samp{x},
10316 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10317 @code{printf} function.
10318 The third parameter is optional and sets the number of positions taken by the output.
10319 It can be used to add padding with zeros from the left.
10322 The time at which the filter is running, expressed in UTC.
10323 It can accept an argument: a strftime() format string.
10326 The time at which the filter is running, expressed in the local time zone.
10327 It can accept an argument: a strftime() format string.
10330 Frame metadata. Takes one or two arguments.
10332 The first argument is mandatory and specifies the metadata key.
10334 The second argument is optional and specifies a default value, used when the
10335 metadata key is not found or empty.
10337 Available metadata can be identified by inspecting entries
10338 starting with TAG included within each frame section
10339 printed by running @code{ffprobe -show_frames}.
10341 String metadata generated in filters leading to
10342 the drawtext filter are also available.
10345 The frame number, starting from 0.
10348 A one character description of the current picture type.
10351 The timestamp of the current frame.
10352 It can take up to three arguments.
10354 The first argument is the format of the timestamp; it defaults to @code{flt}
10355 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10356 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10357 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10358 @code{localtime} stands for the timestamp of the frame formatted as
10359 local time zone time.
10361 The second argument is an offset added to the timestamp.
10363 If the format is set to @code{hms}, a third argument @code{24HH} may be
10364 supplied to present the hour part of the formatted timestamp in 24h format
10367 If the format is set to @code{localtime} or @code{gmtime},
10368 a third argument may be supplied: a strftime() format string.
10369 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10372 @subsection Commands
10374 This filter supports altering parameters via commands:
10377 Alter existing filter parameters.
10379 Syntax for the argument is the same as for filter invocation, e.g.
10382 fontsize=56:fontcolor=green:text='Hello World'
10385 Full filter invocation with sendcmd would look like this:
10388 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10392 If the entire argument can't be parsed or applied as valid values then the filter will
10393 continue with its existing parameters.
10395 @subsection Examples
10399 Draw "Test Text" with font FreeSerif, using the default values for the
10400 optional parameters.
10403 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10407 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10408 and y=50 (counting from the top-left corner of the screen), text is
10409 yellow with a red box around it. Both the text and the box have an
10413 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10414 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10417 Note that the double quotes are not necessary if spaces are not used
10418 within the parameter list.
10421 Show the text at the center of the video frame:
10423 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10427 Show the text at a random position, switching to a new position every 30 seconds:
10429 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)"
10433 Show a text line sliding from right to left in the last row of the video
10434 frame. The file @file{LONG_LINE} is assumed to contain a single line
10437 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10441 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10443 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10447 Draw a single green letter "g", at the center of the input video.
10448 The glyph baseline is placed at half screen height.
10450 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10454 Show text for 1 second every 3 seconds:
10456 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10460 Use fontconfig to set the font. Note that the colons need to be escaped.
10462 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10466 Draw "Test Text" with font size dependent on height of the video.
10468 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10472 Print the date of a real-time encoding (see strftime(3)):
10474 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10478 Show text fading in and out (appearing/disappearing):
10481 DS=1.0 # display start
10482 DE=10.0 # display end
10483 FID=1.5 # fade in duration
10484 FOD=5 # fade out duration
10485 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 @}"
10489 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10490 and the @option{fontsize} value are included in the @option{y} offset.
10492 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10493 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10497 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10498 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10499 must have option @option{-export_path_metadata 1} for the special metadata fields
10500 to be available for filters.
10502 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10507 For more information about libfreetype, check:
10508 @url{http://www.freetype.org/}.
10510 For more information about fontconfig, check:
10511 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10513 For more information about libfribidi, check:
10514 @url{http://fribidi.org/}.
10516 @section edgedetect
10518 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10520 The filter accepts the following options:
10525 Set low and high threshold values used by the Canny thresholding
10528 The high threshold selects the "strong" edge pixels, which are then
10529 connected through 8-connectivity with the "weak" edge pixels selected
10530 by the low threshold.
10532 @var{low} and @var{high} threshold values must be chosen in the range
10533 [0,1], and @var{low} should be lesser or equal to @var{high}.
10535 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10539 Define the drawing mode.
10543 Draw white/gray wires on black background.
10546 Mix the colors to create a paint/cartoon effect.
10549 Apply Canny edge detector on all selected planes.
10551 Default value is @var{wires}.
10554 Select planes for filtering. By default all available planes are filtered.
10557 @subsection Examples
10561 Standard edge detection with custom values for the hysteresis thresholding:
10563 edgedetect=low=0.1:high=0.4
10567 Painting effect without thresholding:
10569 edgedetect=mode=colormix:high=0
10575 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10577 For each input image, the filter will compute the optimal mapping from
10578 the input to the output given the codebook length, that is the number
10579 of distinct output colors.
10581 This filter accepts the following options.
10584 @item codebook_length, l
10585 Set codebook length. The value must be a positive integer, and
10586 represents the number of distinct output colors. Default value is 256.
10589 Set the maximum number of iterations to apply for computing the optimal
10590 mapping. The higher the value the better the result and the higher the
10591 computation time. Default value is 1.
10594 Set a random seed, must be an integer included between 0 and
10595 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10596 will try to use a good random seed on a best effort basis.
10599 Set pal8 output pixel format. This option does not work with codebook
10600 length greater than 256.
10605 Measure graylevel entropy in histogram of color channels of video frames.
10607 It accepts the following parameters:
10611 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10613 @var{diff} mode measures entropy of histogram delta values, absolute differences
10614 between neighbour histogram values.
10618 Set brightness, contrast, saturation and approximate gamma adjustment.
10620 The filter accepts the following options:
10624 Set the contrast expression. The value must be a float value in range
10625 @code{-1000.0} to @code{1000.0}. The default value is "1".
10628 Set the brightness expression. The value must be a float value in
10629 range @code{-1.0} to @code{1.0}. The default value is "0".
10632 Set the saturation expression. The value must be a float in
10633 range @code{0.0} to @code{3.0}. The default value is "1".
10636 Set the gamma expression. The value must be a float in range
10637 @code{0.1} to @code{10.0}. The default value is "1".
10640 Set the gamma expression for red. The value must be a float in
10641 range @code{0.1} to @code{10.0}. The default value is "1".
10644 Set the gamma expression for green. The value must be a float in range
10645 @code{0.1} to @code{10.0}. The default value is "1".
10648 Set the gamma expression for blue. The value must be a float in range
10649 @code{0.1} to @code{10.0}. The default value is "1".
10652 Set the gamma weight expression. It can be used to reduce the effect
10653 of a high gamma value on bright image areas, e.g. keep them from
10654 getting overamplified and just plain white. The value must be a float
10655 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10656 gamma correction all the way down while @code{1.0} leaves it at its
10657 full strength. Default is "1".
10660 Set when the expressions for brightness, contrast, saturation and
10661 gamma expressions are evaluated.
10663 It accepts the following values:
10666 only evaluate expressions once during the filter initialization or
10667 when a command is processed
10670 evaluate expressions for each incoming frame
10673 Default value is @samp{init}.
10676 The expressions accept the following parameters:
10679 frame count of the input frame starting from 0
10682 byte position of the corresponding packet in the input file, NAN if
10686 frame rate of the input video, NAN if the input frame rate is unknown
10689 timestamp expressed in seconds, NAN if the input timestamp is unknown
10692 @subsection Commands
10693 The filter supports the following commands:
10697 Set the contrast expression.
10700 Set the brightness expression.
10703 Set the saturation expression.
10706 Set the gamma expression.
10709 Set the gamma_r expression.
10712 Set gamma_g expression.
10715 Set gamma_b expression.
10718 Set gamma_weight expression.
10720 The command accepts the same syntax of the corresponding option.
10722 If the specified expression is not valid, it is kept at its current
10729 Apply erosion effect to the video.
10731 This filter replaces the pixel by the local(3x3) minimum.
10733 It accepts the following options:
10740 Limit the maximum change for each plane, default is 65535.
10741 If 0, plane will remain unchanged.
10744 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
10747 Flags to local 3x3 coordinates maps like this:
10754 @subsection Commands
10756 This filter supports the all above options as @ref{commands}.
10758 @section extractplanes
10760 Extract color channel components from input video stream into
10761 separate grayscale video streams.
10763 The filter accepts the following option:
10767 Set plane(s) to extract.
10769 Available values for planes are:
10780 Choosing planes not available in the input will result in an error.
10781 That means you cannot select @code{r}, @code{g}, @code{b} planes
10782 with @code{y}, @code{u}, @code{v} planes at same time.
10785 @subsection Examples
10789 Extract luma, u and v color channel component from input video frame
10790 into 3 grayscale outputs:
10792 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
10798 Apply a fade-in/out effect to the input video.
10800 It accepts the following parameters:
10804 The effect type can be either "in" for a fade-in, or "out" for a fade-out
10806 Default is @code{in}.
10808 @item start_frame, s
10809 Specify the number of the frame to start applying the fade
10810 effect at. Default is 0.
10813 The number of frames that the fade effect lasts. At the end of the
10814 fade-in effect, the output video will have the same intensity as the input video.
10815 At the end of the fade-out transition, the output video will be filled with the
10816 selected @option{color}.
10820 If set to 1, fade only alpha channel, if one exists on the input.
10821 Default value is 0.
10823 @item start_time, st
10824 Specify the timestamp (in seconds) of the frame to start to apply the fade
10825 effect. If both start_frame and start_time are specified, the fade will start at
10826 whichever comes last. Default is 0.
10829 The number of seconds for which the fade effect has to last. At the end of the
10830 fade-in effect the output video will have the same intensity as the input video,
10831 at the end of the fade-out transition the output video will be filled with the
10832 selected @option{color}.
10833 If both duration and nb_frames are specified, duration is used. Default is 0
10834 (nb_frames is used by default).
10837 Specify the color of the fade. Default is "black".
10840 @subsection Examples
10844 Fade in the first 30 frames of video:
10849 The command above is equivalent to:
10855 Fade out the last 45 frames of a 200-frame video:
10858 fade=type=out:start_frame=155:nb_frames=45
10862 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
10864 fade=in:0:25, fade=out:975:25
10868 Make the first 5 frames yellow, then fade in from frame 5-24:
10870 fade=in:5:20:color=yellow
10874 Fade in alpha over first 25 frames of video:
10876 fade=in:0:25:alpha=1
10880 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
10882 fade=t=in:st=5.5:d=0.5
10888 Denoise frames using 3D FFT (frequency domain filtering).
10890 The filter accepts the following options:
10894 Set the noise sigma constant. This sets denoising strength.
10895 Default value is 1. Allowed range is from 0 to 30.
10896 Using very high sigma with low overlap may give blocking artifacts.
10899 Set amount of denoising. By default all detected noise is reduced.
10900 Default value is 1. Allowed range is from 0 to 1.
10903 Set size of block, Default is 4, can be 3, 4, 5 or 6.
10904 Actual size of block in pixels is 2 to power of @var{block}, so by default
10905 block size in pixels is 2^4 which is 16.
10908 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
10911 Set number of previous frames to use for denoising. By default is set to 0.
10914 Set number of next frames to to use for denoising. By default is set to 0.
10917 Set planes which will be filtered, by default are all available filtered
10922 Apply arbitrary expressions to samples in frequency domain
10926 Adjust the dc value (gain) of the luma plane of the image. The filter
10927 accepts an integer value in range @code{0} to @code{1000}. The default
10928 value is set to @code{0}.
10931 Adjust the dc value (gain) of the 1st chroma plane of the image. The
10932 filter accepts an integer value in range @code{0} to @code{1000}. The
10933 default value is set to @code{0}.
10936 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
10937 filter accepts an integer value in range @code{0} to @code{1000}. The
10938 default value is set to @code{0}.
10941 Set the frequency domain weight expression for the luma plane.
10944 Set the frequency domain weight expression for the 1st chroma plane.
10947 Set the frequency domain weight expression for the 2nd chroma plane.
10950 Set when the expressions are evaluated.
10952 It accepts the following values:
10955 Only evaluate expressions once during the filter initialization.
10958 Evaluate expressions for each incoming frame.
10961 Default value is @samp{init}.
10963 The filter accepts the following variables:
10966 The coordinates of the current sample.
10970 The width and height of the image.
10973 The number of input frame, starting from 0.
10976 @subsection Examples
10982 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
10988 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
10994 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11000 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11007 Extract a single field from an interlaced image using stride
11008 arithmetic to avoid wasting CPU time. The output frames are marked as
11011 The filter accepts the following options:
11015 Specify whether to extract the top (if the value is @code{0} or
11016 @code{top}) or the bottom field (if the value is @code{1} or
11022 Create new frames by copying the top and bottom fields from surrounding frames
11023 supplied as numbers by the hint file.
11027 Set file containing hints: absolute/relative frame numbers.
11029 There must be one line for each frame in a clip. Each line must contain two
11030 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11031 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11032 is current frame number for @code{absolute} mode or out of [-1, 1] range
11033 for @code{relative} mode. First number tells from which frame to pick up top
11034 field and second number tells from which frame to pick up bottom field.
11036 If optionally followed by @code{+} output frame will be marked as interlaced,
11037 else if followed by @code{-} output frame will be marked as progressive, else
11038 it will be marked same as input frame.
11039 If optionally followed by @code{t} output frame will use only top field, or in
11040 case of @code{b} it will use only bottom field.
11041 If line starts with @code{#} or @code{;} that line is skipped.
11044 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11047 Example of first several lines of @code{hint} file for @code{relative} mode:
11049 0,0 - # first frame
11050 1,0 - # second frame, use third's frame top field and second's frame bottom field
11051 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11066 @section fieldmatch
11068 Field matching filter for inverse telecine. It is meant to reconstruct the
11069 progressive frames from a telecined stream. The filter does not drop duplicated
11070 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11071 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11073 The separation of the field matching and the decimation is notably motivated by
11074 the possibility of inserting a de-interlacing filter fallback between the two.
11075 If the source has mixed telecined and real interlaced content,
11076 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11077 But these remaining combed frames will be marked as interlaced, and thus can be
11078 de-interlaced by a later filter such as @ref{yadif} before decimation.
11080 In addition to the various configuration options, @code{fieldmatch} can take an
11081 optional second stream, activated through the @option{ppsrc} option. If
11082 enabled, the frames reconstruction will be based on the fields and frames from
11083 this second stream. This allows the first input to be pre-processed in order to
11084 help the various algorithms of the filter, while keeping the output lossless
11085 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11086 or brightness/contrast adjustments can help.
11088 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11089 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11090 which @code{fieldmatch} is based on. While the semantic and usage are very
11091 close, some behaviour and options names can differ.
11093 The @ref{decimate} filter currently only works for constant frame rate input.
11094 If your input has mixed telecined (30fps) and progressive content with a lower
11095 framerate like 24fps use the following filterchain to produce the necessary cfr
11096 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11098 The filter accepts the following options:
11102 Specify the assumed field order of the input stream. Available values are:
11106 Auto detect parity (use FFmpeg's internal parity value).
11108 Assume bottom field first.
11110 Assume top field first.
11113 Note that it is sometimes recommended not to trust the parity announced by the
11116 Default value is @var{auto}.
11119 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11120 sense that it won't risk creating jerkiness due to duplicate frames when
11121 possible, but if there are bad edits or blended fields it will end up
11122 outputting combed frames when a good match might actually exist. On the other
11123 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11124 but will almost always find a good frame if there is one. The other values are
11125 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11126 jerkiness and creating duplicate frames versus finding good matches in sections
11127 with bad edits, orphaned fields, blended fields, etc.
11129 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11131 Available values are:
11135 2-way matching (p/c)
11137 2-way matching, and trying 3rd match if still combed (p/c + n)
11139 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11141 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11142 still combed (p/c + n + u/b)
11144 3-way matching (p/c/n)
11146 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11147 detected as combed (p/c/n + u/b)
11150 The parenthesis at the end indicate the matches that would be used for that
11151 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11154 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11157 Default value is @var{pc_n}.
11160 Mark the main input stream as a pre-processed input, and enable the secondary
11161 input stream as the clean source to pick the fields from. See the filter
11162 introduction for more details. It is similar to the @option{clip2} feature from
11165 Default value is @code{0} (disabled).
11168 Set the field to match from. It is recommended to set this to the same value as
11169 @option{order} unless you experience matching failures with that setting. In
11170 certain circumstances changing the field that is used to match from can have a
11171 large impact on matching performance. Available values are:
11175 Automatic (same value as @option{order}).
11177 Match from the bottom field.
11179 Match from the top field.
11182 Default value is @var{auto}.
11185 Set whether or not chroma is included during the match comparisons. In most
11186 cases it is recommended to leave this enabled. You should set this to @code{0}
11187 only if your clip has bad chroma problems such as heavy rainbowing or other
11188 artifacts. Setting this to @code{0} could also be used to speed things up at
11189 the cost of some accuracy.
11191 Default value is @code{1}.
11195 These define an exclusion band which excludes the lines between @option{y0} and
11196 @option{y1} from being included in the field matching decision. An exclusion
11197 band can be used to ignore subtitles, a logo, or other things that may
11198 interfere with the matching. @option{y0} sets the starting scan line and
11199 @option{y1} sets the ending line; all lines in between @option{y0} and
11200 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11201 @option{y0} and @option{y1} to the same value will disable the feature.
11202 @option{y0} and @option{y1} defaults to @code{0}.
11205 Set the scene change detection threshold as a percentage of maximum change on
11206 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11207 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11208 @option{scthresh} is @code{[0.0, 100.0]}.
11210 Default value is @code{12.0}.
11213 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11214 account the combed scores of matches when deciding what match to use as the
11215 final match. Available values are:
11219 No final matching based on combed scores.
11221 Combed scores are only used when a scene change is detected.
11223 Use combed scores all the time.
11226 Default is @var{sc}.
11229 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11230 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11231 Available values are:
11235 No forced calculation.
11237 Force p/c/n calculations.
11239 Force p/c/n/u/b calculations.
11242 Default value is @var{none}.
11245 This is the area combing threshold used for combed frame detection. This
11246 essentially controls how "strong" or "visible" combing must be to be detected.
11247 Larger values mean combing must be more visible and smaller values mean combing
11248 can be less visible or strong and still be detected. Valid settings are from
11249 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11250 be detected as combed). This is basically a pixel difference value. A good
11251 range is @code{[8, 12]}.
11253 Default value is @code{9}.
11256 Sets whether or not chroma is considered in the combed frame decision. Only
11257 disable this if your source has chroma problems (rainbowing, etc.) that are
11258 causing problems for the combed frame detection with chroma enabled. Actually,
11259 using @option{chroma}=@var{0} is usually more reliable, except for the case
11260 where there is chroma only combing in the source.
11262 Default value is @code{0}.
11266 Respectively set the x-axis and y-axis size of the window used during combed
11267 frame detection. This has to do with the size of the area in which
11268 @option{combpel} pixels are required to be detected as combed for a frame to be
11269 declared combed. See the @option{combpel} parameter description for more info.
11270 Possible values are any number that is a power of 2 starting at 4 and going up
11273 Default value is @code{16}.
11276 The number of combed pixels inside any of the @option{blocky} by
11277 @option{blockx} size blocks on the frame for the frame to be detected as
11278 combed. While @option{cthresh} controls how "visible" the combing must be, this
11279 setting controls "how much" combing there must be in any localized area (a
11280 window defined by the @option{blockx} and @option{blocky} settings) on the
11281 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11282 which point no frames will ever be detected as combed). This setting is known
11283 as @option{MI} in TFM/VFM vocabulary.
11285 Default value is @code{80}.
11288 @anchor{p/c/n/u/b meaning}
11289 @subsection p/c/n/u/b meaning
11291 @subsubsection p/c/n
11293 We assume the following telecined stream:
11296 Top fields: 1 2 2 3 4
11297 Bottom fields: 1 2 3 4 4
11300 The numbers correspond to the progressive frame the fields relate to. Here, the
11301 first two frames are progressive, the 3rd and 4th are combed, and so on.
11303 When @code{fieldmatch} is configured to run a matching from bottom
11304 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11309 B 1 2 3 4 4 <-- matching reference
11318 As a result of the field matching, we can see that some frames get duplicated.
11319 To perform a complete inverse telecine, you need to rely on a decimation filter
11320 after this operation. See for instance the @ref{decimate} filter.
11322 The same operation now matching from top fields (@option{field}=@var{top})
11327 T 1 2 2 3 4 <-- matching reference
11337 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11338 basically, they refer to the frame and field of the opposite parity:
11341 @item @var{p} matches the field of the opposite parity in the previous frame
11342 @item @var{c} matches the field of the opposite parity in the current frame
11343 @item @var{n} matches the field of the opposite parity in the next frame
11348 The @var{u} and @var{b} matching are a bit special in the sense that they match
11349 from the opposite parity flag. In the following examples, we assume that we are
11350 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11351 'x' is placed above and below each matched fields.
11353 With bottom matching (@option{field}=@var{bottom}):
11358 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11359 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11367 With top matching (@option{field}=@var{top}):
11372 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11373 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11381 @subsection Examples
11383 Simple IVTC of a top field first telecined stream:
11385 fieldmatch=order=tff:combmatch=none, decimate
11388 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11390 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11393 @section fieldorder
11395 Transform the field order of the input video.
11397 It accepts the following parameters:
11402 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11403 for bottom field first.
11406 The default value is @samp{tff}.
11408 The transformation is done by shifting the picture content up or down
11409 by one line, and filling the remaining line with appropriate picture content.
11410 This method is consistent with most broadcast field order converters.
11412 If the input video is not flagged as being interlaced, or it is already
11413 flagged as being of the required output field order, then this filter does
11414 not alter the incoming video.
11416 It is very useful when converting to or from PAL DV material,
11417 which is bottom field first.
11421 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11424 @section fifo, afifo
11426 Buffer input images and send them when they are requested.
11428 It is mainly useful when auto-inserted by the libavfilter
11431 It does not take parameters.
11433 @section fillborders
11435 Fill borders of the input video, without changing video stream dimensions.
11436 Sometimes video can have garbage at the four edges and you may not want to
11437 crop video input to keep size multiple of some number.
11439 This filter accepts the following options:
11443 Number of pixels to fill from left border.
11446 Number of pixels to fill from right border.
11449 Number of pixels to fill from top border.
11452 Number of pixels to fill from bottom border.
11457 It accepts the following values:
11460 fill pixels using outermost pixels
11463 fill pixels using mirroring
11466 fill pixels with constant value
11469 Default is @var{smear}.
11472 Set color for pixels in fixed mode. Default is @var{black}.
11475 @subsection Commands
11476 This filter supports same @ref{commands} as options.
11477 The command accepts the same syntax of the corresponding option.
11479 If the specified expression is not valid, it is kept at its current
11484 Find a rectangular object
11486 It accepts the following options:
11490 Filepath of the object image, needs to be in gray8.
11493 Detection threshold, default is 0.5.
11496 Number of mipmaps, default is 3.
11498 @item xmin, ymin, xmax, ymax
11499 Specifies the rectangle in which to search.
11502 @subsection Examples
11506 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11508 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11514 Flood area with values of same pixel components with another values.
11516 It accepts the following options:
11519 Set pixel x coordinate.
11522 Set pixel y coordinate.
11525 Set source #0 component value.
11528 Set source #1 component value.
11531 Set source #2 component value.
11534 Set source #3 component value.
11537 Set destination #0 component value.
11540 Set destination #1 component value.
11543 Set destination #2 component value.
11546 Set destination #3 component value.
11552 Convert the input video to one of the specified pixel formats.
11553 Libavfilter will try to pick one that is suitable as input to
11556 It accepts the following parameters:
11560 A '|'-separated list of pixel format names, such as
11561 "pix_fmts=yuv420p|monow|rgb24".
11565 @subsection Examples
11569 Convert the input video to the @var{yuv420p} format
11571 format=pix_fmts=yuv420p
11574 Convert the input video to any of the formats in the list
11576 format=pix_fmts=yuv420p|yuv444p|yuv410p
11583 Convert the video to specified constant frame rate by duplicating or dropping
11584 frames as necessary.
11586 It accepts the following parameters:
11590 The desired output frame rate. The default is @code{25}.
11593 Assume the first PTS should be the given value, in seconds. This allows for
11594 padding/trimming at the start of stream. By default, no assumption is made
11595 about the first frame's expected PTS, so no padding or trimming is done.
11596 For example, this could be set to 0 to pad the beginning with duplicates of
11597 the first frame if a video stream starts after the audio stream or to trim any
11598 frames with a negative PTS.
11601 Timestamp (PTS) rounding method.
11603 Possible values are:
11610 round towards -infinity
11612 round towards +infinity
11616 The default is @code{near}.
11619 Action performed when reading the last frame.
11621 Possible values are:
11624 Use same timestamp rounding method as used for other frames.
11626 Pass through last frame if input duration has not been reached yet.
11628 The default is @code{round}.
11632 Alternatively, the options can be specified as a flat string:
11633 @var{fps}[:@var{start_time}[:@var{round}]].
11635 See also the @ref{setpts} filter.
11637 @subsection Examples
11641 A typical usage in order to set the fps to 25:
11647 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11649 fps=fps=film:round=near
11655 Pack two different video streams into a stereoscopic video, setting proper
11656 metadata on supported codecs. The two views should have the same size and
11657 framerate and processing will stop when the shorter video ends. Please note
11658 that you may conveniently adjust view properties with the @ref{scale} and
11661 It accepts the following parameters:
11665 The desired packing format. Supported values are:
11670 The views are next to each other (default).
11673 The views are on top of each other.
11676 The views are packed by line.
11679 The views are packed by column.
11682 The views are temporally interleaved.
11691 # Convert left and right views into a frame-sequential video
11692 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11694 # Convert views into a side-by-side video with the same output resolution as the input
11695 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
11700 Change the frame rate by interpolating new video output frames from the source
11703 This filter is not designed to function correctly with interlaced media. If
11704 you wish to change the frame rate of interlaced media then you are required
11705 to deinterlace before this filter and re-interlace after this filter.
11707 A description of the accepted options follows.
11711 Specify the output frames per second. This option can also be specified
11712 as a value alone. The default is @code{50}.
11715 Specify the start of a range where the output frame will be created as a
11716 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11717 the default is @code{15}.
11720 Specify the end of a range where the output frame will be created as a
11721 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11722 the default is @code{240}.
11725 Specify the level at which a scene change is detected as a value between
11726 0 and 100 to indicate a new scene; a low value reflects a low
11727 probability for the current frame to introduce a new scene, while a higher
11728 value means the current frame is more likely to be one.
11729 The default is @code{8.2}.
11732 Specify flags influencing the filter process.
11734 Available value for @var{flags} is:
11737 @item scene_change_detect, scd
11738 Enable scene change detection using the value of the option @var{scene}.
11739 This flag is enabled by default.
11745 Select one frame every N-th frame.
11747 This filter accepts the following option:
11750 Select frame after every @code{step} frames.
11751 Allowed values are positive integers higher than 0. Default value is @code{1}.
11754 @section freezedetect
11756 Detect frozen video.
11758 This filter logs a message and sets frame metadata when it detects that the
11759 input video has no significant change in content during a specified duration.
11760 Video freeze detection calculates the mean average absolute difference of all
11761 the components of video frames and compares it to a noise floor.
11763 The printed times and duration are expressed in seconds. The
11764 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
11765 whose timestamp equals or exceeds the detection duration and it contains the
11766 timestamp of the first frame of the freeze. The
11767 @code{lavfi.freezedetect.freeze_duration} and
11768 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
11771 The filter accepts the following options:
11775 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
11776 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
11780 Set freeze duration until notification (default is 2 seconds).
11783 @section freezeframes
11785 Freeze video frames.
11787 This filter freezes video frames using frame from 2nd input.
11789 The filter accepts the following options:
11793 Set number of first frame from which to start freeze.
11796 Set number of last frame from which to end freeze.
11799 Set number of frame from 2nd input which will be used instead of replaced frames.
11805 Apply a frei0r effect to the input video.
11807 To enable the compilation of this filter, you need to install the frei0r
11808 header and configure FFmpeg with @code{--enable-frei0r}.
11810 It accepts the following parameters:
11815 The name of the frei0r effect to load. If the environment variable
11816 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
11817 directories specified by the colon-separated list in @env{FREI0R_PATH}.
11818 Otherwise, the standard frei0r paths are searched, in this order:
11819 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
11820 @file{/usr/lib/frei0r-1/}.
11822 @item filter_params
11823 A '|'-separated list of parameters to pass to the frei0r effect.
11827 A frei0r effect parameter can be a boolean (its value is either
11828 "y" or "n"), a double, a color (specified as
11829 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
11830 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
11831 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
11832 a position (specified as @var{X}/@var{Y}, where
11833 @var{X} and @var{Y} are floating point numbers) and/or a string.
11835 The number and types of parameters depend on the loaded effect. If an
11836 effect parameter is not specified, the default value is set.
11838 @subsection Examples
11842 Apply the distort0r effect, setting the first two double parameters:
11844 frei0r=filter_name=distort0r:filter_params=0.5|0.01
11848 Apply the colordistance effect, taking a color as the first parameter:
11850 frei0r=colordistance:0.2/0.3/0.4
11851 frei0r=colordistance:violet
11852 frei0r=colordistance:0x112233
11856 Apply the perspective effect, specifying the top left and top right image
11859 frei0r=perspective:0.2/0.2|0.8/0.2
11863 For more information, see
11864 @url{http://frei0r.dyne.org}
11866 @subsection Commands
11868 This filter supports the @option{filter_params} option as @ref{commands}.
11872 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
11874 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
11875 processing filter, one of them is performed once per block, not per pixel.
11876 This allows for much higher speed.
11878 The filter accepts the following options:
11882 Set quality. This option defines the number of levels for averaging. It accepts
11883 an integer in the range 4-5. Default value is @code{4}.
11886 Force a constant quantization parameter. It accepts an integer in range 0-63.
11887 If not set, the filter will use the QP from the video stream (if available).
11890 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
11891 more details but also more artifacts, while higher values make the image smoother
11892 but also blurrier. Default value is @code{0} − PSNR optimal.
11894 @item use_bframe_qp
11895 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
11896 option may cause flicker since the B-Frames have often larger QP. Default is
11897 @code{0} (not enabled).
11903 Apply Gaussian blur filter.
11905 The filter accepts the following options:
11909 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
11912 Set number of steps for Gaussian approximation. Default is @code{1}.
11915 Set which planes to filter. By default all planes are filtered.
11918 Set vertical sigma, if negative it will be same as @code{sigma}.
11919 Default is @code{-1}.
11922 @subsection Commands
11923 This filter supports same commands as options.
11924 The command accepts the same syntax of the corresponding option.
11926 If the specified expression is not valid, it is kept at its current
11931 Apply generic equation to each pixel.
11933 The filter accepts the following options:
11936 @item lum_expr, lum
11937 Set the luminance expression.
11939 Set the chrominance blue expression.
11941 Set the chrominance red expression.
11942 @item alpha_expr, a
11943 Set the alpha expression.
11945 Set the red expression.
11946 @item green_expr, g
11947 Set the green expression.
11949 Set the blue expression.
11952 The colorspace is selected according to the specified options. If one
11953 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
11954 options is specified, the filter will automatically select a YCbCr
11955 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
11956 @option{blue_expr} options is specified, it will select an RGB
11959 If one of the chrominance expression is not defined, it falls back on the other
11960 one. If no alpha expression is specified it will evaluate to opaque value.
11961 If none of chrominance expressions are specified, they will evaluate
11962 to the luminance expression.
11964 The expressions can use the following variables and functions:
11968 The sequential number of the filtered frame, starting from @code{0}.
11972 The coordinates of the current sample.
11976 The width and height of the image.
11980 Width and height scale depending on the currently filtered plane. It is the
11981 ratio between the corresponding luma plane number of pixels and the current
11982 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
11983 @code{0.5,0.5} for chroma planes.
11986 Time of the current frame, expressed in seconds.
11989 Return the value of the pixel at location (@var{x},@var{y}) of the current
11993 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
11997 Return the value of the pixel at location (@var{x},@var{y}) of the
11998 blue-difference chroma plane. Return 0 if there is no such plane.
12001 Return the value of the pixel at location (@var{x},@var{y}) of the
12002 red-difference chroma plane. Return 0 if there is no such plane.
12007 Return the value of the pixel at location (@var{x},@var{y}) of the
12008 red/green/blue component. Return 0 if there is no such component.
12011 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12012 plane. Return 0 if there is no such plane.
12014 @item psum(x,y), lumsum(x, y), cbsum(x,y), crsum(x,y), rsum(x,y), gsum(x,y), bsum(x,y), alphasum(x,y)
12015 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12016 sums of samples within a rectangle. See the functions without the sum postfix.
12018 @item interpolation
12019 Set one of interpolation methods:
12024 Default is bilinear.
12027 For functions, if @var{x} and @var{y} are outside the area, the value will be
12028 automatically clipped to the closer edge.
12030 Please note that this filter can use multiple threads in which case each slice
12031 will have its own expression state. If you want to use only a single expression
12032 state because your expressions depend on previous state then you should limit
12033 the number of filter threads to 1.
12035 @subsection Examples
12039 Flip the image horizontally:
12045 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12046 wavelength of 100 pixels:
12048 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12052 Generate a fancy enigmatic moving light:
12054 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
12058 Generate a quick emboss effect:
12060 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12064 Modify RGB components depending on pixel position:
12066 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12070 Create a radial gradient that is the same size as the input (also see
12071 the @ref{vignette} filter):
12073 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12079 Fix the banding artifacts that are sometimes introduced into nearly flat
12080 regions by truncation to 8-bit color depth.
12081 Interpolate the gradients that should go where the bands are, and
12084 It is designed for playback only. Do not use it prior to
12085 lossy compression, because compression tends to lose the dither and
12086 bring back the bands.
12088 It accepts the following parameters:
12093 The maximum amount by which the filter will change any one pixel. This is also
12094 the threshold for detecting nearly flat regions. Acceptable values range from
12095 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12099 The neighborhood to fit the gradient to. A larger radius makes for smoother
12100 gradients, but also prevents the filter from modifying the pixels near detailed
12101 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12102 values will be clipped to the valid range.
12106 Alternatively, the options can be specified as a flat string:
12107 @var{strength}[:@var{radius}]
12109 @subsection Examples
12113 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12119 Specify radius, omitting the strength (which will fall-back to the default
12127 @anchor{graphmonitor}
12128 @section graphmonitor
12129 Show various filtergraph stats.
12131 With this filter one can debug complete filtergraph.
12132 Especially issues with links filling with queued frames.
12134 The filter accepts the following options:
12138 Set video output size. Default is @var{hd720}.
12141 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12144 Set output mode, can be @var{fulll} or @var{compact}.
12145 In @var{compact} mode only filters with some queued frames have displayed stats.
12148 Set flags which enable which stats are shown in video.
12150 Available values for flags are:
12153 Display number of queued frames in each link.
12155 @item frame_count_in
12156 Display number of frames taken from filter.
12158 @item frame_count_out
12159 Display number of frames given out from filter.
12162 Display current filtered frame pts.
12165 Display current filtered frame time.
12168 Display time base for filter link.
12171 Display used format for filter link.
12174 Display video size or number of audio channels in case of audio used by filter link.
12177 Display video frame rate or sample rate in case of audio used by filter link.
12180 Display link output status.
12184 Set upper limit for video rate of output stream, Default value is @var{25}.
12185 This guarantee that output video frame rate will not be higher than this value.
12189 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12190 and corrects the scene colors accordingly.
12192 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12194 The filter accepts the following options:
12198 The order of differentiation to be applied on the scene. Must be chosen in the range
12199 [0,2] and default value is 1.
12202 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12203 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12204 max value instead of calculating Minkowski distance.
12207 The standard deviation of Gaussian blur to be applied on the scene. Must be
12208 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12209 can't be equal to 0 if @var{difford} is greater than 0.
12212 @subsection Examples
12218 greyedge=difford=1:minknorm=5:sigma=2
12224 greyedge=difford=1:minknorm=0:sigma=2
12232 Apply a Hald CLUT to a video stream.
12234 First input is the video stream to process, and second one is the Hald CLUT.
12235 The Hald CLUT input can be a simple picture or a complete video stream.
12237 The filter accepts the following options:
12241 Force termination when the shortest input terminates. Default is @code{0}.
12243 Continue applying the last CLUT after the end of the stream. A value of
12244 @code{0} disable the filter after the last frame of the CLUT is reached.
12245 Default is @code{1}.
12248 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12249 filters share the same internals).
12251 This filter also supports the @ref{framesync} options.
12253 More information about the Hald CLUT can be found on Eskil Steenberg's website
12254 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12256 @subsection Workflow examples
12258 @subsubsection Hald CLUT video stream
12260 Generate an identity Hald CLUT stream altered with various effects:
12262 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
12265 Note: make sure you use a lossless codec.
12267 Then use it with @code{haldclut} to apply it on some random stream:
12269 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12272 The Hald CLUT will be applied to the 10 first seconds (duration of
12273 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12274 to the remaining frames of the @code{mandelbrot} stream.
12276 @subsubsection Hald CLUT with preview
12278 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12279 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12280 biggest possible square starting at the top left of the picture. The remaining
12281 padding pixels (bottom or right) will be ignored. This area can be used to add
12282 a preview of the Hald CLUT.
12284 Typically, the following generated Hald CLUT will be supported by the
12285 @code{haldclut} filter:
12288 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12289 pad=iw+320 [padded_clut];
12290 smptebars=s=320x256, split [a][b];
12291 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12292 [main][b] overlay=W-320" -frames:v 1 clut.png
12295 It contains the original and a preview of the effect of the CLUT: SMPTE color
12296 bars are displayed on the right-top, and below the same color bars processed by
12299 Then, the effect of this Hald CLUT can be visualized with:
12301 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12306 Flip the input video horizontally.
12308 For example, to horizontally flip the input video with @command{ffmpeg}:
12310 ffmpeg -i in.avi -vf "hflip" out.avi
12314 This filter applies a global color histogram equalization on a
12317 It can be used to correct video that has a compressed range of pixel
12318 intensities. The filter redistributes the pixel intensities to
12319 equalize their distribution across the intensity range. It may be
12320 viewed as an "automatically adjusting contrast filter". This filter is
12321 useful only for correcting degraded or poorly captured source
12324 The filter accepts the following options:
12328 Determine the amount of equalization to be applied. As the strength
12329 is reduced, the distribution of pixel intensities more-and-more
12330 approaches that of the input frame. The value must be a float number
12331 in the range [0,1] and defaults to 0.200.
12334 Set the maximum intensity that can generated and scale the output
12335 values appropriately. The strength should be set as desired and then
12336 the intensity can be limited if needed to avoid washing-out. The value
12337 must be a float number in the range [0,1] and defaults to 0.210.
12340 Set the antibanding level. If enabled the filter will randomly vary
12341 the luminance of output pixels by a small amount to avoid banding of
12342 the histogram. Possible values are @code{none}, @code{weak} or
12343 @code{strong}. It defaults to @code{none}.
12349 Compute and draw a color distribution histogram for the input video.
12351 The computed histogram is a representation of the color component
12352 distribution in an image.
12354 Standard histogram displays the color components distribution in an image.
12355 Displays color graph for each color component. Shows distribution of
12356 the Y, U, V, A or R, G, B components, depending on input format, in the
12357 current frame. Below each graph a color component scale meter is shown.
12359 The filter accepts the following options:
12363 Set height of level. Default value is @code{200}.
12364 Allowed range is [50, 2048].
12367 Set height of color scale. Default value is @code{12}.
12368 Allowed range is [0, 40].
12372 It accepts the following values:
12375 Per color component graphs are placed below each other.
12378 Per color component graphs are placed side by side.
12381 Presents information identical to that in the @code{parade}, except
12382 that the graphs representing color components are superimposed directly
12385 Default is @code{stack}.
12388 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12389 Default is @code{linear}.
12392 Set what color components to display.
12393 Default is @code{7}.
12396 Set foreground opacity. Default is @code{0.7}.
12399 Set background opacity. Default is @code{0.5}.
12402 @subsection Examples
12407 Calculate and draw histogram:
12409 ffplay -i input -vf histogram
12417 This is a high precision/quality 3d denoise filter. It aims to reduce
12418 image noise, producing smooth images and making still images really
12419 still. It should enhance compressibility.
12421 It accepts the following optional parameters:
12425 A non-negative floating point number which specifies spatial luma strength.
12426 It defaults to 4.0.
12428 @item chroma_spatial
12429 A non-negative floating point number which specifies spatial chroma strength.
12430 It defaults to 3.0*@var{luma_spatial}/4.0.
12433 A floating point number which specifies luma temporal strength. It defaults to
12434 6.0*@var{luma_spatial}/4.0.
12437 A floating point number which specifies chroma temporal strength. It defaults to
12438 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12441 @subsection Commands
12442 This filter supports same @ref{commands} as options.
12443 The command accepts the same syntax of the corresponding option.
12445 If the specified expression is not valid, it is kept at its current
12448 @anchor{hwdownload}
12449 @section hwdownload
12451 Download hardware frames to system memory.
12453 The input must be in hardware frames, and the output a non-hardware format.
12454 Not all formats will be supported on the output - it may be necessary to insert
12455 an additional @option{format} filter immediately following in the graph to get
12456 the output in a supported format.
12460 Map hardware frames to system memory or to another device.
12462 This filter has several different modes of operation; which one is used depends
12463 on the input and output formats:
12466 Hardware frame input, normal frame output
12468 Map the input frames to system memory and pass them to the output. If the
12469 original hardware frame is later required (for example, after overlaying
12470 something else on part of it), the @option{hwmap} filter can be used again
12471 in the next mode to retrieve it.
12473 Normal frame input, hardware frame output
12475 If the input is actually a software-mapped hardware frame, then unmap it -
12476 that is, return the original hardware frame.
12478 Otherwise, a device must be provided. Create new hardware surfaces on that
12479 device for the output, then map them back to the software format at the input
12480 and give those frames to the preceding filter. This will then act like the
12481 @option{hwupload} filter, but may be able to avoid an additional copy when
12482 the input is already in a compatible format.
12484 Hardware frame input and output
12486 A device must be supplied for the output, either directly or with the
12487 @option{derive_device} option. The input and output devices must be of
12488 different types and compatible - the exact meaning of this is
12489 system-dependent, but typically it means that they must refer to the same
12490 underlying hardware context (for example, refer to the same graphics card).
12492 If the input frames were originally created on the output device, then unmap
12493 to retrieve the original frames.
12495 Otherwise, map the frames to the output device - create new hardware frames
12496 on the output corresponding to the frames on the input.
12499 The following additional parameters are accepted:
12503 Set the frame mapping mode. Some combination of:
12506 The mapped frame should be readable.
12508 The mapped frame should be writeable.
12510 The mapping will always overwrite the entire frame.
12512 This may improve performance in some cases, as the original contents of the
12513 frame need not be loaded.
12515 The mapping must not involve any copying.
12517 Indirect mappings to copies of frames are created in some cases where either
12518 direct mapping is not possible or it would have unexpected properties.
12519 Setting this flag ensures that the mapping is direct and will fail if that is
12522 Defaults to @var{read+write} if not specified.
12524 @item derive_device @var{type}
12525 Rather than using the device supplied at initialisation, instead derive a new
12526 device of type @var{type} from the device the input frames exist on.
12529 In a hardware to hardware mapping, map in reverse - create frames in the sink
12530 and map them back to the source. This may be necessary in some cases where
12531 a mapping in one direction is required but only the opposite direction is
12532 supported by the devices being used.
12534 This option is dangerous - it may break the preceding filter in undefined
12535 ways if there are any additional constraints on that filter's output.
12536 Do not use it without fully understanding the implications of its use.
12542 Upload system memory frames to hardware surfaces.
12544 The device to upload to must be supplied when the filter is initialised. If
12545 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12546 option or with the @option{derive_device} option. The input and output devices
12547 must be of different types and compatible - the exact meaning of this is
12548 system-dependent, but typically it means that they must refer to the same
12549 underlying hardware context (for example, refer to the same graphics card).
12551 The following additional parameters are accepted:
12554 @item derive_device @var{type}
12555 Rather than using the device supplied at initialisation, instead derive a new
12556 device of type @var{type} from the device the input frames exist on.
12559 @anchor{hwupload_cuda}
12560 @section hwupload_cuda
12562 Upload system memory frames to a CUDA device.
12564 It accepts the following optional parameters:
12568 The number of the CUDA device to use
12573 Apply a high-quality magnification filter designed for pixel art. This filter
12574 was originally created by Maxim Stepin.
12576 It accepts the following option:
12580 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12581 @code{hq3x} and @code{4} for @code{hq4x}.
12582 Default is @code{3}.
12586 Stack input videos horizontally.
12588 All streams must be of same pixel format and of same height.
12590 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12591 to create same output.
12593 The filter accepts the following option:
12597 Set number of input streams. Default is 2.
12600 If set to 1, force the output to terminate when the shortest input
12601 terminates. Default value is 0.
12606 Modify the hue and/or the saturation of the input.
12608 It accepts the following parameters:
12612 Specify the hue angle as a number of degrees. It accepts an expression,
12613 and defaults to "0".
12616 Specify the saturation in the [-10,10] range. It accepts an expression and
12620 Specify the hue angle as a number of radians. It accepts an
12621 expression, and defaults to "0".
12624 Specify the brightness in the [-10,10] range. It accepts an expression and
12628 @option{h} and @option{H} are mutually exclusive, and can't be
12629 specified at the same time.
12631 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12632 expressions containing the following constants:
12636 frame count of the input frame starting from 0
12639 presentation timestamp of the input frame expressed in time base units
12642 frame rate of the input video, NAN if the input frame rate is unknown
12645 timestamp expressed in seconds, NAN if the input timestamp is unknown
12648 time base of the input video
12651 @subsection Examples
12655 Set the hue to 90 degrees and the saturation to 1.0:
12661 Same command but expressing the hue in radians:
12667 Rotate hue and make the saturation swing between 0
12668 and 2 over a period of 1 second:
12670 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12674 Apply a 3 seconds saturation fade-in effect starting at 0:
12676 hue="s=min(t/3\,1)"
12679 The general fade-in expression can be written as:
12681 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12685 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12687 hue="s=max(0\, min(1\, (8-t)/3))"
12690 The general fade-out expression can be written as:
12692 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12697 @subsection Commands
12699 This filter supports the following commands:
12705 Modify the hue and/or the saturation and/or brightness of the input video.
12706 The command accepts the same syntax of the corresponding option.
12708 If the specified expression is not valid, it is kept at its current
12712 @section hysteresis
12714 Grow first stream into second stream by connecting components.
12715 This makes it possible to build more robust edge masks.
12717 This filter accepts the following options:
12721 Set which planes will be processed as bitmap, unprocessed planes will be
12722 copied from first stream.
12723 By default value 0xf, all planes will be processed.
12726 Set threshold which is used in filtering. If pixel component value is higher than
12727 this value filter algorithm for connecting components is activated.
12728 By default value is 0.
12731 The @code{hysteresis} filter also supports the @ref{framesync} options.
12735 Detect video interlacing type.
12737 This filter tries to detect if the input frames are interlaced, progressive,
12738 top or bottom field first. It will also try to detect fields that are
12739 repeated between adjacent frames (a sign of telecine).
12741 Single frame detection considers only immediately adjacent frames when classifying each frame.
12742 Multiple frame detection incorporates the classification history of previous frames.
12744 The filter will log these metadata values:
12747 @item single.current_frame
12748 Detected type of current frame using single-frame detection. One of:
12749 ``tff'' (top field first), ``bff'' (bottom field first),
12750 ``progressive'', or ``undetermined''
12753 Cumulative number of frames detected as top field first using single-frame detection.
12756 Cumulative number of frames detected as top field first using multiple-frame detection.
12759 Cumulative number of frames detected as bottom field first using single-frame detection.
12761 @item multiple.current_frame
12762 Detected type of current frame using multiple-frame detection. One of:
12763 ``tff'' (top field first), ``bff'' (bottom field first),
12764 ``progressive'', or ``undetermined''
12767 Cumulative number of frames detected as bottom field first using multiple-frame detection.
12769 @item single.progressive
12770 Cumulative number of frames detected as progressive using single-frame detection.
12772 @item multiple.progressive
12773 Cumulative number of frames detected as progressive using multiple-frame detection.
12775 @item single.undetermined
12776 Cumulative number of frames that could not be classified using single-frame detection.
12778 @item multiple.undetermined
12779 Cumulative number of frames that could not be classified using multiple-frame detection.
12781 @item repeated.current_frame
12782 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
12784 @item repeated.neither
12785 Cumulative number of frames with no repeated field.
12788 Cumulative number of frames with the top field repeated from the previous frame's top field.
12790 @item repeated.bottom
12791 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
12794 The filter accepts the following options:
12798 Set interlacing threshold.
12800 Set progressive threshold.
12802 Threshold for repeated field detection.
12804 Number of frames after which a given frame's contribution to the
12805 statistics is halved (i.e., it contributes only 0.5 to its
12806 classification). The default of 0 means that all frames seen are given
12807 full weight of 1.0 forever.
12808 @item analyze_interlaced_flag
12809 When this is not 0 then idet will use the specified number of frames to determine
12810 if the interlaced flag is accurate, it will not count undetermined frames.
12811 If the flag is found to be accurate it will be used without any further
12812 computations, if it is found to be inaccurate it will be cleared without any
12813 further computations. This allows inserting the idet filter as a low computational
12814 method to clean up the interlaced flag
12819 Deinterleave or interleave fields.
12821 This filter allows one to process interlaced images fields without
12822 deinterlacing them. Deinterleaving splits the input frame into 2
12823 fields (so called half pictures). Odd lines are moved to the top
12824 half of the output image, even lines to the bottom half.
12825 You can process (filter) them independently and then re-interleave them.
12827 The filter accepts the following options:
12831 @item chroma_mode, c
12832 @item alpha_mode, a
12833 Available values for @var{luma_mode}, @var{chroma_mode} and
12834 @var{alpha_mode} are:
12840 @item deinterleave, d
12841 Deinterleave fields, placing one above the other.
12843 @item interleave, i
12844 Interleave fields. Reverse the effect of deinterleaving.
12846 Default value is @code{none}.
12848 @item luma_swap, ls
12849 @item chroma_swap, cs
12850 @item alpha_swap, as
12851 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
12854 @subsection Commands
12856 This filter supports the all above options as @ref{commands}.
12860 Apply inflate effect to the video.
12862 This filter replaces the pixel by the local(3x3) average by taking into account
12863 only values higher than the pixel.
12865 It accepts the following options:
12872 Limit the maximum change for each plane, default is 65535.
12873 If 0, plane will remain unchanged.
12876 @subsection Commands
12878 This filter supports the all above options as @ref{commands}.
12882 Simple interlacing filter from progressive contents. This interleaves upper (or
12883 lower) lines from odd frames with lower (or upper) lines from even frames,
12884 halving the frame rate and preserving image height.
12887 Original Original New Frame
12888 Frame 'j' Frame 'j+1' (tff)
12889 ========== =========== ==================
12890 Line 0 --------------------> Frame 'j' Line 0
12891 Line 1 Line 1 ----> Frame 'j+1' Line 1
12892 Line 2 ---------------------> Frame 'j' Line 2
12893 Line 3 Line 3 ----> Frame 'j+1' Line 3
12895 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
12898 It accepts the following optional parameters:
12902 This determines whether the interlaced frame is taken from the even
12903 (tff - default) or odd (bff) lines of the progressive frame.
12906 Vertical lowpass filter to avoid twitter interlacing and
12907 reduce moire patterns.
12911 Disable vertical lowpass filter
12914 Enable linear filter (default)
12917 Enable complex filter. This will slightly less reduce twitter and moire
12918 but better retain detail and subjective sharpness impression.
12925 Deinterlace input video by applying Donald Graft's adaptive kernel
12926 deinterling. Work on interlaced parts of a video to produce
12927 progressive frames.
12929 The description of the accepted parameters follows.
12933 Set the threshold which affects the filter's tolerance when
12934 determining if a pixel line must be processed. It must be an integer
12935 in the range [0,255] and defaults to 10. A value of 0 will result in
12936 applying the process on every pixels.
12939 Paint pixels exceeding the threshold value to white if set to 1.
12943 Set the fields order. Swap fields if set to 1, leave fields alone if
12947 Enable additional sharpening if set to 1. Default is 0.
12950 Enable twoway sharpening if set to 1. Default is 0.
12953 @subsection Examples
12957 Apply default values:
12959 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
12963 Enable additional sharpening:
12969 Paint processed pixels in white:
12977 Slowly update darker pixels.
12979 This filter makes short flashes of light appear longer.
12980 This filter accepts the following options:
12984 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
12987 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
12990 @section lenscorrection
12992 Correct radial lens distortion
12994 This filter can be used to correct for radial distortion as can result from the use
12995 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
12996 one can use tools available for example as part of opencv or simply trial-and-error.
12997 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
12998 and extract the k1 and k2 coefficients from the resulting matrix.
13000 Note that effectively the same filter is available in the open-source tools Krita and
13001 Digikam from the KDE project.
13003 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13004 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13005 brightness distribution, so you may want to use both filters together in certain
13006 cases, though you will have to take care of ordering, i.e. whether vignetting should
13007 be applied before or after lens correction.
13009 @subsection Options
13011 The filter accepts the following options:
13015 Relative x-coordinate of the focal point of the image, and thereby the center of the
13016 distortion. This value has a range [0,1] and is expressed as fractions of the image
13017 width. Default is 0.5.
13019 Relative y-coordinate of the focal point of the image, and thereby the center of the
13020 distortion. This value has a range [0,1] and is expressed as fractions of the image
13021 height. Default is 0.5.
13023 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13024 no correction. Default is 0.
13026 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13027 0 means no correction. Default is 0.
13030 The formula that generates the correction is:
13032 @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)
13034 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13035 distances from the focal point in the source and target images, respectively.
13039 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13041 The @code{lensfun} filter requires the camera make, camera model, and lens model
13042 to apply the lens correction. The filter will load the lensfun database and
13043 query it to find the corresponding camera and lens entries in the database. As
13044 long as these entries can be found with the given options, the filter can
13045 perform corrections on frames. Note that incomplete strings will result in the
13046 filter choosing the best match with the given options, and the filter will
13047 output the chosen camera and lens models (logged with level "info"). You must
13048 provide the make, camera model, and lens model as they are required.
13050 The filter accepts the following options:
13054 The make of the camera (for example, "Canon"). This option is required.
13057 The model of the camera (for example, "Canon EOS 100D"). This option is
13061 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13062 option is required.
13065 The type of correction to apply. The following values are valid options:
13069 Enables fixing lens vignetting.
13072 Enables fixing lens geometry. This is the default.
13075 Enables fixing chromatic aberrations.
13078 Enables fixing lens vignetting and lens geometry.
13081 Enables fixing lens vignetting and chromatic aberrations.
13084 Enables fixing both lens geometry and chromatic aberrations.
13087 Enables all possible corrections.
13091 The focal length of the image/video (zoom; expected constant for video). For
13092 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13093 range should be chosen when using that lens. Default 18.
13096 The aperture of the image/video (expected constant for video). Note that
13097 aperture is only used for vignetting correction. Default 3.5.
13099 @item focus_distance
13100 The focus distance of the image/video (expected constant for video). Note that
13101 focus distance is only used for vignetting and only slightly affects the
13102 vignetting correction process. If unknown, leave it at the default value (which
13106 The scale factor which is applied after transformation. After correction the
13107 video is no longer necessarily rectangular. This parameter controls how much of
13108 the resulting image is visible. The value 0 means that a value will be chosen
13109 automatically such that there is little or no unmapped area in the output
13110 image. 1.0 means that no additional scaling is done. Lower values may result
13111 in more of the corrected image being visible, while higher values may avoid
13112 unmapped areas in the output.
13114 @item target_geometry
13115 The target geometry of the output image/video. The following values are valid
13119 @item rectilinear (default)
13122 @item equirectangular
13123 @item fisheye_orthographic
13124 @item fisheye_stereographic
13125 @item fisheye_equisolid
13126 @item fisheye_thoby
13129 Apply the reverse of image correction (instead of correcting distortion, apply
13132 @item interpolation
13133 The type of interpolation used when correcting distortion. The following values
13138 @item linear (default)
13143 @subsection Examples
13147 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13148 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13152 ffmpeg -i input.mov -vf lensfun=make=Canon:model="Canon EOS 100D":lens_model="Canon EF-S 18-55mm f/3.5-5.6 IS STM":focal_length=18:aperture=8 -c:v h264 -b:v 8000k output.mov
13156 Apply the same as before, but only for the first 5 seconds of video.
13159 ffmpeg -i input.mov -vf lensfun=make=Canon:model="Canon EOS 100D":lens_model="Canon EF-S 18-55mm f/3.5-5.6 IS STM":focal_length=18:aperture=8:enable='lte(t\,5)' -c:v h264 -b:v 8000k output.mov
13166 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13167 score between two input videos.
13169 The obtained VMAF score is printed through the logging system.
13171 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13172 After installing the library it can be enabled using:
13173 @code{./configure --enable-libvmaf}.
13174 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13176 The filter has following options:
13180 Set the model path which is to be used for SVM.
13181 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13184 Set the file path to be used to store logs.
13187 Set the format of the log file (csv, json or xml).
13189 @item enable_transform
13190 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13191 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13192 Default value: @code{false}
13195 Invokes the phone model which will generate VMAF scores higher than in the
13196 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13197 Default value: @code{false}
13200 Enables computing psnr along with vmaf.
13201 Default value: @code{false}
13204 Enables computing ssim along with vmaf.
13205 Default value: @code{false}
13208 Enables computing ms_ssim along with vmaf.
13209 Default value: @code{false}
13212 Set the pool method to be used for computing vmaf.
13213 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13216 Set number of threads to be used when computing vmaf.
13217 Default value: @code{0}, which makes use of all available logical processors.
13220 Set interval for frame subsampling used when computing vmaf.
13221 Default value: @code{1}
13223 @item enable_conf_interval
13224 Enables confidence interval.
13225 Default value: @code{false}
13228 This filter also supports the @ref{framesync} options.
13230 @subsection Examples
13233 On the below examples the input file @file{main.mpg} being processed is
13234 compared with the reference file @file{ref.mpg}.
13237 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13241 Example with options:
13243 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13247 Example with options and different containers:
13249 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]libvmaf=psnr=1:log_fmt=json" -f null -
13255 Limits the pixel components values to the specified range [min, max].
13257 The filter accepts the following options:
13261 Lower bound. Defaults to the lowest allowed value for the input.
13264 Upper bound. Defaults to the highest allowed value for the input.
13267 Specify which planes will be processed. Defaults to all available.
13274 The filter accepts the following options:
13278 Set the number of loops. Setting this value to -1 will result in infinite loops.
13282 Set maximal size in number of frames. Default is 0.
13285 Set first frame of loop. Default is 0.
13288 @subsection Examples
13292 Loop single first frame infinitely:
13294 loop=loop=-1:size=1:start=0
13298 Loop single first frame 10 times:
13300 loop=loop=10:size=1:start=0
13304 Loop 10 first frames 5 times:
13306 loop=loop=5:size=10:start=0
13312 Apply a 1D LUT to an input video.
13314 The filter accepts the following options:
13318 Set the 1D LUT file name.
13320 Currently supported formats:
13329 Select interpolation mode.
13331 Available values are:
13335 Use values from the nearest defined point.
13337 Interpolate values using the linear interpolation.
13339 Interpolate values using the cosine interpolation.
13341 Interpolate values using the cubic interpolation.
13343 Interpolate values using the spline interpolation.
13350 Apply a 3D LUT to an input video.
13352 The filter accepts the following options:
13356 Set the 3D LUT file name.
13358 Currently supported formats:
13372 Select interpolation mode.
13374 Available values are:
13378 Use values from the nearest defined point.
13380 Interpolate values using the 8 points defining a cube.
13382 Interpolate values using a tetrahedron.
13388 Turn certain luma values into transparency.
13390 The filter accepts the following options:
13394 Set the luma which will be used as base for transparency.
13395 Default value is @code{0}.
13398 Set the range of luma values to be keyed out.
13399 Default value is @code{0.01}.
13402 Set the range of softness. Default value is @code{0}.
13403 Use this to control gradual transition from zero to full transparency.
13406 @subsection Commands
13407 This filter supports same @ref{commands} as options.
13408 The command accepts the same syntax of the corresponding option.
13410 If the specified expression is not valid, it is kept at its current
13413 @section lut, lutrgb, lutyuv
13415 Compute a look-up table for binding each pixel component input value
13416 to an output value, and apply it to the input video.
13418 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13419 to an RGB input video.
13421 These filters accept the following parameters:
13424 set first pixel component expression
13426 set second pixel component expression
13428 set third pixel component expression
13430 set fourth pixel component expression, corresponds to the alpha component
13433 set red component expression
13435 set green component expression
13437 set blue component expression
13439 alpha component expression
13442 set Y/luminance component expression
13444 set U/Cb component expression
13446 set V/Cr component expression
13449 Each of them specifies the expression to use for computing the lookup table for
13450 the corresponding pixel component values.
13452 The exact component associated to each of the @var{c*} options depends on the
13455 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13456 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13458 The expressions can contain the following constants and functions:
13463 The input width and height.
13466 The input value for the pixel component.
13469 The input value, clipped to the @var{minval}-@var{maxval} range.
13472 The maximum value for the pixel component.
13475 The minimum value for the pixel component.
13478 The negated value for the pixel component value, clipped to the
13479 @var{minval}-@var{maxval} range; it corresponds to the expression
13480 "maxval-clipval+minval".
13483 The computed value in @var{val}, clipped to the
13484 @var{minval}-@var{maxval} range.
13486 @item gammaval(gamma)
13487 The computed gamma correction value of the pixel component value,
13488 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13490 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13494 All expressions default to "val".
13496 @subsection Examples
13500 Negate input video:
13502 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13503 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13506 The above is the same as:
13508 lutrgb="r=negval:g=negval:b=negval"
13509 lutyuv="y=negval:u=negval:v=negval"
13519 Remove chroma components, turning the video into a graytone image:
13521 lutyuv="u=128:v=128"
13525 Apply a luma burning effect:
13531 Remove green and blue components:
13537 Set a constant alpha channel value on input:
13539 format=rgba,lutrgb=a="maxval-minval/2"
13543 Correct luminance gamma by a factor of 0.5:
13545 lutyuv=y=gammaval(0.5)
13549 Discard least significant bits of luma:
13551 lutyuv=y='bitand(val, 128+64+32)'
13555 Technicolor like effect:
13557 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13561 @section lut2, tlut2
13563 The @code{lut2} filter takes two input streams and outputs one
13566 The @code{tlut2} (time lut2) filter takes two consecutive frames
13567 from one single stream.
13569 This filter accepts the following parameters:
13572 set first pixel component expression
13574 set second pixel component expression
13576 set third pixel component expression
13578 set fourth pixel component expression, corresponds to the alpha component
13581 set output bit depth, only available for @code{lut2} filter. By default is 0,
13582 which means bit depth is automatically picked from first input format.
13585 The @code{lut2} filter also supports the @ref{framesync} options.
13587 Each of them specifies the expression to use for computing the lookup table for
13588 the corresponding pixel component values.
13590 The exact component associated to each of the @var{c*} options depends on the
13593 The expressions can contain the following constants:
13598 The input width and height.
13601 The first input value for the pixel component.
13604 The second input value for the pixel component.
13607 The first input video bit depth.
13610 The second input video bit depth.
13613 All expressions default to "x".
13615 @subsection Examples
13619 Highlight differences between two RGB video streams:
13621 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)'
13625 Highlight differences between two YUV video streams:
13627 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)'
13631 Show max difference between two video streams:
13633 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)))'
13637 @section maskedclamp
13639 Clamp the first input stream with the second input and third input stream.
13641 Returns the value of first stream to be between second input
13642 stream - @code{undershoot} and third input stream + @code{overshoot}.
13644 This filter accepts the following options:
13647 Default value is @code{0}.
13650 Default value is @code{0}.
13653 Set which planes will be processed as bitmap, unprocessed planes will be
13654 copied from first stream.
13655 By default value 0xf, all planes will be processed.
13660 Merge the second and third input stream into output stream using absolute differences
13661 between second input stream and first input stream and absolute difference between
13662 third input stream and first input stream. The picked value will be from second input
13663 stream if second absolute difference is greater than first one or from third input stream
13666 This filter accepts the following options:
13669 Set which planes will be processed as bitmap, unprocessed planes will be
13670 copied from first stream.
13671 By default value 0xf, all planes will be processed.
13674 @section maskedmerge
13676 Merge the first input stream with the second input stream using per pixel
13677 weights in the third input stream.
13679 A value of 0 in the third stream pixel component means that pixel component
13680 from first stream is returned unchanged, while maximum value (eg. 255 for
13681 8-bit videos) means that pixel component from second stream is returned
13682 unchanged. Intermediate values define the amount of merging between both
13683 input stream's pixel components.
13685 This filter accepts the following options:
13688 Set which planes will be processed as bitmap, unprocessed planes will be
13689 copied from first stream.
13690 By default value 0xf, all planes will be processed.
13695 Merge the second and third input stream into output stream using absolute differences
13696 between second input stream and first input stream and absolute difference between
13697 third input stream and first input stream. The picked value will be from second input
13698 stream if second absolute difference is less than first one or from third input stream
13701 This filter accepts the following options:
13704 Set which planes will be processed as bitmap, unprocessed planes will be
13705 copied from first stream.
13706 By default value 0xf, all planes will be processed.
13709 @section maskedthreshold
13710 Pick pixels comparing absolute difference of two video streams with fixed
13713 If absolute difference between pixel component of first and second video
13714 stream is equal or lower than user supplied threshold than pixel component
13715 from first video stream is picked, otherwise pixel component from second
13716 video stream is picked.
13718 This filter accepts the following options:
13721 Set threshold used when picking pixels from absolute difference from two input
13725 Set which planes will be processed as bitmap, unprocessed planes will be
13726 copied from second stream.
13727 By default value 0xf, all planes will be processed.
13731 Create mask from input video.
13733 For example it is useful to create motion masks after @code{tblend} filter.
13735 This filter accepts the following options:
13739 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
13742 Set high threshold. Any pixel component higher than this value will be set to max value
13743 allowed for current pixel format.
13746 Set planes to filter, by default all available planes are filtered.
13749 Fill all frame pixels with this value.
13752 Set max average pixel value for frame. If sum of all pixel components is higher that this
13753 average, output frame will be completely filled with value set by @var{fill} option.
13754 Typically useful for scene changes when used in combination with @code{tblend} filter.
13759 Apply motion-compensation deinterlacing.
13761 It needs one field per frame as input and must thus be used together
13762 with yadif=1/3 or equivalent.
13764 This filter accepts the following options:
13767 Set the deinterlacing mode.
13769 It accepts one of the following values:
13774 use iterative motion estimation
13776 like @samp{slow}, but use multiple reference frames.
13778 Default value is @samp{fast}.
13781 Set the picture field parity assumed for the input video. It must be
13782 one of the following values:
13786 assume top field first
13788 assume bottom field first
13791 Default value is @samp{bff}.
13794 Set per-block quantization parameter (QP) used by the internal
13797 Higher values should result in a smoother motion vector field but less
13798 optimal individual vectors. Default value is 1.
13803 Pick median pixel from certain rectangle defined by radius.
13805 This filter accepts the following options:
13809 Set horizontal radius size. Default value is @code{1}.
13810 Allowed range is integer from 1 to 127.
13813 Set which planes to process. Default is @code{15}, which is all available planes.
13816 Set vertical radius size. Default value is @code{0}.
13817 Allowed range is integer from 0 to 127.
13818 If it is 0, value will be picked from horizontal @code{radius} option.
13821 Set median percentile. Default value is @code{0.5}.
13822 Default value of @code{0.5} will pick always median values, while @code{0} will pick
13823 minimum values, and @code{1} maximum values.
13826 @subsection Commands
13827 This filter supports same @ref{commands} as options.
13828 The command accepts the same syntax of the corresponding option.
13830 If the specified expression is not valid, it is kept at its current
13833 @section mergeplanes
13835 Merge color channel components from several video streams.
13837 The filter accepts up to 4 input streams, and merge selected input
13838 planes to the output video.
13840 This filter accepts the following options:
13843 Set input to output plane mapping. Default is @code{0}.
13845 The mappings is specified as a bitmap. It should be specified as a
13846 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
13847 mapping for the first plane of the output stream. 'A' sets the number of
13848 the input stream to use (from 0 to 3), and 'a' the plane number of the
13849 corresponding input to use (from 0 to 3). The rest of the mappings is
13850 similar, 'Bb' describes the mapping for the output stream second
13851 plane, 'Cc' describes the mapping for the output stream third plane and
13852 'Dd' describes the mapping for the output stream fourth plane.
13855 Set output pixel format. Default is @code{yuva444p}.
13858 @subsection Examples
13862 Merge three gray video streams of same width and height into single video stream:
13864 [a0][a1][a2]mergeplanes=0x001020:yuv444p
13868 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
13870 [a0][a1]mergeplanes=0x00010210:yuva444p
13874 Swap Y and A plane in yuva444p stream:
13876 format=yuva444p,mergeplanes=0x03010200:yuva444p
13880 Swap U and V plane in yuv420p stream:
13882 format=yuv420p,mergeplanes=0x000201:yuv420p
13886 Cast a rgb24 clip to yuv444p:
13888 format=rgb24,mergeplanes=0x000102:yuv444p
13894 Estimate and export motion vectors using block matching algorithms.
13895 Motion vectors are stored in frame side data to be used by other filters.
13897 This filter accepts the following options:
13900 Specify the motion estimation method. Accepts one of the following values:
13904 Exhaustive search algorithm.
13906 Three step search algorithm.
13908 Two dimensional logarithmic search algorithm.
13910 New three step search algorithm.
13912 Four step search algorithm.
13914 Diamond search algorithm.
13916 Hexagon-based search algorithm.
13918 Enhanced predictive zonal search algorithm.
13920 Uneven multi-hexagon search algorithm.
13922 Default value is @samp{esa}.
13925 Macroblock size. Default @code{16}.
13928 Search parameter. Default @code{7}.
13931 @section midequalizer
13933 Apply Midway Image Equalization effect using two video streams.
13935 Midway Image Equalization adjusts a pair of images to have the same
13936 histogram, while maintaining their dynamics as much as possible. It's
13937 useful for e.g. matching exposures from a pair of stereo cameras.
13939 This filter has two inputs and one output, which must be of same pixel format, but
13940 may be of different sizes. The output of filter is first input adjusted with
13941 midway histogram of both inputs.
13943 This filter accepts the following option:
13947 Set which planes to process. Default is @code{15}, which is all available planes.
13950 @section minterpolate
13952 Convert the video to specified frame rate using motion interpolation.
13954 This filter accepts the following options:
13957 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}.
13960 Motion interpolation mode. Following values are accepted:
13963 Duplicate previous or next frame for interpolating new ones.
13965 Blend source frames. Interpolated frame is mean of previous and next frames.
13967 Motion compensated interpolation. Following options are effective when this mode is selected:
13971 Motion compensation mode. Following values are accepted:
13974 Overlapped block motion compensation.
13976 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
13978 Default mode is @samp{obmc}.
13981 Motion estimation mode. Following values are accepted:
13984 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
13986 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
13988 Default mode is @samp{bilat}.
13991 The algorithm to be used for motion estimation. Following values are accepted:
13994 Exhaustive search algorithm.
13996 Three step search algorithm.
13998 Two dimensional logarithmic search algorithm.
14000 New three step search algorithm.
14002 Four step search algorithm.
14004 Diamond search algorithm.
14006 Hexagon-based search algorithm.
14008 Enhanced predictive zonal search algorithm.
14010 Uneven multi-hexagon search algorithm.
14012 Default algorithm is @samp{epzs}.
14015 Macroblock size. Default @code{16}.
14018 Motion estimation search parameter. Default @code{32}.
14021 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).
14026 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:
14029 Disable scene change detection.
14031 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14033 Default method is @samp{fdiff}.
14035 @item scd_threshold
14036 Scene change detection threshold. Default is @code{10.}.
14041 Mix several video input streams into one video stream.
14043 A description of the accepted options follows.
14047 The number of inputs. If unspecified, it defaults to 2.
14050 Specify weight of each input video stream as sequence.
14051 Each weight is separated by space. If number of weights
14052 is smaller than number of @var{frames} last specified
14053 weight will be used for all remaining unset weights.
14056 Specify scale, if it is set it will be multiplied with sum
14057 of each weight multiplied with pixel values to give final destination
14058 pixel value. By default @var{scale} is auto scaled to sum of weights.
14061 Specify how end of stream is determined.
14064 The duration of the longest input. (default)
14067 The duration of the shortest input.
14070 The duration of the first input.
14074 @section mpdecimate
14076 Drop frames that do not differ greatly from the previous frame in
14077 order to reduce frame rate.
14079 The main use of this filter is for very-low-bitrate encoding
14080 (e.g. streaming over dialup modem), but it could in theory be used for
14081 fixing movies that were inverse-telecined incorrectly.
14083 A description of the accepted options follows.
14087 Set the maximum number of consecutive frames which can be dropped (if
14088 positive), or the minimum interval between dropped frames (if
14089 negative). If the value is 0, the frame is dropped disregarding the
14090 number of previous sequentially dropped frames.
14092 Default value is 0.
14097 Set the dropping threshold values.
14099 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14100 represent actual pixel value differences, so a threshold of 64
14101 corresponds to 1 unit of difference for each pixel, or the same spread
14102 out differently over the block.
14104 A frame is a candidate for dropping if no 8x8 blocks differ by more
14105 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14106 meaning the whole image) differ by more than a threshold of @option{lo}.
14108 Default value for @option{hi} is 64*12, default value for @option{lo} is
14109 64*5, and default value for @option{frac} is 0.33.
14115 Negate (invert) the input video.
14117 It accepts the following option:
14122 With value 1, it negates the alpha component, if present. Default value is 0.
14128 Denoise frames using Non-Local Means algorithm.
14130 Each pixel is adjusted by looking for other pixels with similar contexts. This
14131 context similarity is defined by comparing their surrounding patches of size
14132 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14135 Note that the research area defines centers for patches, which means some
14136 patches will be made of pixels outside that research area.
14138 The filter accepts the following options.
14142 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14145 Set patch size. Default is 7. Must be odd number in range [0, 99].
14148 Same as @option{p} but for chroma planes.
14150 The default value is @var{0} and means automatic.
14153 Set research size. Default is 15. Must be odd number in range [0, 99].
14156 Same as @option{r} but for chroma planes.
14158 The default value is @var{0} and means automatic.
14163 Deinterlace video using neural network edge directed interpolation.
14165 This filter accepts the following options:
14169 Mandatory option, without binary file filter can not work.
14170 Currently file can be found here:
14171 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14174 Set which frames to deinterlace, by default it is @code{all}.
14175 Can be @code{all} or @code{interlaced}.
14178 Set mode of operation.
14180 Can be one of the following:
14184 Use frame flags, both fields.
14186 Use frame flags, single field.
14188 Use top field only.
14190 Use bottom field only.
14192 Use both fields, top first.
14194 Use both fields, bottom first.
14198 Set which planes to process, by default filter process all frames.
14201 Set size of local neighborhood around each pixel, used by the predictor neural
14204 Can be one of the following:
14217 Set the number of neurons in predictor neural network.
14218 Can be one of the following:
14229 Controls the number of different neural network predictions that are blended
14230 together to compute the final output value. Can be @code{fast}, default or
14234 Set which set of weights to use in the predictor.
14235 Can be one of the following:
14239 weights trained to minimize absolute error
14241 weights trained to minimize squared error
14245 Controls whether or not the prescreener neural network is used to decide
14246 which pixels should be processed by the predictor neural network and which
14247 can be handled by simple cubic interpolation.
14248 The prescreener is trained to know whether cubic interpolation will be
14249 sufficient for a pixel or whether it should be predicted by the predictor nn.
14250 The computational complexity of the prescreener nn is much less than that of
14251 the predictor nn. Since most pixels can be handled by cubic interpolation,
14252 using the prescreener generally results in much faster processing.
14253 The prescreener is pretty accurate, so the difference between using it and not
14254 using it is almost always unnoticeable.
14256 Can be one of the following:
14264 Default is @code{new}.
14267 Set various debugging flags.
14272 Force libavfilter not to use any of the specified pixel formats for the
14273 input to the next filter.
14275 It accepts the following parameters:
14279 A '|'-separated list of pixel format names, such as
14280 pix_fmts=yuv420p|monow|rgb24".
14284 @subsection Examples
14288 Force libavfilter to use a format different from @var{yuv420p} for the
14289 input to the vflip filter:
14291 noformat=pix_fmts=yuv420p,vflip
14295 Convert the input video to any of the formats not contained in the list:
14297 noformat=yuv420p|yuv444p|yuv410p
14303 Add noise on video input frame.
14305 The filter accepts the following options:
14313 Set noise seed for specific pixel component or all pixel components in case
14314 of @var{all_seed}. Default value is @code{123457}.
14316 @item all_strength, alls
14317 @item c0_strength, c0s
14318 @item c1_strength, c1s
14319 @item c2_strength, c2s
14320 @item c3_strength, c3s
14321 Set noise strength for specific pixel component or all pixel components in case
14322 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14324 @item all_flags, allf
14325 @item c0_flags, c0f
14326 @item c1_flags, c1f
14327 @item c2_flags, c2f
14328 @item c3_flags, c3f
14329 Set pixel component flags or set flags for all components if @var{all_flags}.
14330 Available values for component flags are:
14333 averaged temporal noise (smoother)
14335 mix random noise with a (semi)regular pattern
14337 temporal noise (noise pattern changes between frames)
14339 uniform noise (gaussian otherwise)
14343 @subsection Examples
14345 Add temporal and uniform noise to input video:
14347 noise=alls=20:allf=t+u
14352 Normalize RGB video (aka histogram stretching, contrast stretching).
14353 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14355 For each channel of each frame, the filter computes the input range and maps
14356 it linearly to the user-specified output range. The output range defaults
14357 to the full dynamic range from pure black to pure white.
14359 Temporal smoothing can be used on the input range to reduce flickering (rapid
14360 changes in brightness) caused when small dark or bright objects enter or leave
14361 the scene. This is similar to the auto-exposure (automatic gain control) on a
14362 video camera, and, like a video camera, it may cause a period of over- or
14363 under-exposure of the video.
14365 The R,G,B channels can be normalized independently, which may cause some
14366 color shifting, or linked together as a single channel, which prevents
14367 color shifting. Linked normalization preserves hue. Independent normalization
14368 does not, so it can be used to remove some color casts. Independent and linked
14369 normalization can be combined in any ratio.
14371 The normalize filter accepts the following options:
14376 Colors which define the output range. The minimum input value is mapped to
14377 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14378 The defaults are black and white respectively. Specifying white for
14379 @var{blackpt} and black for @var{whitept} will give color-inverted,
14380 normalized video. Shades of grey can be used to reduce the dynamic range
14381 (contrast). Specifying saturated colors here can create some interesting
14385 The number of previous frames to use for temporal smoothing. The input range
14386 of each channel is smoothed using a rolling average over the current frame
14387 and the @var{smoothing} previous frames. The default is 0 (no temporal
14391 Controls the ratio of independent (color shifting) channel normalization to
14392 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14393 independent. Defaults to 1.0 (fully independent).
14396 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14397 expensive no-op. Defaults to 1.0 (full strength).
14401 @subsection Commands
14402 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14403 The command accepts the same syntax of the corresponding option.
14405 If the specified expression is not valid, it is kept at its current
14408 @subsection Examples
14410 Stretch video contrast to use the full dynamic range, with no temporal
14411 smoothing; may flicker depending on the source content:
14413 normalize=blackpt=black:whitept=white:smoothing=0
14416 As above, but with 50 frames of temporal smoothing; flicker should be
14417 reduced, depending on the source content:
14419 normalize=blackpt=black:whitept=white:smoothing=50
14422 As above, but with hue-preserving linked channel normalization:
14424 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14427 As above, but with half strength:
14429 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14432 Map the darkest input color to red, the brightest input color to cyan:
14434 normalize=blackpt=red:whitept=cyan
14439 Pass the video source unchanged to the output.
14442 Optical Character Recognition
14444 This filter uses Tesseract for optical character recognition. To enable
14445 compilation of this filter, you need to configure FFmpeg with
14446 @code{--enable-libtesseract}.
14448 It accepts the following options:
14452 Set datapath to tesseract data. Default is to use whatever was
14453 set at installation.
14456 Set language, default is "eng".
14459 Set character whitelist.
14462 Set character blacklist.
14465 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14466 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14470 Apply a video transform using libopencv.
14472 To enable this filter, install the libopencv library and headers and
14473 configure FFmpeg with @code{--enable-libopencv}.
14475 It accepts the following parameters:
14480 The name of the libopencv filter to apply.
14482 @item filter_params
14483 The parameters to pass to the libopencv filter. If not specified, the default
14484 values are assumed.
14488 Refer to the official libopencv documentation for more precise
14490 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14492 Several libopencv filters are supported; see the following subsections.
14497 Dilate an image by using a specific structuring element.
14498 It corresponds to the libopencv function @code{cvDilate}.
14500 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14502 @var{struct_el} represents a structuring element, and has the syntax:
14503 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14505 @var{cols} and @var{rows} represent the number of columns and rows of
14506 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14507 point, and @var{shape} the shape for the structuring element. @var{shape}
14508 must be "rect", "cross", "ellipse", or "custom".
14510 If the value for @var{shape} is "custom", it must be followed by a
14511 string of the form "=@var{filename}". The file with name
14512 @var{filename} is assumed to represent a binary image, with each
14513 printable character corresponding to a bright pixel. When a custom
14514 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14515 or columns and rows of the read file are assumed instead.
14517 The default value for @var{struct_el} is "3x3+0x0/rect".
14519 @var{nb_iterations} specifies the number of times the transform is
14520 applied to the image, and defaults to 1.
14524 # Use the default values
14527 # Dilate using a structuring element with a 5x5 cross, iterating two times
14528 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14530 # Read the shape from the file diamond.shape, iterating two times.
14531 # The file diamond.shape may contain a pattern of characters like this
14537 # The specified columns and rows are ignored
14538 # but the anchor point coordinates are not
14539 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14544 Erode an image by using a specific structuring element.
14545 It corresponds to the libopencv function @code{cvErode}.
14547 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14548 with the same syntax and semantics as the @ref{dilate} filter.
14552 Smooth the input video.
14554 The filter takes the following parameters:
14555 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14557 @var{type} is the type of smooth filter to apply, and must be one of
14558 the following values: "blur", "blur_no_scale", "median", "gaussian",
14559 or "bilateral". The default value is "gaussian".
14561 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14562 depends on the smooth type. @var{param1} and
14563 @var{param2} accept integer positive values or 0. @var{param3} and
14564 @var{param4} accept floating point values.
14566 The default value for @var{param1} is 3. The default value for the
14567 other parameters is 0.
14569 These parameters correspond to the parameters assigned to the
14570 libopencv function @code{cvSmooth}.
14572 @section oscilloscope
14574 2D Video Oscilloscope.
14576 Useful to measure spatial impulse, step responses, chroma delays, etc.
14578 It accepts the following parameters:
14582 Set scope center x position.
14585 Set scope center y position.
14588 Set scope size, relative to frame diagonal.
14591 Set scope tilt/rotation.
14597 Set trace center x position.
14600 Set trace center y position.
14603 Set trace width, relative to width of frame.
14606 Set trace height, relative to height of frame.
14609 Set which components to trace. By default it traces first three components.
14612 Draw trace grid. By default is enabled.
14615 Draw some statistics. By default is enabled.
14618 Draw scope. By default is enabled.
14621 @subsection Commands
14622 This filter supports same @ref{commands} as options.
14623 The command accepts the same syntax of the corresponding option.
14625 If the specified expression is not valid, it is kept at its current
14628 @subsection Examples
14632 Inspect full first row of video frame.
14634 oscilloscope=x=0.5:y=0:s=1
14638 Inspect full last row of video frame.
14640 oscilloscope=x=0.5:y=1:s=1
14644 Inspect full 5th line of video frame of height 1080.
14646 oscilloscope=x=0.5:y=5/1080:s=1
14650 Inspect full last column of video frame.
14652 oscilloscope=x=1:y=0.5:s=1:t=1
14660 Overlay one video on top of another.
14662 It takes two inputs and has one output. The first input is the "main"
14663 video on which the second input is overlaid.
14665 It accepts the following parameters:
14667 A description of the accepted options follows.
14672 Set the expression for the x and y coordinates of the overlaid video
14673 on the main video. Default value is "0" for both expressions. In case
14674 the expression is invalid, it is set to a huge value (meaning that the
14675 overlay will not be displayed within the output visible area).
14678 See @ref{framesync}.
14681 Set when the expressions for @option{x}, and @option{y} are evaluated.
14683 It accepts the following values:
14686 only evaluate expressions once during the filter initialization or
14687 when a command is processed
14690 evaluate expressions for each incoming frame
14693 Default value is @samp{frame}.
14696 See @ref{framesync}.
14699 Set the format for the output video.
14701 It accepts the following values:
14704 force YUV420 output
14707 force YUV420p10 output
14710 force YUV422 output
14713 force YUV422p10 output
14716 force YUV444 output
14719 force packed RGB output
14722 force planar RGB output
14725 automatically pick format
14728 Default value is @samp{yuv420}.
14731 See @ref{framesync}.
14734 Set format of alpha of the overlaid video, it can be @var{straight} or
14735 @var{premultiplied}. Default is @var{straight}.
14738 The @option{x}, and @option{y} expressions can contain the following
14744 The main input width and height.
14748 The overlay input width and height.
14752 The computed values for @var{x} and @var{y}. They are evaluated for
14757 horizontal and vertical chroma subsample values of the output
14758 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
14762 the number of input frame, starting from 0
14765 the position in the file of the input frame, NAN if unknown
14768 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
14772 This filter also supports the @ref{framesync} options.
14774 Note that the @var{n}, @var{pos}, @var{t} variables are available only
14775 when evaluation is done @emph{per frame}, and will evaluate to NAN
14776 when @option{eval} is set to @samp{init}.
14778 Be aware that frames are taken from each input video in timestamp
14779 order, hence, if their initial timestamps differ, it is a good idea
14780 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
14781 have them begin in the same zero timestamp, as the example for
14782 the @var{movie} filter does.
14784 You can chain together more overlays but you should test the
14785 efficiency of such approach.
14787 @subsection Commands
14789 This filter supports the following commands:
14793 Modify the x and y of the overlay input.
14794 The command accepts the same syntax of the corresponding option.
14796 If the specified expression is not valid, it is kept at its current
14800 @subsection Examples
14804 Draw the overlay at 10 pixels from the bottom right corner of the main
14807 overlay=main_w-overlay_w-10:main_h-overlay_h-10
14810 Using named options the example above becomes:
14812 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
14816 Insert a transparent PNG logo in the bottom left corner of the input,
14817 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
14819 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
14823 Insert 2 different transparent PNG logos (second logo on bottom
14824 right corner) using the @command{ffmpeg} tool:
14826 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
14830 Add a transparent color layer on top of the main video; @code{WxH}
14831 must specify the size of the main input to the overlay filter:
14833 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
14837 Play an original video and a filtered version (here with the deshake
14838 filter) side by side using the @command{ffplay} tool:
14840 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
14843 The above command is the same as:
14845 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
14849 Make a sliding overlay appearing from the left to the right top part of the
14850 screen starting since time 2:
14852 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
14856 Compose output by putting two input videos side to side:
14858 ffmpeg -i left.avi -i right.avi -filter_complex "
14859 nullsrc=size=200x100 [background];
14860 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
14861 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
14862 [background][left] overlay=shortest=1 [background+left];
14863 [background+left][right] overlay=shortest=1:x=100 [left+right]
14868 Mask 10-20 seconds of a video by applying the delogo filter to a section
14870 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
14871 -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]'
14876 Chain several overlays in cascade:
14878 nullsrc=s=200x200 [bg];
14879 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
14880 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
14881 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
14882 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
14883 [in3] null, [mid2] overlay=100:100 [out0]
14888 @anchor{overlay_cuda}
14889 @section overlay_cuda
14891 Overlay one video on top of another.
14893 This is the CUDA variant of the @ref{overlay} filter.
14894 It only accepts CUDA frames. The underlying input pixel formats have to match.
14896 It takes two inputs and has one output. The first input is the "main"
14897 video on which the second input is overlaid.
14899 It accepts the following parameters:
14904 Set the x and y coordinates of the overlaid video on the main video.
14905 Default value is "0" for both expressions.
14908 See @ref{framesync}.
14911 See @ref{framesync}.
14914 See @ref{framesync}.
14918 This filter also supports the @ref{framesync} options.
14922 Apply Overcomplete Wavelet denoiser.
14924 The filter accepts the following options:
14930 Larger depth values will denoise lower frequency components more, but
14931 slow down filtering.
14933 Must be an int in the range 8-16, default is @code{8}.
14935 @item luma_strength, ls
14938 Must be a double value in the range 0-1000, default is @code{1.0}.
14940 @item chroma_strength, cs
14941 Set chroma strength.
14943 Must be a double value in the range 0-1000, default is @code{1.0}.
14949 Add paddings to the input image, and place the original input at the
14950 provided @var{x}, @var{y} coordinates.
14952 It accepts the following parameters:
14957 Specify an expression for the size of the output image with the
14958 paddings added. If the value for @var{width} or @var{height} is 0, the
14959 corresponding input size is used for the output.
14961 The @var{width} expression can reference the value set by the
14962 @var{height} expression, and vice versa.
14964 The default value of @var{width} and @var{height} is 0.
14968 Specify the offsets to place the input image at within the padded area,
14969 with respect to the top/left border of the output image.
14971 The @var{x} expression can reference the value set by the @var{y}
14972 expression, and vice versa.
14974 The default value of @var{x} and @var{y} is 0.
14976 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
14977 so the input image is centered on the padded area.
14980 Specify the color of the padded area. For the syntax of this option,
14981 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
14982 manual,ffmpeg-utils}.
14984 The default value of @var{color} is "black".
14987 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
14989 It accepts the following values:
14993 Only evaluate expressions once during the filter initialization or when
14994 a command is processed.
14997 Evaluate expressions for each incoming frame.
15001 Default value is @samp{init}.
15004 Pad to aspect instead to a resolution.
15008 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15009 options are expressions containing the following constants:
15014 The input video width and height.
15018 These are the same as @var{in_w} and @var{in_h}.
15022 The output width and height (the size of the padded area), as
15023 specified by the @var{width} and @var{height} expressions.
15027 These are the same as @var{out_w} and @var{out_h}.
15031 The x and y offsets as specified by the @var{x} and @var{y}
15032 expressions, or NAN if not yet specified.
15035 same as @var{iw} / @var{ih}
15038 input sample aspect ratio
15041 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15045 The horizontal and vertical chroma subsample values. For example for the
15046 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15049 @subsection Examples
15053 Add paddings with the color "violet" to the input video. The output video
15054 size is 640x480, and the top-left corner of the input video is placed at
15057 pad=640:480:0:40:violet
15060 The example above is equivalent to the following command:
15062 pad=width=640:height=480:x=0:y=40:color=violet
15066 Pad the input to get an output with dimensions increased by 3/2,
15067 and put the input video at the center of the padded area:
15069 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15073 Pad the input to get a squared output with size equal to the maximum
15074 value between the input width and height, and put the input video at
15075 the center of the padded area:
15077 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15081 Pad the input to get a final w/h ratio of 16:9:
15083 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15087 In case of anamorphic video, in order to set the output display aspect
15088 correctly, it is necessary to use @var{sar} in the expression,
15089 according to the relation:
15091 (ih * X / ih) * sar = output_dar
15092 X = output_dar / sar
15095 Thus the previous example needs to be modified to:
15097 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15101 Double the output size and put the input video in the bottom-right
15102 corner of the output padded area:
15104 pad="2*iw:2*ih:ow-iw:oh-ih"
15108 @anchor{palettegen}
15109 @section palettegen
15111 Generate one palette for a whole video stream.
15113 It accepts the following options:
15117 Set the maximum number of colors to quantize in the palette.
15118 Note: the palette will still contain 256 colors; the unused palette entries
15121 @item reserve_transparent
15122 Create a palette of 255 colors maximum and reserve the last one for
15123 transparency. Reserving the transparency color is useful for GIF optimization.
15124 If not set, the maximum of colors in the palette will be 256. You probably want
15125 to disable this option for a standalone image.
15128 @item transparency_color
15129 Set the color that will be used as background for transparency.
15132 Set statistics mode.
15134 It accepts the following values:
15137 Compute full frame histograms.
15139 Compute histograms only for the part that differs from previous frame. This
15140 might be relevant to give more importance to the moving part of your input if
15141 the background is static.
15143 Compute new histogram for each frame.
15146 Default value is @var{full}.
15149 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15150 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15151 color quantization of the palette. This information is also visible at
15152 @var{info} logging level.
15154 @subsection Examples
15158 Generate a representative palette of a given video using @command{ffmpeg}:
15160 ffmpeg -i input.mkv -vf palettegen palette.png
15164 @section paletteuse
15166 Use a palette to downsample an input video stream.
15168 The filter takes two inputs: one video stream and a palette. The palette must
15169 be a 256 pixels image.
15171 It accepts the following options:
15175 Select dithering mode. Available algorithms are:
15178 Ordered 8x8 bayer dithering (deterministic)
15180 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15181 Note: this dithering is sometimes considered "wrong" and is included as a
15183 @item floyd_steinberg
15184 Floyd and Steingberg dithering (error diffusion)
15186 Frankie Sierra dithering v2 (error diffusion)
15188 Frankie Sierra dithering v2 "Lite" (error diffusion)
15191 Default is @var{sierra2_4a}.
15194 When @var{bayer} dithering is selected, this option defines the scale of the
15195 pattern (how much the crosshatch pattern is visible). A low value means more
15196 visible pattern for less banding, and higher value means less visible pattern
15197 at the cost of more banding.
15199 The option must be an integer value in the range [0,5]. Default is @var{2}.
15202 If set, define the zone to process
15206 Only the changing rectangle will be reprocessed. This is similar to GIF
15207 cropping/offsetting compression mechanism. This option can be useful for speed
15208 if only a part of the image is changing, and has use cases such as limiting the
15209 scope of the error diffusal @option{dither} to the rectangle that bounds the
15210 moving scene (it leads to more deterministic output if the scene doesn't change
15211 much, and as a result less moving noise and better GIF compression).
15214 Default is @var{none}.
15217 Take new palette for each output frame.
15219 @item alpha_threshold
15220 Sets the alpha threshold for transparency. Alpha values above this threshold
15221 will be treated as completely opaque, and values below this threshold will be
15222 treated as completely transparent.
15224 The option must be an integer value in the range [0,255]. Default is @var{128}.
15227 @subsection Examples
15231 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15232 using @command{ffmpeg}:
15234 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15238 @section perspective
15240 Correct perspective of video not recorded perpendicular to the screen.
15242 A description of the accepted parameters follows.
15253 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15254 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15255 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15256 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15257 then the corners of the source will be sent to the specified coordinates.
15259 The expressions can use the following variables:
15264 the width and height of video frame.
15268 Output frame count.
15271 @item interpolation
15272 Set interpolation for perspective correction.
15274 It accepts the following values:
15280 Default value is @samp{linear}.
15283 Set interpretation of coordinate options.
15285 It accepts the following values:
15289 Send point in the source specified by the given coordinates to
15290 the corners of the destination.
15292 @item 1, destination
15294 Send the corners of the source to the point in the destination specified
15295 by the given coordinates.
15297 Default value is @samp{source}.
15301 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15303 It accepts the following values:
15306 only evaluate expressions once during the filter initialization or
15307 when a command is processed
15310 evaluate expressions for each incoming frame
15313 Default value is @samp{init}.
15318 Delay interlaced video by one field time so that the field order changes.
15320 The intended use is to fix PAL movies that have been captured with the
15321 opposite field order to the film-to-video transfer.
15323 A description of the accepted parameters follows.
15329 It accepts the following values:
15332 Capture field order top-first, transfer bottom-first.
15333 Filter will delay the bottom field.
15336 Capture field order bottom-first, transfer top-first.
15337 Filter will delay the top field.
15340 Capture and transfer with the same field order. This mode only exists
15341 for the documentation of the other options to refer to, but if you
15342 actually select it, the filter will faithfully do nothing.
15345 Capture field order determined automatically by field flags, transfer
15347 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15348 basis using field flags. If no field information is available,
15349 then this works just like @samp{u}.
15352 Capture unknown or varying, transfer opposite.
15353 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15354 analyzing the images and selecting the alternative that produces best
15355 match between the fields.
15358 Capture top-first, transfer unknown or varying.
15359 Filter selects among @samp{t} and @samp{p} using image analysis.
15362 Capture bottom-first, transfer unknown or varying.
15363 Filter selects among @samp{b} and @samp{p} using image analysis.
15366 Capture determined by field flags, transfer unknown or varying.
15367 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15368 image analysis. If no field information is available, then this works just
15369 like @samp{U}. This is the default mode.
15372 Both capture and transfer unknown or varying.
15373 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15377 @section photosensitivity
15378 Reduce various flashes in video, so to help users with epilepsy.
15380 It accepts the following options:
15383 Set how many frames to use when filtering. Default is 30.
15386 Set detection threshold factor. Default is 1.
15390 Set how many pixels to skip when sampling frames. Default is 1.
15391 Allowed range is from 1 to 1024.
15394 Leave frames unchanged. Default is disabled.
15397 @section pixdesctest
15399 Pixel format descriptor test filter, mainly useful for internal
15400 testing. The output video should be equal to the input video.
15404 format=monow, pixdesctest
15407 can be used to test the monowhite pixel format descriptor definition.
15411 Display sample values of color channels. Mainly useful for checking color
15412 and levels. Minimum supported resolution is 640x480.
15414 The filters accept the following options:
15418 Set scope X position, relative offset on X axis.
15421 Set scope Y position, relative offset on Y axis.
15430 Set window opacity. This window also holds statistics about pixel area.
15433 Set window X position, relative offset on X axis.
15436 Set window Y position, relative offset on Y axis.
15441 Enable the specified chain of postprocessing subfilters using libpostproc. This
15442 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15443 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15444 Each subfilter and some options have a short and a long name that can be used
15445 interchangeably, i.e. dr/dering are the same.
15447 The filters accept the following options:
15451 Set postprocessing subfilters string.
15454 All subfilters share common options to determine their scope:
15458 Honor the quality commands for this subfilter.
15461 Do chrominance filtering, too (default).
15464 Do luminance filtering only (no chrominance).
15467 Do chrominance filtering only (no luminance).
15470 These options can be appended after the subfilter name, separated by a '|'.
15472 Available subfilters are:
15475 @item hb/hdeblock[|difference[|flatness]]
15476 Horizontal deblocking filter
15479 Difference factor where higher values mean more deblocking (default: @code{32}).
15481 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15484 @item vb/vdeblock[|difference[|flatness]]
15485 Vertical deblocking filter
15488 Difference factor where higher values mean more deblocking (default: @code{32}).
15490 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15493 @item ha/hadeblock[|difference[|flatness]]
15494 Accurate horizontal deblocking filter
15497 Difference factor where higher values mean more deblocking (default: @code{32}).
15499 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15502 @item va/vadeblock[|difference[|flatness]]
15503 Accurate vertical deblocking filter
15506 Difference factor where higher values mean more deblocking (default: @code{32}).
15508 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15512 The horizontal and vertical deblocking filters share the difference and
15513 flatness values so you cannot set different horizontal and vertical
15517 @item h1/x1hdeblock
15518 Experimental horizontal deblocking filter
15520 @item v1/x1vdeblock
15521 Experimental vertical deblocking filter
15526 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15529 larger -> stronger filtering
15531 larger -> stronger filtering
15533 larger -> stronger filtering
15536 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15539 Stretch luminance to @code{0-255}.
15542 @item lb/linblenddeint
15543 Linear blend deinterlacing filter that deinterlaces the given block by
15544 filtering all lines with a @code{(1 2 1)} filter.
15546 @item li/linipoldeint
15547 Linear interpolating deinterlacing filter that deinterlaces the given block by
15548 linearly interpolating every second line.
15550 @item ci/cubicipoldeint
15551 Cubic interpolating deinterlacing filter deinterlaces the given block by
15552 cubically interpolating every second line.
15554 @item md/mediandeint
15555 Median deinterlacing filter that deinterlaces the given block by applying a
15556 median filter to every second line.
15558 @item fd/ffmpegdeint
15559 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15560 second line with a @code{(-1 4 2 4 -1)} filter.
15563 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15564 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15566 @item fq/forceQuant[|quantizer]
15567 Overrides the quantizer table from the input with the constant quantizer you
15575 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15578 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15581 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15584 @subsection Examples
15588 Apply horizontal and vertical deblocking, deringing and automatic
15589 brightness/contrast:
15595 Apply default filters without brightness/contrast correction:
15601 Apply default filters and temporal denoiser:
15603 pp=default/tmpnoise|1|2|3
15607 Apply deblocking on luminance only, and switch vertical deblocking on or off
15608 automatically depending on available CPU time:
15615 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15616 similar to spp = 6 with 7 point DCT, where only the center sample is
15619 The filter accepts the following options:
15623 Force a constant quantization parameter. It accepts an integer in range
15624 0 to 63. If not set, the filter will use the QP from the video stream
15628 Set thresholding mode. Available modes are:
15632 Set hard thresholding.
15634 Set soft thresholding (better de-ringing effect, but likely blurrier).
15636 Set medium thresholding (good results, default).
15640 @section premultiply
15641 Apply alpha premultiply effect to input video stream using first plane
15642 of second stream as alpha.
15644 Both streams must have same dimensions and same pixel format.
15646 The filter accepts the following option:
15650 Set which planes will be processed, unprocessed planes will be copied.
15651 By default value 0xf, all planes will be processed.
15654 Do not require 2nd input for processing, instead use alpha plane from input stream.
15658 Apply prewitt operator to input video stream.
15660 The filter accepts the following option:
15664 Set which planes will be processed, unprocessed planes will be copied.
15665 By default value 0xf, all planes will be processed.
15668 Set value which will be multiplied with filtered result.
15671 Set value which will be added to filtered result.
15674 @section pseudocolor
15676 Alter frame colors in video with pseudocolors.
15678 This filter accepts the following options:
15682 set pixel first component expression
15685 set pixel second component expression
15688 set pixel third component expression
15691 set pixel fourth component expression, corresponds to the alpha component
15694 set component to use as base for altering colors
15697 Each of them specifies the expression to use for computing the lookup table for
15698 the corresponding pixel component values.
15700 The expressions can contain the following constants and functions:
15705 The input width and height.
15708 The input value for the pixel component.
15710 @item ymin, umin, vmin, amin
15711 The minimum allowed component value.
15713 @item ymax, umax, vmax, amax
15714 The maximum allowed component value.
15717 All expressions default to "val".
15719 @subsection Examples
15723 Change too high luma values to gradient:
15725 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'"
15731 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
15732 Ratio) between two input videos.
15734 This filter takes in input two input videos, the first input is
15735 considered the "main" source and is passed unchanged to the
15736 output. The second input is used as a "reference" video for computing
15739 Both video inputs must have the same resolution and pixel format for
15740 this filter to work correctly. Also it assumes that both inputs
15741 have the same number of frames, which are compared one by one.
15743 The obtained average PSNR is printed through the logging system.
15745 The filter stores the accumulated MSE (mean squared error) of each
15746 frame, and at the end of the processing it is averaged across all frames
15747 equally, and the following formula is applied to obtain the PSNR:
15750 PSNR = 10*log10(MAX^2/MSE)
15753 Where MAX is the average of the maximum values of each component of the
15756 The description of the accepted parameters follows.
15759 @item stats_file, f
15760 If specified the filter will use the named file to save the PSNR of
15761 each individual frame. When filename equals "-" the data is sent to
15764 @item stats_version
15765 Specifies which version of the stats file format to use. Details of
15766 each format are written below.
15767 Default value is 1.
15769 @item stats_add_max
15770 Determines whether the max value is output to the stats log.
15771 Default value is 0.
15772 Requires stats_version >= 2. If this is set and stats_version < 2,
15773 the filter will return an error.
15776 This filter also supports the @ref{framesync} options.
15778 The file printed if @var{stats_file} is selected, contains a sequence of
15779 key/value pairs of the form @var{key}:@var{value} for each compared
15782 If a @var{stats_version} greater than 1 is specified, a header line precedes
15783 the list of per-frame-pair stats, with key value pairs following the frame
15784 format with the following parameters:
15787 @item psnr_log_version
15788 The version of the log file format. Will match @var{stats_version}.
15791 A comma separated list of the per-frame-pair parameters included in
15795 A description of each shown per-frame-pair parameter follows:
15799 sequential number of the input frame, starting from 1
15802 Mean Square Error pixel-by-pixel average difference of the compared
15803 frames, averaged over all the image components.
15805 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
15806 Mean Square Error pixel-by-pixel average difference of the compared
15807 frames for the component specified by the suffix.
15809 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
15810 Peak Signal to Noise ratio of the compared frames for the component
15811 specified by the suffix.
15813 @item max_avg, max_y, max_u, max_v
15814 Maximum allowed value for each channel, and average over all
15818 @subsection Examples
15823 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
15824 [main][ref] psnr="stats_file=stats.log" [out]
15827 On this example the input file being processed is compared with the
15828 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
15829 is stored in @file{stats.log}.
15832 Another example with different containers:
15834 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]psnr" -f null -
15841 Pulldown reversal (inverse telecine) filter, capable of handling mixed
15842 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
15845 The pullup filter is designed to take advantage of future context in making
15846 its decisions. This filter is stateless in the sense that it does not lock
15847 onto a pattern to follow, but it instead looks forward to the following
15848 fields in order to identify matches and rebuild progressive frames.
15850 To produce content with an even framerate, insert the fps filter after
15851 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
15852 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
15854 The filter accepts the following options:
15861 These options set the amount of "junk" to ignore at the left, right, top, and
15862 bottom of the image, respectively. Left and right are in units of 8 pixels,
15863 while top and bottom are in units of 2 lines.
15864 The default is 8 pixels on each side.
15867 Set the strict breaks. Setting this option to 1 will reduce the chances of
15868 filter generating an occasional mismatched frame, but it may also cause an
15869 excessive number of frames to be dropped during high motion sequences.
15870 Conversely, setting it to -1 will make filter match fields more easily.
15871 This may help processing of video where there is slight blurring between
15872 the fields, but may also cause there to be interlaced frames in the output.
15873 Default value is @code{0}.
15876 Set the metric plane to use. It accepts the following values:
15882 Use chroma blue plane.
15885 Use chroma red plane.
15888 This option may be set to use chroma plane instead of the default luma plane
15889 for doing filter's computations. This may improve accuracy on very clean
15890 source material, but more likely will decrease accuracy, especially if there
15891 is chroma noise (rainbow effect) or any grayscale video.
15892 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
15893 load and make pullup usable in realtime on slow machines.
15896 For best results (without duplicated frames in the output file) it is
15897 necessary to change the output frame rate. For example, to inverse
15898 telecine NTSC input:
15900 ffmpeg -i input -vf pullup -r 24000/1001 ...
15905 Change video quantization parameters (QP).
15907 The filter accepts the following option:
15911 Set expression for quantization parameter.
15914 The expression is evaluated through the eval API and can contain, among others,
15915 the following constants:
15919 1 if index is not 129, 0 otherwise.
15922 Sequential index starting from -129 to 128.
15925 @subsection Examples
15929 Some equation like:
15937 Flush video frames from internal cache of frames into a random order.
15938 No frame is discarded.
15939 Inspired by @ref{frei0r} nervous filter.
15943 Set size in number of frames of internal cache, in range from @code{2} to
15944 @code{512}. Default is @code{30}.
15947 Set seed for random number generator, must be an integer included between
15948 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
15949 less than @code{0}, the filter will try to use a good random seed on a
15953 @section readeia608
15955 Read closed captioning (EIA-608) information from the top lines of a video frame.
15957 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
15958 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
15959 with EIA-608 data (starting from 0). A description of each metadata value follows:
15962 @item lavfi.readeia608.X.cc
15963 The two bytes stored as EIA-608 data (printed in hexadecimal).
15965 @item lavfi.readeia608.X.line
15966 The number of the line on which the EIA-608 data was identified and read.
15969 This filter accepts the following options:
15973 Set the line to start scanning for EIA-608 data. Default is @code{0}.
15976 Set the line to end scanning for EIA-608 data. Default is @code{29}.
15979 Set the ratio of width reserved for sync code detection.
15980 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
15983 Enable checking the parity bit. In the event of a parity error, the filter will output
15984 @code{0x00} for that character. Default is false.
15987 Lowpass lines prior to further processing. Default is enabled.
15990 @subsection Commands
15992 This filter supports the all above options as @ref{commands}.
15994 @subsection Examples
15998 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16000 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
16006 Read vertical interval timecode (VITC) information from the top lines of a
16009 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16010 timecode value, if a valid timecode has been detected. Further metadata key
16011 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16012 timecode data has been found or not.
16014 This filter accepts the following options:
16018 Set the maximum number of lines to scan for VITC data. If the value is set to
16019 @code{-1} the full video frame is scanned. Default is @code{45}.
16022 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16023 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16026 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16027 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16030 @subsection Examples
16034 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16035 draw @code{--:--:--:--} as a placeholder:
16037 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16043 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16045 Destination pixel at position (X, Y) will be picked from source (x, y) position
16046 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16047 value for pixel will be used for destination pixel.
16049 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16050 will have Xmap/Ymap video stream dimensions.
16051 Xmap and Ymap input video streams are 16bit depth, single channel.
16055 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16056 Default is @code{color}.
16059 Specify the color of the unmapped pixels. For the syntax of this option,
16060 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16061 manual,ffmpeg-utils}. Default color is @code{black}.
16064 @section removegrain
16066 The removegrain filter is a spatial denoiser for progressive video.
16070 Set mode for the first plane.
16073 Set mode for the second plane.
16076 Set mode for the third plane.
16079 Set mode for the fourth plane.
16082 Range of mode is from 0 to 24. Description of each mode follows:
16086 Leave input plane unchanged. Default.
16089 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16092 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16095 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16098 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16099 This is equivalent to a median filter.
16102 Line-sensitive clipping giving the minimal change.
16105 Line-sensitive clipping, intermediate.
16108 Line-sensitive clipping, intermediate.
16111 Line-sensitive clipping, intermediate.
16114 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16117 Replaces the target pixel with the closest neighbour.
16120 [1 2 1] horizontal and vertical kernel blur.
16126 Bob mode, interpolates top field from the line where the neighbours
16127 pixels are the closest.
16130 Bob mode, interpolates bottom field from the line where the neighbours
16131 pixels are the closest.
16134 Bob mode, interpolates top field. Same as 13 but with a more complicated
16135 interpolation formula.
16138 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16139 interpolation formula.
16142 Clips the pixel with the minimum and maximum of respectively the maximum and
16143 minimum of each pair of opposite neighbour pixels.
16146 Line-sensitive clipping using opposite neighbours whose greatest distance from
16147 the current pixel is minimal.
16150 Replaces the pixel with the average of its 8 neighbours.
16153 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16156 Clips pixels using the averages of opposite neighbour.
16159 Same as mode 21 but simpler and faster.
16162 Small edge and halo removal, but reputed useless.
16168 @section removelogo
16170 Suppress a TV station logo, using an image file to determine which
16171 pixels comprise the logo. It works by filling in the pixels that
16172 comprise the logo with neighboring pixels.
16174 The filter accepts the following options:
16178 Set the filter bitmap file, which can be any image format supported by
16179 libavformat. The width and height of the image file must match those of the
16180 video stream being processed.
16183 Pixels in the provided bitmap image with a value of zero are not
16184 considered part of the logo, non-zero pixels are considered part of
16185 the logo. If you use white (255) for the logo and black (0) for the
16186 rest, you will be safe. For making the filter bitmap, it is
16187 recommended to take a screen capture of a black frame with the logo
16188 visible, and then using a threshold filter followed by the erode
16189 filter once or twice.
16191 If needed, little splotches can be fixed manually. Remember that if
16192 logo pixels are not covered, the filter quality will be much
16193 reduced. Marking too many pixels as part of the logo does not hurt as
16194 much, but it will increase the amount of blurring needed to cover over
16195 the image and will destroy more information than necessary, and extra
16196 pixels will slow things down on a large logo.
16198 @section repeatfields
16200 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16201 fields based on its value.
16205 Reverse a video clip.
16207 Warning: This filter requires memory to buffer the entire clip, so trimming
16210 @subsection Examples
16214 Take the first 5 seconds of a clip, and reverse it.
16221 Shift R/G/B/A pixels horizontally and/or vertically.
16223 The filter accepts the following options:
16226 Set amount to shift red horizontally.
16228 Set amount to shift red vertically.
16230 Set amount to shift green horizontally.
16232 Set amount to shift green vertically.
16234 Set amount to shift blue horizontally.
16236 Set amount to shift blue vertically.
16238 Set amount to shift alpha horizontally.
16240 Set amount to shift alpha vertically.
16242 Set edge mode, can be @var{smear}, default, or @var{warp}.
16245 @subsection Commands
16247 This filter supports the all above options as @ref{commands}.
16250 Apply roberts cross operator to input video stream.
16252 The filter accepts the following option:
16256 Set which planes will be processed, unprocessed planes will be copied.
16257 By default value 0xf, all planes will be processed.
16260 Set value which will be multiplied with filtered result.
16263 Set value which will be added to filtered result.
16268 Rotate video by an arbitrary angle expressed in radians.
16270 The filter accepts the following options:
16272 A description of the optional parameters follows.
16275 Set an expression for the angle by which to rotate the input video
16276 clockwise, expressed as a number of radians. A negative value will
16277 result in a counter-clockwise rotation. By default it is set to "0".
16279 This expression is evaluated for each frame.
16282 Set the output width expression, default value is "iw".
16283 This expression is evaluated just once during configuration.
16286 Set the output height expression, default value is "ih".
16287 This expression is evaluated just once during configuration.
16290 Enable bilinear interpolation if set to 1, a value of 0 disables
16291 it. Default value is 1.
16294 Set the color used to fill the output area not covered by the rotated
16295 image. For the general syntax of this option, check the
16296 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16297 If the special value "none" is selected then no
16298 background is printed (useful for example if the background is never shown).
16300 Default value is "black".
16303 The expressions for the angle and the output size can contain the
16304 following constants and functions:
16308 sequential number of the input frame, starting from 0. It is always NAN
16309 before the first frame is filtered.
16312 time in seconds of the input frame, it is set to 0 when the filter is
16313 configured. It is always NAN before the first frame is filtered.
16317 horizontal and vertical chroma subsample values. For example for the
16318 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16322 the input video width and height
16326 the output width and height, that is the size of the padded area as
16327 specified by the @var{width} and @var{height} expressions
16331 the minimal width/height required for completely containing the input
16332 video rotated by @var{a} radians.
16334 These are only available when computing the @option{out_w} and
16335 @option{out_h} expressions.
16338 @subsection Examples
16342 Rotate the input by PI/6 radians clockwise:
16348 Rotate the input by PI/6 radians counter-clockwise:
16354 Rotate the input by 45 degrees clockwise:
16360 Apply a constant rotation with period T, starting from an angle of PI/3:
16362 rotate=PI/3+2*PI*t/T
16366 Make the input video rotation oscillating with a period of T
16367 seconds and an amplitude of A radians:
16369 rotate=A*sin(2*PI/T*t)
16373 Rotate the video, output size is chosen so that the whole rotating
16374 input video is always completely contained in the output:
16376 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16380 Rotate the video, reduce the output size so that no background is ever
16383 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16387 @subsection Commands
16389 The filter supports the following commands:
16393 Set the angle expression.
16394 The command accepts the same syntax of the corresponding option.
16396 If the specified expression is not valid, it is kept at its current
16402 Apply Shape Adaptive Blur.
16404 The filter accepts the following options:
16407 @item luma_radius, lr
16408 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16409 value is 1.0. A greater value will result in a more blurred image, and
16410 in slower processing.
16412 @item luma_pre_filter_radius, lpfr
16413 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16416 @item luma_strength, ls
16417 Set luma maximum difference between pixels to still be considered, must
16418 be a value in the 0.1-100.0 range, default value is 1.0.
16420 @item chroma_radius, cr
16421 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16422 greater value will result in a more blurred image, and in slower
16425 @item chroma_pre_filter_radius, cpfr
16426 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16428 @item chroma_strength, cs
16429 Set chroma maximum difference between pixels to still be considered,
16430 must be a value in the -0.9-100.0 range.
16433 Each chroma option value, if not explicitly specified, is set to the
16434 corresponding luma option value.
16439 Scale (resize) the input video, using the libswscale library.
16441 The scale filter forces the output display aspect ratio to be the same
16442 of the input, by changing the output sample aspect ratio.
16444 If the input image format is different from the format requested by
16445 the next filter, the scale filter will convert the input to the
16448 @subsection Options
16449 The filter accepts the following options, or any of the options
16450 supported by the libswscale scaler.
16452 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16453 the complete list of scaler options.
16458 Set the output video dimension expression. Default value is the input
16461 If the @var{width} or @var{w} value is 0, the input width is used for
16462 the output. If the @var{height} or @var{h} value is 0, the input height
16463 is used for the output.
16465 If one and only one of the values is -n with n >= 1, the scale filter
16466 will use a value that maintains the aspect ratio of the input image,
16467 calculated from the other specified dimension. After that it will,
16468 however, make sure that the calculated dimension is divisible by n and
16469 adjust the value if necessary.
16471 If both values are -n with n >= 1, the behavior will be identical to
16472 both values being set to 0 as previously detailed.
16474 See below for the list of accepted constants for use in the dimension
16478 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16482 Only evaluate expressions once during the filter initialization or when a command is processed.
16485 Evaluate expressions for each incoming frame.
16489 Default value is @samp{init}.
16493 Set the interlacing mode. It accepts the following values:
16497 Force interlaced aware scaling.
16500 Do not apply interlaced scaling.
16503 Select interlaced aware scaling depending on whether the source frames
16504 are flagged as interlaced or not.
16507 Default value is @samp{0}.
16510 Set libswscale scaling flags. See
16511 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16512 complete list of values. If not explicitly specified the filter applies
16516 @item param0, param1
16517 Set libswscale input parameters for scaling algorithms that need them. See
16518 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16519 complete documentation. If not explicitly specified the filter applies
16525 Set the video size. For the syntax of this option, check the
16526 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16528 @item in_color_matrix
16529 @item out_color_matrix
16530 Set in/output YCbCr color space type.
16532 This allows the autodetected value to be overridden as well as allows forcing
16533 a specific value used for the output and encoder.
16535 If not specified, the color space type depends on the pixel format.
16541 Choose automatically.
16544 Format conforming to International Telecommunication Union (ITU)
16545 Recommendation BT.709.
16548 Set color space conforming to the United States Federal Communications
16549 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16554 Set color space conforming to:
16558 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16561 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16564 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16569 Set color space conforming to SMPTE ST 240:1999.
16572 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16577 Set in/output YCbCr sample range.
16579 This allows the autodetected value to be overridden as well as allows forcing
16580 a specific value used for the output and encoder. If not specified, the
16581 range depends on the pixel format. Possible values:
16585 Choose automatically.
16588 Set full range (0-255 in case of 8-bit luma).
16590 @item mpeg/limited/tv
16591 Set "MPEG" range (16-235 in case of 8-bit luma).
16594 @item force_original_aspect_ratio
16595 Enable decreasing or increasing output video width or height if necessary to
16596 keep the original aspect ratio. Possible values:
16600 Scale the video as specified and disable this feature.
16603 The output video dimensions will automatically be decreased if needed.
16606 The output video dimensions will automatically be increased if needed.
16610 One useful instance of this option is that when you know a specific device's
16611 maximum allowed resolution, you can use this to limit the output video to
16612 that, while retaining the aspect ratio. For example, device A allows
16613 1280x720 playback, and your video is 1920x800. Using this option (set it to
16614 decrease) and specifying 1280x720 to the command line makes the output
16617 Please note that this is a different thing than specifying -1 for @option{w}
16618 or @option{h}, you still need to specify the output resolution for this option
16621 @item force_divisible_by
16622 Ensures that both the output dimensions, width and height, are divisible by the
16623 given integer when used together with @option{force_original_aspect_ratio}. This
16624 works similar to using @code{-n} in the @option{w} and @option{h} options.
16626 This option respects the value set for @option{force_original_aspect_ratio},
16627 increasing or decreasing the resolution accordingly. The video's aspect ratio
16628 may be slightly modified.
16630 This option can be handy if you need to have a video fit within or exceed
16631 a defined resolution using @option{force_original_aspect_ratio} but also have
16632 encoder restrictions on width or height divisibility.
16636 The values of the @option{w} and @option{h} options are expressions
16637 containing the following constants:
16642 The input width and height
16646 These are the same as @var{in_w} and @var{in_h}.
16650 The output (scaled) width and height
16654 These are the same as @var{out_w} and @var{out_h}
16657 The same as @var{iw} / @var{ih}
16660 input sample aspect ratio
16663 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16667 horizontal and vertical input chroma subsample values. For example for the
16668 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16672 horizontal and vertical output chroma subsample values. For example for the
16673 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16676 The (sequential) number of the input frame, starting from 0.
16677 Only available with @code{eval=frame}.
16680 The presentation timestamp of the input frame, expressed as a number of
16681 seconds. Only available with @code{eval=frame}.
16684 The position (byte offset) of the frame in the input stream, or NaN if
16685 this information is unavailable and/or meaningless (for example in case of synthetic video).
16686 Only available with @code{eval=frame}.
16689 @subsection Examples
16693 Scale the input video to a size of 200x100
16698 This is equivalent to:
16709 Specify a size abbreviation for the output size:
16714 which can also be written as:
16720 Scale the input to 2x:
16722 scale=w=2*iw:h=2*ih
16726 The above is the same as:
16728 scale=2*in_w:2*in_h
16732 Scale the input to 2x with forced interlaced scaling:
16734 scale=2*iw:2*ih:interl=1
16738 Scale the input to half size:
16740 scale=w=iw/2:h=ih/2
16744 Increase the width, and set the height to the same size:
16750 Seek Greek harmony:
16757 Increase the height, and set the width to 3/2 of the height:
16759 scale=w=3/2*oh:h=3/5*ih
16763 Increase the size, making the size a multiple of the chroma
16766 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
16770 Increase the width to a maximum of 500 pixels,
16771 keeping the same aspect ratio as the input:
16773 scale=w='min(500\, iw*3/2):h=-1'
16777 Make pixels square by combining scale and setsar:
16779 scale='trunc(ih*dar):ih',setsar=1/1
16783 Make pixels square by combining scale and setsar,
16784 making sure the resulting resolution is even (required by some codecs):
16786 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
16790 @subsection Commands
16792 This filter supports the following commands:
16796 Set the output video dimension expression.
16797 The command accepts the same syntax of the corresponding option.
16799 If the specified expression is not valid, it is kept at its current
16805 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
16806 format conversion on CUDA video frames. Setting the output width and height
16807 works in the same way as for the @var{scale} filter.
16809 The following additional options are accepted:
16812 The pixel format of the output CUDA frames. If set to the string "same" (the
16813 default), the input format will be kept. Note that automatic format negotiation
16814 and conversion is not yet supported for hardware frames
16817 The interpolation algorithm used for resizing. One of the following:
16824 @item cubic2p_bspline
16825 2-parameter cubic (B=1, C=0)
16827 @item cubic2p_catmullrom
16828 2-parameter cubic (B=0, C=1/2)
16830 @item cubic2p_b05c03
16831 2-parameter cubic (B=1/2, C=3/10)
16839 @item force_original_aspect_ratio
16840 Enable decreasing or increasing output video width or height if necessary to
16841 keep the original aspect ratio. Possible values:
16845 Scale the video as specified and disable this feature.
16848 The output video dimensions will automatically be decreased if needed.
16851 The output video dimensions will automatically be increased if needed.
16855 One useful instance of this option is that when you know a specific device's
16856 maximum allowed resolution, you can use this to limit the output video to
16857 that, while retaining the aspect ratio. For example, device A allows
16858 1280x720 playback, and your video is 1920x800. Using this option (set it to
16859 decrease) and specifying 1280x720 to the command line makes the output
16862 Please note that this is a different thing than specifying -1 for @option{w}
16863 or @option{h}, you still need to specify the output resolution for this option
16866 @item force_divisible_by
16867 Ensures that both the output dimensions, width and height, are divisible by the
16868 given integer when used together with @option{force_original_aspect_ratio}. This
16869 works similar to using @code{-n} in the @option{w} and @option{h} options.
16871 This option respects the value set for @option{force_original_aspect_ratio},
16872 increasing or decreasing the resolution accordingly. The video's aspect ratio
16873 may be slightly modified.
16875 This option can be handy if you need to have a video fit within or exceed
16876 a defined resolution using @option{force_original_aspect_ratio} but also have
16877 encoder restrictions on width or height divisibility.
16883 Scale (resize) the input video, based on a reference video.
16885 See the scale filter for available options, scale2ref supports the same but
16886 uses the reference video instead of the main input as basis. scale2ref also
16887 supports the following additional constants for the @option{w} and
16888 @option{h} options:
16893 The main input video's width and height
16896 The same as @var{main_w} / @var{main_h}
16899 The main input video's sample aspect ratio
16901 @item main_dar, mdar
16902 The main input video's display aspect ratio. Calculated from
16903 @code{(main_w / main_h) * main_sar}.
16907 The main input video's horizontal and vertical chroma subsample values.
16908 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
16912 The (sequential) number of the main input frame, starting from 0.
16913 Only available with @code{eval=frame}.
16916 The presentation timestamp of the main input frame, expressed as a number of
16917 seconds. Only available with @code{eval=frame}.
16920 The position (byte offset) of the frame in the main input stream, or NaN if
16921 this information is unavailable and/or meaningless (for example in case of synthetic video).
16922 Only available with @code{eval=frame}.
16925 @subsection Examples
16929 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
16931 'scale2ref[b][a];[a][b]overlay'
16935 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
16937 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
16941 @subsection Commands
16943 This filter supports the following commands:
16947 Set the output video dimension expression.
16948 The command accepts the same syntax of the corresponding option.
16950 If the specified expression is not valid, it is kept at its current
16955 Scroll input video horizontally and/or vertically by constant speed.
16957 The filter accepts the following options:
16959 @item horizontal, h
16960 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
16961 Negative values changes scrolling direction.
16964 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
16965 Negative values changes scrolling direction.
16968 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
16971 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
16974 @subsection Commands
16976 This filter supports the following @ref{commands}:
16978 @item horizontal, h
16979 Set the horizontal scrolling speed.
16981 Set the vertical scrolling speed.
16987 Detect video scene change.
16989 This filter sets frame metadata with mafd between frame, the scene score, and
16990 forward the frame to the next filter, so they can use these metadata to detect
16991 scene change or others.
16993 In addition, this filter logs a message and sets frame metadata when it detects
16994 a scene change by @option{threshold}.
16996 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
16998 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
16999 to detect scene change.
17001 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17002 detect scene change with @option{threshold}.
17004 The filter accepts the following options:
17008 Set the scene change detection threshold as a percentage of maximum change. Good
17009 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17012 Default value is @code{10.}.
17015 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17016 You can enable it if you want to get snapshot of scene change frames only.
17019 @anchor{selectivecolor}
17020 @section selectivecolor
17022 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17023 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17024 by the "purity" of the color (that is, how saturated it already is).
17026 This filter is similar to the Adobe Photoshop Selective Color tool.
17028 The filter accepts the following options:
17031 @item correction_method
17032 Select color correction method.
17034 Available values are:
17037 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17040 Specified adjustments are relative to the original component value.
17042 Default is @code{absolute}.
17044 Adjustments for red pixels (pixels where the red component is the maximum)
17046 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17048 Adjustments for green pixels (pixels where the green component is the maximum)
17050 Adjustments for cyan pixels (pixels where the red component is the minimum)
17052 Adjustments for blue pixels (pixels where the blue component is the maximum)
17054 Adjustments for magenta pixels (pixels where the green component is the minimum)
17056 Adjustments for white pixels (pixels where all components are greater than 128)
17058 Adjustments for all pixels except pure black and pure white
17060 Adjustments for black pixels (pixels where all components are lesser than 128)
17062 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17065 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17066 4 space separated floating point adjustment values in the [-1,1] range,
17067 respectively to adjust the amount of cyan, magenta, yellow and black for the
17068 pixels of its range.
17070 @subsection Examples
17074 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17075 increase magenta by 27% in blue areas:
17077 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17081 Use a Photoshop selective color preset:
17083 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17087 @anchor{separatefields}
17088 @section separatefields
17090 The @code{separatefields} takes a frame-based video input and splits
17091 each frame into its components fields, producing a new half height clip
17092 with twice the frame rate and twice the frame count.
17094 This filter use field-dominance information in frame to decide which
17095 of each pair of fields to place first in the output.
17096 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17098 @section setdar, setsar
17100 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17103 This is done by changing the specified Sample (aka Pixel) Aspect
17104 Ratio, according to the following equation:
17106 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17109 Keep in mind that the @code{setdar} filter does not modify the pixel
17110 dimensions of the video frame. Also, the display aspect ratio set by
17111 this filter may be changed by later filters in the filterchain,
17112 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17115 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17116 the filter output video.
17118 Note that as a consequence of the application of this filter, the
17119 output display aspect ratio will change according to the equation
17122 Keep in mind that the sample aspect ratio set by the @code{setsar}
17123 filter may be changed by later filters in the filterchain, e.g. if
17124 another "setsar" or a "setdar" filter is applied.
17126 It accepts the following parameters:
17129 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17130 Set the aspect ratio used by the filter.
17132 The parameter can be a floating point number string, an expression, or
17133 a string of the form @var{num}:@var{den}, where @var{num} and
17134 @var{den} are the numerator and denominator of the aspect ratio. If
17135 the parameter is not specified, it is assumed the value "0".
17136 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17140 Set the maximum integer value to use for expressing numerator and
17141 denominator when reducing the expressed aspect ratio to a rational.
17142 Default value is @code{100}.
17146 The parameter @var{sar} is an expression containing
17147 the following constants:
17151 These are approximated values for the mathematical constants e
17152 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17155 The input width and height.
17158 These are the same as @var{w} / @var{h}.
17161 The input sample aspect ratio.
17164 The input display aspect ratio. It is the same as
17165 (@var{w} / @var{h}) * @var{sar}.
17168 Horizontal and vertical chroma subsample values. For example, for the
17169 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17172 @subsection Examples
17177 To change the display aspect ratio to 16:9, specify one of the following:
17184 To change the sample aspect ratio to 10:11, specify:
17190 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17191 1000 in the aspect ratio reduction, use the command:
17193 setdar=ratio=16/9:max=1000
17201 Force field for the output video frame.
17203 The @code{setfield} filter marks the interlace type field for the
17204 output frames. It does not change the input frame, but only sets the
17205 corresponding property, which affects how the frame is treated by
17206 following filters (e.g. @code{fieldorder} or @code{yadif}).
17208 The filter accepts the following options:
17213 Available values are:
17217 Keep the same field property.
17220 Mark the frame as bottom-field-first.
17223 Mark the frame as top-field-first.
17226 Mark the frame as progressive.
17233 Force frame parameter for the output video frame.
17235 The @code{setparams} filter marks interlace and color range for the
17236 output frames. It does not change the input frame, but only sets the
17237 corresponding property, which affects how the frame is treated by
17242 Available values are:
17246 Keep the same field property (default).
17249 Mark the frame as bottom-field-first.
17252 Mark the frame as top-field-first.
17255 Mark the frame as progressive.
17259 Available values are:
17263 Keep the same color range property (default).
17265 @item unspecified, unknown
17266 Mark the frame as unspecified color range.
17268 @item limited, tv, mpeg
17269 Mark the frame as limited range.
17271 @item full, pc, jpeg
17272 Mark the frame as full range.
17275 @item color_primaries
17276 Set the color primaries.
17277 Available values are:
17281 Keep the same color primaries property (default).
17298 Set the color transfer.
17299 Available values are:
17303 Keep the same color trc property (default).
17325 Set the colorspace.
17326 Available values are:
17330 Keep the same colorspace property (default).
17343 @item chroma-derived-nc
17344 @item chroma-derived-c
17351 Show a line containing various information for each input video frame.
17352 The input video is not modified.
17354 This filter supports the following options:
17358 Calculate checksums of each plane. By default enabled.
17361 The shown line contains a sequence of key/value pairs of the form
17362 @var{key}:@var{value}.
17364 The following values are shown in the output:
17368 The (sequential) number of the input frame, starting from 0.
17371 The Presentation TimeStamp of the input frame, expressed as a number of
17372 time base units. The time base unit depends on the filter input pad.
17375 The Presentation TimeStamp of the input frame, expressed as a number of
17379 The position of the frame in the input stream, or -1 if this information is
17380 unavailable and/or meaningless (for example in case of synthetic video).
17383 The pixel format name.
17386 The sample aspect ratio of the input frame, expressed in the form
17387 @var{num}/@var{den}.
17390 The size of the input frame. For the syntax of this option, check the
17391 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17394 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17395 for bottom field first).
17398 This is 1 if the frame is a key frame, 0 otherwise.
17401 The picture type of the input frame ("I" for an I-frame, "P" for a
17402 P-frame, "B" for a B-frame, or "?" for an unknown type).
17403 Also refer to the documentation of the @code{AVPictureType} enum and of
17404 the @code{av_get_picture_type_char} function defined in
17405 @file{libavutil/avutil.h}.
17408 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17410 @item plane_checksum
17411 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17412 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17415 The mean value of pixels in each plane of the input frame, expressed in the form
17416 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17419 The standard deviation of pixel values in each plane of the input frame, expressed
17420 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17424 @section showpalette
17426 Displays the 256 colors palette of each frame. This filter is only relevant for
17427 @var{pal8} pixel format frames.
17429 It accepts the following option:
17433 Set the size of the box used to represent one palette color entry. Default is
17434 @code{30} (for a @code{30x30} pixel box).
17437 @section shuffleframes
17439 Reorder and/or duplicate and/or drop video frames.
17441 It accepts the following parameters:
17445 Set the destination indexes of input frames.
17446 This is space or '|' separated list of indexes that maps input frames to output
17447 frames. Number of indexes also sets maximal value that each index may have.
17448 '-1' index have special meaning and that is to drop frame.
17451 The first frame has the index 0. The default is to keep the input unchanged.
17453 @subsection Examples
17457 Swap second and third frame of every three frames of the input:
17459 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17463 Swap 10th and 1st frame of every ten frames of the input:
17465 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17469 @section shuffleplanes
17471 Reorder and/or duplicate video planes.
17473 It accepts the following parameters:
17478 The index of the input plane to be used as the first output plane.
17481 The index of the input plane to be used as the second output plane.
17484 The index of the input plane to be used as the third output plane.
17487 The index of the input plane to be used as the fourth output plane.
17491 The first plane has the index 0. The default is to keep the input unchanged.
17493 @subsection Examples
17497 Swap the second and third planes of the input:
17499 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17503 @anchor{signalstats}
17504 @section signalstats
17505 Evaluate various visual metrics that assist in determining issues associated
17506 with the digitization of analog video media.
17508 By default the filter will log these metadata values:
17512 Display the minimal Y value contained within the input frame. Expressed in
17516 Display the Y value at the 10% percentile within the input frame. Expressed in
17520 Display the average Y value within the input frame. Expressed in range of
17524 Display the Y value at the 90% percentile within the input frame. Expressed in
17528 Display the maximum Y value contained within the input frame. Expressed in
17532 Display the minimal U value contained within the input frame. Expressed in
17536 Display the U value at the 10% percentile within the input frame. Expressed in
17540 Display the average U value within the input frame. Expressed in range of
17544 Display the U value at the 90% percentile within the input frame. Expressed in
17548 Display the maximum U value contained within the input frame. Expressed in
17552 Display the minimal V value contained within the input frame. Expressed in
17556 Display the V value at the 10% percentile within the input frame. Expressed in
17560 Display the average V value within the input frame. Expressed in range of
17564 Display the V value at the 90% percentile within the input frame. Expressed in
17568 Display the maximum V value contained within the input frame. Expressed in
17572 Display the minimal saturation value contained within the input frame.
17573 Expressed in range of [0-~181.02].
17576 Display the saturation value at the 10% percentile within the input frame.
17577 Expressed in range of [0-~181.02].
17580 Display the average saturation value within the input frame. Expressed in range
17584 Display the saturation value at the 90% percentile within the input frame.
17585 Expressed in range of [0-~181.02].
17588 Display the maximum saturation value contained within the input frame.
17589 Expressed in range of [0-~181.02].
17592 Display the median value for hue within the input frame. Expressed in range of
17596 Display the average value for hue within the input frame. Expressed in range of
17600 Display the average of sample value difference between all values of the Y
17601 plane in the current frame and corresponding values of the previous input frame.
17602 Expressed in range of [0-255].
17605 Display the average of sample value difference between all values of the U
17606 plane in the current frame and corresponding values of the previous input frame.
17607 Expressed in range of [0-255].
17610 Display the average of sample value difference between all values of the V
17611 plane in the current frame and corresponding values of the previous input frame.
17612 Expressed in range of [0-255].
17615 Display bit depth of Y plane in current frame.
17616 Expressed in range of [0-16].
17619 Display bit depth of U plane in current frame.
17620 Expressed in range of [0-16].
17623 Display bit depth of V plane in current frame.
17624 Expressed in range of [0-16].
17627 The filter accepts the following options:
17633 @option{stat} specify an additional form of image analysis.
17634 @option{out} output video with the specified type of pixel highlighted.
17636 Both options accept the following values:
17640 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17641 unlike the neighboring pixels of the same field. Examples of temporal outliers
17642 include the results of video dropouts, head clogs, or tape tracking issues.
17645 Identify @var{vertical line repetition}. Vertical line repetition includes
17646 similar rows of pixels within a frame. In born-digital video vertical line
17647 repetition is common, but this pattern is uncommon in video digitized from an
17648 analog source. When it occurs in video that results from the digitization of an
17649 analog source it can indicate concealment from a dropout compensator.
17652 Identify pixels that fall outside of legal broadcast range.
17656 Set the highlight color for the @option{out} option. The default color is
17660 @subsection Examples
17664 Output data of various video metrics:
17666 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17670 Output specific data about the minimum and maximum values of the Y plane per frame:
17672 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17676 Playback video while highlighting pixels that are outside of broadcast range in red.
17678 ffplay example.mov -vf signalstats="out=brng:color=red"
17682 Playback video with signalstats metadata drawn over the frame.
17684 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17687 The contents of signalstat_drawtext.txt used in the command are:
17690 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17691 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17692 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17693 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17701 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17702 input. In this case the matching between the inputs can be calculated additionally.
17703 The filter always passes through the first input. The signature of each stream can
17704 be written into a file.
17706 It accepts the following options:
17710 Enable or disable the matching process.
17712 Available values are:
17716 Disable the calculation of a matching (default).
17718 Calculate the matching for the whole video and output whether the whole video
17719 matches or only parts.
17721 Calculate only until a matching is found or the video ends. Should be faster in
17726 Set the number of inputs. The option value must be a non negative integer.
17727 Default value is 1.
17730 Set the path to which the output is written. If there is more than one input,
17731 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
17732 integer), that will be replaced with the input number. If no filename is
17733 specified, no output will be written. This is the default.
17736 Choose the output format.
17738 Available values are:
17742 Use the specified binary representation (default).
17744 Use the specified xml representation.
17748 Set threshold to detect one word as similar. The option value must be an integer
17749 greater than zero. The default value is 9000.
17752 Set threshold to detect all words as similar. The option value must be an integer
17753 greater than zero. The default value is 60000.
17756 Set threshold to detect frames as similar. The option value must be an integer
17757 greater than zero. The default value is 116.
17760 Set the minimum length of a sequence in frames to recognize it as matching
17761 sequence. The option value must be a non negative integer value.
17762 The default value is 0.
17765 Set the minimum relation, that matching frames to all frames must have.
17766 The option value must be a double value between 0 and 1. The default value is 0.5.
17769 @subsection Examples
17773 To calculate the signature of an input video and store it in signature.bin:
17775 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
17779 To detect whether two videos match and store the signatures in XML format in
17780 signature0.xml and signature1.xml:
17782 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 -
17790 Blur the input video without impacting the outlines.
17792 It accepts the following options:
17795 @item luma_radius, lr
17796 Set the luma radius. The option value must be a float number in
17797 the range [0.1,5.0] that specifies the variance of the gaussian filter
17798 used to blur the image (slower if larger). Default value is 1.0.
17800 @item luma_strength, ls
17801 Set the luma strength. The option value must be a float number
17802 in the range [-1.0,1.0] that configures the blurring. A value included
17803 in [0.0,1.0] will blur the image whereas a value included in
17804 [-1.0,0.0] will sharpen the image. Default value is 1.0.
17806 @item luma_threshold, lt
17807 Set the luma threshold used as a coefficient to determine
17808 whether a pixel should be blurred or not. The option value must be an
17809 integer in the range [-30,30]. A value of 0 will filter all the image,
17810 a value included in [0,30] will filter flat areas and a value included
17811 in [-30,0] will filter edges. Default value is 0.
17813 @item chroma_radius, cr
17814 Set the chroma radius. The option value must be a float number in
17815 the range [0.1,5.0] that specifies the variance of the gaussian filter
17816 used to blur the image (slower if larger). Default value is @option{luma_radius}.
17818 @item chroma_strength, cs
17819 Set the chroma strength. The option value must be a float number
17820 in the range [-1.0,1.0] that configures the blurring. A value included
17821 in [0.0,1.0] will blur the image whereas a value included in
17822 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
17824 @item chroma_threshold, ct
17825 Set the chroma threshold used as a coefficient to determine
17826 whether a pixel should be blurred or not. The option value must be an
17827 integer in the range [-30,30]. A value of 0 will filter all the image,
17828 a value included in [0,30] will filter flat areas and a value included
17829 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
17832 If a chroma option is not explicitly set, the corresponding luma value
17836 Apply sobel operator to input video stream.
17838 The filter accepts the following option:
17842 Set which planes will be processed, unprocessed planes will be copied.
17843 By default value 0xf, all planes will be processed.
17846 Set value which will be multiplied with filtered result.
17849 Set value which will be added to filtered result.
17855 Apply a simple postprocessing filter that compresses and decompresses the image
17856 at several (or - in the case of @option{quality} level @code{6} - all) shifts
17857 and average the results.
17859 The filter accepts the following options:
17863 Set quality. This option defines the number of levels for averaging. It accepts
17864 an integer in the range 0-6. If set to @code{0}, the filter will have no
17865 effect. A value of @code{6} means the higher quality. For each increment of
17866 that value the speed drops by a factor of approximately 2. Default value is
17870 Force a constant quantization parameter. If not set, the filter will use the QP
17871 from the video stream (if available).
17874 Set thresholding mode. Available modes are:
17878 Set hard thresholding (default).
17880 Set soft thresholding (better de-ringing effect, but likely blurrier).
17883 @item use_bframe_qp
17884 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
17885 option may cause flicker since the B-Frames have often larger QP. Default is
17886 @code{0} (not enabled).
17889 @subsection Commands
17891 This filter supports the following commands:
17893 @item quality, level
17894 Set quality level. The value @code{max} can be used to set the maximum level,
17895 currently @code{6}.
17901 Scale the input by applying one of the super-resolution methods based on
17902 convolutional neural networks. Supported models:
17906 Super-Resolution Convolutional Neural Network model (SRCNN).
17907 See @url{https://arxiv.org/abs/1501.00092}.
17910 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
17911 See @url{https://arxiv.org/abs/1609.05158}.
17914 Training scripts as well as scripts for model file (.pb) saving can be found at
17915 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
17916 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
17918 Native model files (.model) can be generated from TensorFlow model
17919 files (.pb) by using tools/python/convert.py
17921 The filter accepts the following options:
17925 Specify which DNN backend to use for model loading and execution. This option accepts
17926 the following values:
17930 Native implementation of DNN loading and execution.
17933 TensorFlow backend. To enable this backend you
17934 need to install the TensorFlow for C library (see
17935 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
17936 @code{--enable-libtensorflow}
17939 Default value is @samp{native}.
17942 Set path to model file specifying network architecture and its parameters.
17943 Note that different backends use different file formats. TensorFlow backend
17944 can load files for both formats, while native backend can load files for only
17948 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
17949 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
17950 input upscaled using bicubic upscaling with proper scale factor.
17953 This feature can also be finished with @ref{dnn_processing} filter.
17957 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
17959 This filter takes in input two input videos, the first input is
17960 considered the "main" source and is passed unchanged to the
17961 output. The second input is used as a "reference" video for computing
17964 Both video inputs must have the same resolution and pixel format for
17965 this filter to work correctly. Also it assumes that both inputs
17966 have the same number of frames, which are compared one by one.
17968 The filter stores the calculated SSIM of each frame.
17970 The description of the accepted parameters follows.
17973 @item stats_file, f
17974 If specified the filter will use the named file to save the SSIM of
17975 each individual frame. When filename equals "-" the data is sent to
17979 The file printed if @var{stats_file} is selected, contains a sequence of
17980 key/value pairs of the form @var{key}:@var{value} for each compared
17983 A description of each shown parameter follows:
17987 sequential number of the input frame, starting from 1
17989 @item Y, U, V, R, G, B
17990 SSIM of the compared frames for the component specified by the suffix.
17993 SSIM of the compared frames for the whole frame.
17996 Same as above but in dB representation.
17999 This filter also supports the @ref{framesync} options.
18001 @subsection Examples
18006 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18007 [main][ref] ssim="stats_file=stats.log" [out]
18010 On this example the input file being processed is compared with the
18011 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18012 is stored in @file{stats.log}.
18015 Another example with both psnr and ssim at same time:
18017 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18021 Another example with different containers:
18023 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]ssim" -f null -
18029 Convert between different stereoscopic image formats.
18031 The filters accept the following options:
18035 Set stereoscopic image format of input.
18037 Available values for input image formats are:
18040 side by side parallel (left eye left, right eye right)
18043 side by side crosseye (right eye left, left eye right)
18046 side by side parallel with half width resolution
18047 (left eye left, right eye right)
18050 side by side crosseye with half width resolution
18051 (right eye left, left eye right)
18055 above-below (left eye above, right eye below)
18059 above-below (right eye above, left eye below)
18063 above-below with half height resolution
18064 (left eye above, right eye below)
18068 above-below with half height resolution
18069 (right eye above, left eye below)
18072 alternating frames (left eye first, right eye second)
18075 alternating frames (right eye first, left eye second)
18078 interleaved rows (left eye has top row, right eye starts on next row)
18081 interleaved rows (right eye has top row, left eye starts on next row)
18084 interleaved columns, left eye first
18087 interleaved columns, right eye first
18089 Default value is @samp{sbsl}.
18093 Set stereoscopic image format of output.
18097 side by side parallel (left eye left, right eye right)
18100 side by side crosseye (right eye left, left eye right)
18103 side by side parallel with half width resolution
18104 (left eye left, right eye right)
18107 side by side crosseye with half width resolution
18108 (right eye left, left eye right)
18112 above-below (left eye above, right eye below)
18116 above-below (right eye above, left eye below)
18120 above-below with half height resolution
18121 (left eye above, right eye below)
18125 above-below with half height resolution
18126 (right eye above, left eye below)
18129 alternating frames (left eye first, right eye second)
18132 alternating frames (right eye first, left eye second)
18135 interleaved rows (left eye has top row, right eye starts on next row)
18138 interleaved rows (right eye has top row, left eye starts on next row)
18141 anaglyph red/blue gray
18142 (red filter on left eye, blue filter on right eye)
18145 anaglyph red/green gray
18146 (red filter on left eye, green filter on right eye)
18149 anaglyph red/cyan gray
18150 (red filter on left eye, cyan filter on right eye)
18153 anaglyph red/cyan half colored
18154 (red filter on left eye, cyan filter on right eye)
18157 anaglyph red/cyan color
18158 (red filter on left eye, cyan filter on right eye)
18161 anaglyph red/cyan color optimized with the least squares projection of dubois
18162 (red filter on left eye, cyan filter on right eye)
18165 anaglyph green/magenta gray
18166 (green filter on left eye, magenta filter on right eye)
18169 anaglyph green/magenta half colored
18170 (green filter on left eye, magenta filter on right eye)
18173 anaglyph green/magenta colored
18174 (green filter on left eye, magenta filter on right eye)
18177 anaglyph green/magenta color optimized with the least squares projection of dubois
18178 (green filter on left eye, magenta filter on right eye)
18181 anaglyph yellow/blue gray
18182 (yellow filter on left eye, blue filter on right eye)
18185 anaglyph yellow/blue half colored
18186 (yellow filter on left eye, blue filter on right eye)
18189 anaglyph yellow/blue colored
18190 (yellow filter on left eye, blue filter on right eye)
18193 anaglyph yellow/blue color optimized with the least squares projection of dubois
18194 (yellow filter on left eye, blue filter on right eye)
18197 mono output (left eye only)
18200 mono output (right eye only)
18203 checkerboard, left eye first
18206 checkerboard, right eye first
18209 interleaved columns, left eye first
18212 interleaved columns, right eye first
18218 Default value is @samp{arcd}.
18221 @subsection Examples
18225 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18231 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18237 @section streamselect, astreamselect
18238 Select video or audio streams.
18240 The filter accepts the following options:
18244 Set number of inputs. Default is 2.
18247 Set input indexes to remap to outputs.
18250 @subsection Commands
18252 The @code{streamselect} and @code{astreamselect} filter supports the following
18257 Set input indexes to remap to outputs.
18260 @subsection Examples
18264 Select first 5 seconds 1st stream and rest of time 2nd stream:
18266 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18270 Same as above, but for audio:
18272 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18279 Draw subtitles on top of input video using the libass library.
18281 To enable compilation of this filter you need to configure FFmpeg with
18282 @code{--enable-libass}. This filter also requires a build with libavcodec and
18283 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18284 Alpha) subtitles format.
18286 The filter accepts the following options:
18290 Set the filename of the subtitle file to read. It must be specified.
18292 @item original_size
18293 Specify the size of the original video, the video for which the ASS file
18294 was composed. For the syntax of this option, check the
18295 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18296 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18297 correctly scale the fonts if the aspect ratio has been changed.
18300 Set a directory path containing fonts that can be used by the filter.
18301 These fonts will be used in addition to whatever the font provider uses.
18304 Process alpha channel, by default alpha channel is untouched.
18307 Set subtitles input character encoding. @code{subtitles} filter only. Only
18308 useful if not UTF-8.
18310 @item stream_index, si
18311 Set subtitles stream index. @code{subtitles} filter only.
18314 Override default style or script info parameters of the subtitles. It accepts a
18315 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18318 If the first key is not specified, it is assumed that the first value
18319 specifies the @option{filename}.
18321 For example, to render the file @file{sub.srt} on top of the input
18322 video, use the command:
18327 which is equivalent to:
18329 subtitles=filename=sub.srt
18332 To render the default subtitles stream from file @file{video.mkv}, use:
18334 subtitles=video.mkv
18337 To render the second subtitles stream from that file, use:
18339 subtitles=video.mkv:si=1
18342 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18343 @code{DejaVu Serif}, use:
18345 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18348 @section super2xsai
18350 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18351 Interpolate) pixel art scaling algorithm.
18353 Useful for enlarging pixel art images without reducing sharpness.
18357 Swap two rectangular objects in video.
18359 This filter accepts the following options:
18369 Set 1st rect x coordinate.
18372 Set 1st rect y coordinate.
18375 Set 2nd rect x coordinate.
18378 Set 2nd rect y coordinate.
18380 All expressions are evaluated once for each frame.
18383 The all options are expressions containing the following constants:
18388 The input width and height.
18391 same as @var{w} / @var{h}
18394 input sample aspect ratio
18397 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18400 The number of the input frame, starting from 0.
18403 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18406 the position in the file of the input frame, NAN if unknown
18413 Blend successive video frames.
18419 Apply telecine process to the video.
18421 This filter accepts the following options:
18430 The default value is @code{top}.
18434 A string of numbers representing the pulldown pattern you wish to apply.
18435 The default value is @code{23}.
18439 Some typical patterns:
18444 24p: 2332 (preferred)
18451 24p: 222222222223 ("Euro pulldown")
18456 @section thistogram
18458 Compute and draw a color distribution histogram for the input video across time.
18460 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18461 at certain time, this filter shows also past histograms of number of frames defined
18462 by @code{width} option.
18464 The computed histogram is a representation of the color component
18465 distribution in an image.
18467 The filter accepts the following options:
18471 Set width of single color component output. Default value is @code{0}.
18472 Value of @code{0} means width will be picked from input video.
18473 This also set number of passed histograms to keep.
18474 Allowed range is [0, 8192].
18476 @item display_mode, d
18478 It accepts the following values:
18481 Per color component graphs are placed below each other.
18484 Per color component graphs are placed side by side.
18487 Presents information identical to that in the @code{parade}, except
18488 that the graphs representing color components are superimposed directly
18491 Default is @code{stack}.
18493 @item levels_mode, m
18494 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18495 Default is @code{linear}.
18497 @item components, c
18498 Set what color components to display.
18499 Default is @code{7}.
18502 Set background opacity. Default is @code{0.9}.
18505 Show envelope. Default is disabled.
18508 Set envelope color. Default is @code{gold}.
18513 Available values for slide is:
18516 Draw new frame when right border is reached.
18519 Replace old columns with new ones.
18522 Scroll from right to left.
18525 Scroll from left to right.
18528 Draw single picture.
18531 Default is @code{replace}.
18536 Apply threshold effect to video stream.
18538 This filter needs four video streams to perform thresholding.
18539 First stream is stream we are filtering.
18540 Second stream is holding threshold values, third stream is holding min values,
18541 and last, fourth stream is holding max values.
18543 The filter accepts the following option:
18547 Set which planes will be processed, unprocessed planes will be copied.
18548 By default value 0xf, all planes will be processed.
18551 For example if first stream pixel's component value is less then threshold value
18552 of pixel component from 2nd threshold stream, third stream value will picked,
18553 otherwise fourth stream pixel component value will be picked.
18555 Using color source filter one can perform various types of thresholding:
18557 @subsection Examples
18561 Binary threshold, using gray color as threshold:
18563 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18567 Inverted binary threshold, using gray color as threshold:
18569 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18573 Truncate binary threshold, using gray color as threshold:
18575 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18579 Threshold to zero, using gray color as threshold:
18581 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18585 Inverted threshold to zero, using gray color as threshold:
18587 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18592 Select the most representative frame in a given sequence of consecutive frames.
18594 The filter accepts the following options:
18598 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18599 will pick one of them, and then handle the next batch of @var{n} frames until
18600 the end. Default is @code{100}.
18603 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18604 value will result in a higher memory usage, so a high value is not recommended.
18606 @subsection Examples
18610 Extract one picture each 50 frames:
18616 Complete example of a thumbnail creation with @command{ffmpeg}:
18618 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18625 Tile several successive frames together.
18627 The @ref{untile} filter can do the reverse.
18629 The filter accepts the following options:
18634 Set the grid size (i.e. the number of lines and columns). For the syntax of
18635 this option, check the
18636 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18639 Set the maximum number of frames to render in the given area. It must be less
18640 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18641 the area will be used.
18644 Set the outer border margin in pixels.
18647 Set the inner border thickness (i.e. the number of pixels between frames). For
18648 more advanced padding options (such as having different values for the edges),
18649 refer to the pad video filter.
18652 Specify the color of the unused area. For the syntax of this option, check the
18653 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18654 The default value of @var{color} is "black".
18657 Set the number of frames to overlap when tiling several successive frames together.
18658 The value must be between @code{0} and @var{nb_frames - 1}.
18661 Set the number of frames to initially be empty before displaying first output frame.
18662 This controls how soon will one get first output frame.
18663 The value must be between @code{0} and @var{nb_frames - 1}.
18666 @subsection Examples
18670 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18672 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18674 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18675 duplicating each output frame to accommodate the originally detected frame
18679 Display @code{5} pictures in an area of @code{3x2} frames,
18680 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18681 mixed flat and named options:
18683 tile=3x2:nb_frames=5:padding=7:margin=2
18687 @section tinterlace
18689 Perform various types of temporal field interlacing.
18691 Frames are counted starting from 1, so the first input frame is
18694 The filter accepts the following options:
18699 Specify the mode of the interlacing. This option can also be specified
18700 as a value alone. See below for a list of values for this option.
18702 Available values are:
18706 Move odd frames into the upper field, even into the lower field,
18707 generating a double height frame at half frame rate.
18711 Frame 1 Frame 2 Frame 3 Frame 4
18713 11111 22222 33333 44444
18714 11111 22222 33333 44444
18715 11111 22222 33333 44444
18716 11111 22222 33333 44444
18730 Only output odd frames, even frames are dropped, generating a frame with
18731 unchanged height at half frame rate.
18736 Frame 1 Frame 2 Frame 3 Frame 4
18738 11111 22222 33333 44444
18739 11111 22222 33333 44444
18740 11111 22222 33333 44444
18741 11111 22222 33333 44444
18751 Only output even frames, odd frames are dropped, generating a frame with
18752 unchanged height at half frame rate.
18757 Frame 1 Frame 2 Frame 3 Frame 4
18759 11111 22222 33333 44444
18760 11111 22222 33333 44444
18761 11111 22222 33333 44444
18762 11111 22222 33333 44444
18772 Expand each frame to full height, but pad alternate lines with black,
18773 generating a frame with double height at the same input frame rate.
18778 Frame 1 Frame 2 Frame 3 Frame 4
18780 11111 22222 33333 44444
18781 11111 22222 33333 44444
18782 11111 22222 33333 44444
18783 11111 22222 33333 44444
18786 11111 ..... 33333 .....
18787 ..... 22222 ..... 44444
18788 11111 ..... 33333 .....
18789 ..... 22222 ..... 44444
18790 11111 ..... 33333 .....
18791 ..... 22222 ..... 44444
18792 11111 ..... 33333 .....
18793 ..... 22222 ..... 44444
18797 @item interleave_top, 4
18798 Interleave the upper field from odd frames with the lower field from
18799 even frames, generating a frame with unchanged height at half frame rate.
18804 Frame 1 Frame 2 Frame 3 Frame 4
18806 11111<- 22222 33333<- 44444
18807 11111 22222<- 33333 44444<-
18808 11111<- 22222 33333<- 44444
18809 11111 22222<- 33333 44444<-
18819 @item interleave_bottom, 5
18820 Interleave the lower field from odd frames with the upper field from
18821 even frames, generating a frame with unchanged height at half frame rate.
18826 Frame 1 Frame 2 Frame 3 Frame 4
18828 11111 22222<- 33333 44444<-
18829 11111<- 22222 33333<- 44444
18830 11111 22222<- 33333 44444<-
18831 11111<- 22222 33333<- 44444
18841 @item interlacex2, 6
18842 Double frame rate with unchanged height. Frames are inserted each
18843 containing the second temporal field from the previous input frame and
18844 the first temporal field from the next input frame. This mode relies on
18845 the top_field_first flag. Useful for interlaced video displays with no
18846 field synchronisation.
18851 Frame 1 Frame 2 Frame 3 Frame 4
18853 11111 22222 33333 44444
18854 11111 22222 33333 44444
18855 11111 22222 33333 44444
18856 11111 22222 33333 44444
18859 11111 22222 22222 33333 33333 44444 44444
18860 11111 11111 22222 22222 33333 33333 44444
18861 11111 22222 22222 33333 33333 44444 44444
18862 11111 11111 22222 22222 33333 33333 44444
18867 Move odd frames into the upper field, even into the lower field,
18868 generating a double height frame at same frame rate.
18873 Frame 1 Frame 2 Frame 3 Frame 4
18875 11111 22222 33333 44444
18876 11111 22222 33333 44444
18877 11111 22222 33333 44444
18878 11111 22222 33333 44444
18881 11111 33333 33333 55555
18882 22222 22222 44444 44444
18883 11111 33333 33333 55555
18884 22222 22222 44444 44444
18885 11111 33333 33333 55555
18886 22222 22222 44444 44444
18887 11111 33333 33333 55555
18888 22222 22222 44444 44444
18893 Numeric values are deprecated but are accepted for backward
18894 compatibility reasons.
18896 Default mode is @code{merge}.
18899 Specify flags influencing the filter process.
18901 Available value for @var{flags} is:
18904 @item low_pass_filter, vlpf
18905 Enable linear vertical low-pass filtering in the filter.
18906 Vertical low-pass filtering is required when creating an interlaced
18907 destination from a progressive source which contains high-frequency
18908 vertical detail. Filtering will reduce interlace 'twitter' and Moire
18911 @item complex_filter, cvlpf
18912 Enable complex vertical low-pass filtering.
18913 This will slightly less reduce interlace 'twitter' and Moire
18914 patterning but better retain detail and subjective sharpness impression.
18917 Bypass already interlaced frames, only adjust the frame rate.
18920 Vertical low-pass filtering and bypassing already interlaced frames can only be
18921 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
18926 Pick median pixels from several successive input video frames.
18928 The filter accepts the following options:
18932 Set radius of median filter.
18933 Default is 1. Allowed range is from 1 to 127.
18936 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
18939 Set median percentile. Default value is @code{0.5}.
18940 Default value of @code{0.5} will pick always median values, while @code{0} will pick
18941 minimum values, and @code{1} maximum values.
18946 Mix successive video frames.
18948 A description of the accepted options follows.
18952 The number of successive frames to mix. If unspecified, it defaults to 3.
18955 Specify weight of each input video frame.
18956 Each weight is separated by space. If number of weights is smaller than
18957 number of @var{frames} last specified weight will be used for all remaining
18961 Specify scale, if it is set it will be multiplied with sum
18962 of each weight multiplied with pixel values to give final destination
18963 pixel value. By default @var{scale} is auto scaled to sum of weights.
18966 @subsection Examples
18970 Average 7 successive frames:
18972 tmix=frames=7:weights="1 1 1 1 1 1 1"
18976 Apply simple temporal convolution:
18978 tmix=frames=3:weights="-1 3 -1"
18982 Similar as above but only showing temporal differences:
18984 tmix=frames=3:weights="-1 2 -1":scale=1
18990 Tone map colors from different dynamic ranges.
18992 This filter expects data in single precision floating point, as it needs to
18993 operate on (and can output) out-of-range values. Another filter, such as
18994 @ref{zscale}, is needed to convert the resulting frame to a usable format.
18996 The tonemapping algorithms implemented only work on linear light, so input
18997 data should be linearized beforehand (and possibly correctly tagged).
19000 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19003 @subsection Options
19004 The filter accepts the following options.
19008 Set the tone map algorithm to use.
19010 Possible values are:
19013 Do not apply any tone map, only desaturate overbright pixels.
19016 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19017 in-range values, while distorting out-of-range values.
19020 Stretch the entire reference gamut to a linear multiple of the display.
19023 Fit a logarithmic transfer between the tone curves.
19026 Preserve overall image brightness with a simple curve, using nonlinear
19027 contrast, which results in flattening details and degrading color accuracy.
19030 Preserve both dark and bright details better than @var{reinhard}, at the cost
19031 of slightly darkening everything. Use it when detail preservation is more
19032 important than color and brightness accuracy.
19035 Smoothly map out-of-range values, while retaining contrast and colors for
19036 in-range material as much as possible. Use it when color accuracy is more
19037 important than detail preservation.
19043 Tune the tone mapping algorithm.
19045 This affects the following algorithms:
19051 Specifies the scale factor to use while stretching.
19055 Specifies the exponent of the function.
19059 Specify an extra linear coefficient to multiply into the signal before clipping.
19063 Specify the local contrast coefficient at the display peak.
19064 Default to 0.5, which means that in-gamut values will be about half as bright
19071 Specify the transition point from linear to mobius transform. Every value
19072 below this point is guaranteed to be mapped 1:1. The higher the value, the
19073 more accurate the result will be, at the cost of losing bright details.
19074 Default to 0.3, which due to the steep initial slope still preserves in-range
19075 colors fairly accurately.
19079 Apply desaturation for highlights that exceed this level of brightness. The
19080 higher the parameter, the more color information will be preserved. This
19081 setting helps prevent unnaturally blown-out colors for super-highlights, by
19082 (smoothly) turning into white instead. This makes images feel more natural,
19083 at the cost of reducing information about out-of-range colors.
19085 The default of 2.0 is somewhat conservative and will mostly just apply to
19086 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19088 This option works only if the input frame has a supported color tag.
19091 Override signal/nominal/reference peak with this value. Useful when the
19092 embedded peak information in display metadata is not reliable or when tone
19093 mapping from a lower range to a higher range.
19098 Temporarily pad video frames.
19100 The filter accepts the following options:
19104 Specify number of delay frames before input video stream. Default is 0.
19107 Specify number of padding frames after input video stream.
19108 Set to -1 to pad indefinitely. Default is 0.
19111 Set kind of frames added to beginning of stream.
19112 Can be either @var{add} or @var{clone}.
19113 With @var{add} frames of solid-color are added.
19114 With @var{clone} frames are clones of first frame.
19115 Default is @var{add}.
19118 Set kind of frames added to end of stream.
19119 Can be either @var{add} or @var{clone}.
19120 With @var{add} frames of solid-color are added.
19121 With @var{clone} frames are clones of last frame.
19122 Default is @var{add}.
19124 @item start_duration, stop_duration
19125 Specify the duration of the start/stop delay. See
19126 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19127 for the accepted syntax.
19128 These options override @var{start} and @var{stop}. Default is 0.
19131 Specify the color of the padded area. For the syntax of this option,
19132 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19133 manual,ffmpeg-utils}.
19135 The default value of @var{color} is "black".
19141 Transpose rows with columns in the input video and optionally flip it.
19143 It accepts the following parameters:
19148 Specify the transposition direction.
19150 Can assume the following values:
19152 @item 0, 4, cclock_flip
19153 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19161 Rotate by 90 degrees clockwise, that is:
19169 Rotate by 90 degrees counterclockwise, that is:
19176 @item 3, 7, clock_flip
19177 Rotate by 90 degrees clockwise and vertically flip, that is:
19185 For values between 4-7, the transposition is only done if the input
19186 video geometry is portrait and not landscape. These values are
19187 deprecated, the @code{passthrough} option should be used instead.
19189 Numerical values are deprecated, and should be dropped in favor of
19190 symbolic constants.
19193 Do not apply the transposition if the input geometry matches the one
19194 specified by the specified value. It accepts the following values:
19197 Always apply transposition.
19199 Preserve portrait geometry (when @var{height} >= @var{width}).
19201 Preserve landscape geometry (when @var{width} >= @var{height}).
19204 Default value is @code{none}.
19207 For example to rotate by 90 degrees clockwise and preserve portrait
19210 transpose=dir=1:passthrough=portrait
19213 The command above can also be specified as:
19215 transpose=1:portrait
19218 @section transpose_npp
19220 Transpose rows with columns in the input video and optionally flip it.
19221 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19223 It accepts the following parameters:
19228 Specify the transposition direction.
19230 Can assume the following values:
19233 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19236 Rotate by 90 degrees clockwise.
19239 Rotate by 90 degrees counterclockwise.
19242 Rotate by 90 degrees clockwise and vertically flip.
19246 Do not apply the transposition if the input geometry matches the one
19247 specified by the specified value. It accepts the following values:
19250 Always apply transposition. (default)
19252 Preserve portrait geometry (when @var{height} >= @var{width}).
19254 Preserve landscape geometry (when @var{width} >= @var{height}).
19260 Trim the input so that the output contains one continuous subpart of the input.
19262 It accepts the following parameters:
19265 Specify the time of the start of the kept section, i.e. the frame with the
19266 timestamp @var{start} will be the first frame in the output.
19269 Specify the time of the first frame that will be dropped, i.e. the frame
19270 immediately preceding the one with the timestamp @var{end} will be the last
19271 frame in the output.
19274 This is the same as @var{start}, except this option sets the start timestamp
19275 in timebase units instead of seconds.
19278 This is the same as @var{end}, except this option sets the end timestamp
19279 in timebase units instead of seconds.
19282 The maximum duration of the output in seconds.
19285 The number of the first frame that should be passed to the output.
19288 The number of the first frame that should be dropped.
19291 @option{start}, @option{end}, and @option{duration} are expressed as time
19292 duration specifications; see
19293 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19294 for the accepted syntax.
19296 Note that the first two sets of the start/end options and the @option{duration}
19297 option look at the frame timestamp, while the _frame variants simply count the
19298 frames that pass through the filter. Also note that this filter does not modify
19299 the timestamps. If you wish for the output timestamps to start at zero, insert a
19300 setpts filter after the trim filter.
19302 If multiple start or end options are set, this filter tries to be greedy and
19303 keep all the frames that match at least one of the specified constraints. To keep
19304 only the part that matches all the constraints at once, chain multiple trim
19307 The defaults are such that all the input is kept. So it is possible to set e.g.
19308 just the end values to keep everything before the specified time.
19313 Drop everything except the second minute of input:
19315 ffmpeg -i INPUT -vf trim=60:120
19319 Keep only the first second:
19321 ffmpeg -i INPUT -vf trim=duration=1
19326 @section unpremultiply
19327 Apply alpha unpremultiply effect to input video stream using first plane
19328 of second stream as alpha.
19330 Both streams must have same dimensions and same pixel format.
19332 The filter accepts the following option:
19336 Set which planes will be processed, unprocessed planes will be copied.
19337 By default value 0xf, all planes will be processed.
19339 If the format has 1 or 2 components, then luma is bit 0.
19340 If the format has 3 or 4 components:
19341 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19342 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19343 If present, the alpha channel is always the last bit.
19346 Do not require 2nd input for processing, instead use alpha plane from input stream.
19352 Sharpen or blur the input video.
19354 It accepts the following parameters:
19357 @item luma_msize_x, lx
19358 Set the luma matrix horizontal size. It must be an odd integer between
19359 3 and 23. The default value is 5.
19361 @item luma_msize_y, ly
19362 Set the luma matrix vertical size. It must be an odd integer between 3
19363 and 23. The default value is 5.
19365 @item luma_amount, la
19366 Set the luma effect strength. It must be a floating point number, reasonable
19367 values lay between -1.5 and 1.5.
19369 Negative values will blur the input video, while positive values will
19370 sharpen it, a value of zero will disable the effect.
19372 Default value is 1.0.
19374 @item chroma_msize_x, cx
19375 Set the chroma matrix horizontal size. It must be an odd integer
19376 between 3 and 23. The default value is 5.
19378 @item chroma_msize_y, cy
19379 Set the chroma matrix vertical size. It must be an odd integer
19380 between 3 and 23. The default value is 5.
19382 @item chroma_amount, ca
19383 Set the chroma effect strength. It must be a floating point number, reasonable
19384 values lay between -1.5 and 1.5.
19386 Negative values will blur the input video, while positive values will
19387 sharpen it, a value of zero will disable the effect.
19389 Default value is 0.0.
19393 All parameters are optional and default to the equivalent of the
19394 string '5:5:1.0:5:5:0.0'.
19396 @subsection Examples
19400 Apply strong luma sharpen effect:
19402 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19406 Apply a strong blur of both luma and chroma parameters:
19408 unsharp=7:7:-2:7:7:-2
19415 Decompose a video made of tiled images into the individual images.
19417 The frame rate of the output video is the frame rate of the input video
19418 multiplied by the number of tiles.
19420 This filter does the reverse of @ref{tile}.
19422 The filter accepts the following options:
19427 Set the grid size (i.e. the number of lines and columns). For the syntax of
19428 this option, check the
19429 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19432 @subsection Examples
19436 Produce a 1-second video from a still image file made of 25 frames stacked
19437 vertically, like an analogic film reel:
19439 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19445 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19446 the image at several (or - in the case of @option{quality} level @code{8} - all)
19447 shifts and average the results.
19449 The way this differs from the behavior of spp is that uspp actually encodes &
19450 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19451 DCT similar to MJPEG.
19453 The filter accepts the following options:
19457 Set quality. This option defines the number of levels for averaging. It accepts
19458 an integer in the range 0-8. If set to @code{0}, the filter will have no
19459 effect. A value of @code{8} means the higher quality. For each increment of
19460 that value the speed drops by a factor of approximately 2. Default value is
19464 Force a constant quantization parameter. If not set, the filter will use the QP
19465 from the video stream (if available).
19470 Convert 360 videos between various formats.
19472 The filter accepts the following options:
19478 Set format of the input/output video.
19486 Equirectangular projection.
19491 Cubemap with 3x2/6x1/1x6 layout.
19493 Format specific options:
19498 Set padding proportion for the input/output cubemap. Values in decimals.
19505 1% of face is padding. For example, with 1920x1280 resolution face size would be 640x640 and padding would be 3 pixels from each side. (640 * 0.01 = 6 pixels)
19508 Default value is @b{@samp{0}}.
19509 Maximum value is @b{@samp{0.1}}.
19513 Set fixed padding for the input/output cubemap. Values in pixels.
19515 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19519 Set order of faces for the input/output cubemap. Choose one direction for each position.
19521 Designation of directions:
19537 Default value is @b{@samp{rludfb}}.
19541 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19543 Designation of angles:
19546 0 degrees clockwise
19548 90 degrees clockwise
19550 180 degrees clockwise
19552 270 degrees clockwise
19555 Default value is @b{@samp{000000}}.
19559 Equi-Angular Cubemap.
19566 Format specific options:
19571 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19573 If diagonal field of view is set it overrides horizontal and vertical field of view.
19578 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19580 If diagonal field of view is set it overrides horizontal and vertical field of view.
19586 Format specific options:
19591 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19593 If diagonal field of view is set it overrides horizontal and vertical field of view.
19598 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19600 If diagonal field of view is set it overrides horizontal and vertical field of view.
19606 Facebook's 360 formats.
19609 Stereographic format.
19611 Format specific options:
19616 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19618 If diagonal field of view is set it overrides horizontal and vertical field of view.
19623 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19625 If diagonal field of view is set it overrides horizontal and vertical field of view.
19632 Ball format, gives significant distortion toward the back.
19635 Hammer-Aitoff map projection format.
19638 Sinusoidal map projection format.
19641 Fisheye projection.
19643 Format specific options:
19648 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19650 If diagonal field of view is set it overrides horizontal and vertical field of view.
19655 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19657 If diagonal field of view is set it overrides horizontal and vertical field of view.
19661 Pannini projection.
19663 Format specific options:
19666 Set output pannini parameter.
19669 Set input pannini parameter.
19673 Cylindrical projection.
19675 Format specific options:
19680 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19682 If diagonal field of view is set it overrides horizontal and vertical field of view.
19687 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19689 If diagonal field of view is set it overrides horizontal and vertical field of view.
19693 Perspective projection. @i{(output only)}
19695 Format specific options:
19698 Set perspective parameter.
19702 Tetrahedron projection.
19705 Truncated square pyramid projection.
19709 Half equirectangular projection.
19714 Format specific options:
19719 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19721 If diagonal field of view is set it overrides horizontal and vertical field of view.
19726 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19728 If diagonal field of view is set it overrides horizontal and vertical field of view.
19732 Orthographic format.
19734 Format specific options:
19739 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19741 If diagonal field of view is set it overrides horizontal and vertical field of view.
19746 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19748 If diagonal field of view is set it overrides horizontal and vertical field of view.
19752 Octahedron projection.
19756 Set interpolation method.@*
19757 @i{Note: more complex interpolation methods require much more memory to run.}
19767 Bilinear interpolation.
19769 Lagrange9 interpolation.
19772 Bicubic interpolation.
19775 Lanczos interpolation.
19778 Spline16 interpolation.
19781 Gaussian interpolation.
19783 Mitchell interpolation.
19786 Default value is @b{@samp{line}}.
19790 Set the output video resolution.
19792 Default resolution depends on formats.
19796 Set the input/output stereo format.
19807 Default value is @b{@samp{2d}} for input and output format.
19812 Set rotation for the output video. Values in degrees.
19815 Set rotation order for the output video. Choose one item for each position.
19826 Default value is @b{@samp{ypr}}.
19831 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
19835 Set if input video is flipped horizontally/vertically. Boolean values.
19838 Set if input video is transposed. Boolean value, by default disabled.
19841 Set if output video needs to be transposed. Boolean value, by default disabled.
19844 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
19847 @subsection Examples
19851 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
19853 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
19856 Extract back view of Equi-Angular Cubemap:
19858 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
19861 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
19863 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
19867 @subsection Commands
19869 This filter supports subset of above options as @ref{commands}.
19871 @section vaguedenoiser
19873 Apply a wavelet based denoiser.
19875 It transforms each frame from the video input into the wavelet domain,
19876 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
19877 the obtained coefficients. It does an inverse wavelet transform after.
19878 Due to wavelet properties, it should give a nice smoothed result, and
19879 reduced noise, without blurring picture features.
19881 This filter accepts the following options:
19885 The filtering strength. The higher, the more filtered the video will be.
19886 Hard thresholding can use a higher threshold than soft thresholding
19887 before the video looks overfiltered. Default value is 2.
19890 The filtering method the filter will use.
19892 It accepts the following values:
19895 All values under the threshold will be zeroed.
19898 All values under the threshold will be zeroed. All values above will be
19899 reduced by the threshold.
19902 Scales or nullifies coefficients - intermediary between (more) soft and
19903 (less) hard thresholding.
19906 Default is garrote.
19909 Number of times, the wavelet will decompose the picture. Picture can't
19910 be decomposed beyond a particular point (typically, 8 for a 640x480
19911 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
19914 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
19917 A list of the planes to process. By default all planes are processed.
19920 The threshold type the filter will use.
19922 It accepts the following values:
19925 Threshold used is same for all decompositions.
19928 Threshold used depends also on each decomposition coefficients.
19931 Default is universal.
19934 @section vectorscope
19936 Display 2 color component values in the two dimensional graph (which is called
19939 This filter accepts the following options:
19943 Set vectorscope mode.
19945 It accepts the following values:
19949 Gray values are displayed on graph, higher brightness means more pixels have
19950 same component color value on location in graph. This is the default mode.
19953 Gray values are displayed on graph. Surrounding pixels values which are not
19954 present in video frame are drawn in gradient of 2 color components which are
19955 set by option @code{x} and @code{y}. The 3rd color component is static.
19958 Actual color components values present in video frame are displayed on graph.
19961 Similar as color2 but higher frequency of same values @code{x} and @code{y}
19962 on graph increases value of another color component, which is luminance by
19963 default values of @code{x} and @code{y}.
19966 Actual colors present in video frame are displayed on graph. If two different
19967 colors map to same position on graph then color with higher value of component
19968 not present in graph is picked.
19971 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
19972 component picked from radial gradient.
19976 Set which color component will be represented on X-axis. Default is @code{1}.
19979 Set which color component will be represented on Y-axis. Default is @code{2}.
19982 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
19983 of color component which represents frequency of (X, Y) location in graph.
19988 No envelope, this is default.
19991 Instant envelope, even darkest single pixel will be clearly highlighted.
19994 Hold maximum and minimum values presented in graph over time. This way you
19995 can still spot out of range values without constantly looking at vectorscope.
19998 Peak and instant envelope combined together.
20002 Set what kind of graticule to draw.
20011 Set graticule opacity.
20014 Set graticule flags.
20018 Draw graticule for white point.
20021 Draw graticule for black point.
20024 Draw color points short names.
20028 Set background opacity.
20030 @item lthreshold, l
20031 Set low threshold for color component not represented on X or Y axis.
20032 Values lower than this value will be ignored. Default is 0.
20033 Note this value is multiplied with actual max possible value one pixel component
20034 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20037 @item hthreshold, h
20038 Set high threshold for color component not represented on X or Y axis.
20039 Values higher than this value will be ignored. Default is 1.
20040 Note this value is multiplied with actual max possible value one pixel component
20041 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20042 is 0.9 * 255 = 230.
20044 @item colorspace, c
20045 Set what kind of colorspace to use when drawing graticule.
20055 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20056 This means no tint, and output will remain gray.
20059 @anchor{vidstabdetect}
20060 @section vidstabdetect
20062 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20063 @ref{vidstabtransform} for pass 2.
20065 This filter generates a file with relative translation and rotation
20066 transform information about subsequent frames, which is then used by
20067 the @ref{vidstabtransform} filter.
20069 To enable compilation of this filter you need to configure FFmpeg with
20070 @code{--enable-libvidstab}.
20072 This filter accepts the following options:
20076 Set the path to the file used to write the transforms information.
20077 Default value is @file{transforms.trf}.
20080 Set how shaky the video is and how quick the camera is. It accepts an
20081 integer in the range 1-10, a value of 1 means little shakiness, a
20082 value of 10 means strong shakiness. Default value is 5.
20085 Set the accuracy of the detection process. It must be a value in the
20086 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20087 accuracy. Default value is 15.
20090 Set stepsize of the search process. The region around minimum is
20091 scanned with 1 pixel resolution. Default value is 6.
20094 Set minimum contrast. Below this value a local measurement field is
20095 discarded. Must be a floating point value in the range 0-1. Default
20099 Set reference frame number for tripod mode.
20101 If enabled, the motion of the frames is compared to a reference frame
20102 in the filtered stream, identified by the specified number. The idea
20103 is to compensate all movements in a more-or-less static scene and keep
20104 the camera view absolutely still.
20106 If set to 0, it is disabled. The frames are counted starting from 1.
20109 Show fields and transforms in the resulting frames. It accepts an
20110 integer in the range 0-2. Default value is 0, which disables any
20114 @subsection Examples
20118 Use default values:
20124 Analyze strongly shaky movie and put the results in file
20125 @file{mytransforms.trf}:
20127 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20131 Visualize the result of internal transformations in the resulting
20134 vidstabdetect=show=1
20138 Analyze a video with medium shakiness using @command{ffmpeg}:
20140 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20144 @anchor{vidstabtransform}
20145 @section vidstabtransform
20147 Video stabilization/deshaking: pass 2 of 2,
20148 see @ref{vidstabdetect} for pass 1.
20150 Read a file with transform information for each frame and
20151 apply/compensate them. Together with the @ref{vidstabdetect}
20152 filter this can be used to deshake videos. See also
20153 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20154 the @ref{unsharp} filter, see below.
20156 To enable compilation of this filter you need to configure FFmpeg with
20157 @code{--enable-libvidstab}.
20159 @subsection Options
20163 Set path to the file used to read the transforms. Default value is
20164 @file{transforms.trf}.
20167 Set the number of frames (value*2 + 1) used for lowpass filtering the
20168 camera movements. Default value is 10.
20170 For example a number of 10 means that 21 frames are used (10 in the
20171 past and 10 in the future) to smoothen the motion in the video. A
20172 larger value leads to a smoother video, but limits the acceleration of
20173 the camera (pan/tilt movements). 0 is a special case where a static
20174 camera is simulated.
20177 Set the camera path optimization algorithm.
20179 Accepted values are:
20182 gaussian kernel low-pass filter on camera motion (default)
20184 averaging on transformations
20188 Set maximal number of pixels to translate frames. Default value is -1,
20192 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20193 value is -1, meaning no limit.
20196 Specify how to deal with borders that may be visible due to movement
20199 Available values are:
20202 keep image information from previous frame (default)
20204 fill the border black
20208 Invert transforms if set to 1. Default value is 0.
20211 Consider transforms as relative to previous frame if set to 1,
20212 absolute if set to 0. Default value is 0.
20215 Set percentage to zoom. A positive value will result in a zoom-in
20216 effect, a negative value in a zoom-out effect. Default value is 0 (no
20220 Set optimal zooming to avoid borders.
20222 Accepted values are:
20227 optimal static zoom value is determined (only very strong movements
20228 will lead to visible borders) (default)
20230 optimal adaptive zoom value is determined (no borders will be
20231 visible), see @option{zoomspeed}
20234 Note that the value given at zoom is added to the one calculated here.
20237 Set percent to zoom maximally each frame (enabled when
20238 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20242 Specify type of interpolation.
20244 Available values are:
20249 linear only horizontal
20251 linear in both directions (default)
20253 cubic in both directions (slow)
20257 Enable virtual tripod mode if set to 1, which is equivalent to
20258 @code{relative=0:smoothing=0}. Default value is 0.
20260 Use also @code{tripod} option of @ref{vidstabdetect}.
20263 Increase log verbosity if set to 1. Also the detected global motions
20264 are written to the temporary file @file{global_motions.trf}. Default
20268 @subsection Examples
20272 Use @command{ffmpeg} for a typical stabilization with default values:
20274 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20277 Note the use of the @ref{unsharp} filter which is always recommended.
20280 Zoom in a bit more and load transform data from a given file:
20282 vidstabtransform=zoom=5:input="mytransforms.trf"
20286 Smoothen the video even more:
20288 vidstabtransform=smoothing=30
20294 Flip the input video vertically.
20296 For example, to vertically flip a video with @command{ffmpeg}:
20298 ffmpeg -i in.avi -vf "vflip" out.avi
20303 Detect variable frame rate video.
20305 This filter tries to detect if the input is variable or constant frame rate.
20307 At end it will output number of frames detected as having variable delta pts,
20308 and ones with constant delta pts.
20309 If there was frames with variable delta, than it will also show min, max and
20310 average delta encountered.
20314 Boost or alter saturation.
20316 The filter accepts the following options:
20319 Set strength of boost if positive value or strength of alter if negative value.
20320 Default is 0. Allowed range is from -2 to 2.
20323 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20326 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20329 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20332 Set the red luma coefficient.
20335 Set the green luma coefficient.
20338 Set the blue luma coefficient.
20341 If @code{intensity} is negative and this is set to 1, colors will change,
20342 otherwise colors will be less saturated, more towards gray.
20345 @subsection Commands
20347 This filter supports the all above options as @ref{commands}.
20352 Make or reverse a natural vignetting effect.
20354 The filter accepts the following options:
20358 Set lens angle expression as a number of radians.
20360 The value is clipped in the @code{[0,PI/2]} range.
20362 Default value: @code{"PI/5"}
20366 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20370 Set forward/backward mode.
20372 Available modes are:
20375 The larger the distance from the central point, the darker the image becomes.
20378 The larger the distance from the central point, the brighter the image becomes.
20379 This can be used to reverse a vignette effect, though there is no automatic
20380 detection to extract the lens @option{angle} and other settings (yet). It can
20381 also be used to create a burning effect.
20384 Default value is @samp{forward}.
20387 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20389 It accepts the following values:
20392 Evaluate expressions only once during the filter initialization.
20395 Evaluate expressions for each incoming frame. This is way slower than the
20396 @samp{init} mode since it requires all the scalers to be re-computed, but it
20397 allows advanced dynamic expressions.
20400 Default value is @samp{init}.
20403 Set dithering to reduce the circular banding effects. Default is @code{1}
20407 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20408 Setting this value to the SAR of the input will make a rectangular vignetting
20409 following the dimensions of the video.
20411 Default is @code{1/1}.
20414 @subsection Expressions
20416 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20417 following parameters.
20422 input width and height
20425 the number of input frame, starting from 0
20428 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20429 @var{TB} units, NAN if undefined
20432 frame rate of the input video, NAN if the input frame rate is unknown
20435 the PTS (Presentation TimeStamp) of the filtered video frame,
20436 expressed in seconds, NAN if undefined
20439 time base of the input video
20443 @subsection Examples
20447 Apply simple strong vignetting effect:
20453 Make a flickering vignetting:
20455 vignette='PI/4+random(1)*PI/50':eval=frame
20460 @section vmafmotion
20462 Obtain the average VMAF motion score of a video.
20463 It is one of the component metrics of VMAF.
20465 The obtained average motion score is printed through the logging system.
20467 The filter accepts the following options:
20471 If specified, the filter will use the named file to save the motion score of
20472 each frame with respect to the previous frame.
20473 When filename equals "-" the data is sent to standard output.
20478 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20482 Stack input videos vertically.
20484 All streams must be of same pixel format and of same width.
20486 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20487 to create same output.
20489 The filter accepts the following options:
20493 Set number of input streams. Default is 2.
20496 If set to 1, force the output to terminate when the shortest input
20497 terminates. Default value is 0.
20502 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20503 Deinterlacing Filter").
20505 Based on the process described by Martin Weston for BBC R&D, and
20506 implemented based on the de-interlace algorithm written by Jim
20507 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20508 uses filter coefficients calculated by BBC R&D.
20510 This filter uses field-dominance information in frame to decide which
20511 of each pair of fields to place first in the output.
20512 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20514 There are two sets of filter coefficients, so called "simple"
20515 and "complex". Which set of filter coefficients is used can
20516 be set by passing an optional parameter:
20520 Set the interlacing filter coefficients. Accepts one of the following values:
20524 Simple filter coefficient set.
20526 More-complex filter coefficient set.
20528 Default value is @samp{complex}.
20531 Specify which frames to deinterlace. Accepts one of the following values:
20535 Deinterlace all frames,
20537 Only deinterlace frames marked as interlaced.
20540 Default value is @samp{all}.
20544 Video waveform monitor.
20546 The waveform monitor plots color component intensity. By default luminance
20547 only. Each column of the waveform corresponds to a column of pixels in the
20550 It accepts the following options:
20554 Can be either @code{row}, or @code{column}. Default is @code{column}.
20555 In row mode, the graph on the left side represents color component value 0 and
20556 the right side represents value = 255. In column mode, the top side represents
20557 color component value = 0 and bottom side represents value = 255.
20560 Set intensity. Smaller values are useful to find out how many values of the same
20561 luminance are distributed across input rows/columns.
20562 Default value is @code{0.04}. Allowed range is [0, 1].
20565 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20566 In mirrored mode, higher values will be represented on the left
20567 side for @code{row} mode and at the top for @code{column} mode. Default is
20568 @code{1} (mirrored).
20572 It accepts the following values:
20575 Presents information identical to that in the @code{parade}, except
20576 that the graphs representing color components are superimposed directly
20579 This display mode makes it easier to spot relative differences or similarities
20580 in overlapping areas of the color components that are supposed to be identical,
20581 such as neutral whites, grays, or blacks.
20584 Display separate graph for the color components side by side in
20585 @code{row} mode or one below the other in @code{column} mode.
20588 Display separate graph for the color components side by side in
20589 @code{column} mode or one below the other in @code{row} mode.
20591 Using this display mode makes it easy to spot color casts in the highlights
20592 and shadows of an image, by comparing the contours of the top and the bottom
20593 graphs of each waveform. Since whites, grays, and blacks are characterized
20594 by exactly equal amounts of red, green, and blue, neutral areas of the picture
20595 should display three waveforms of roughly equal width/height. If not, the
20596 correction is easy to perform by making level adjustments the three waveforms.
20598 Default is @code{stack}.
20600 @item components, c
20601 Set which color components to display. Default is 1, which means only luminance
20602 or red color component if input is in RGB colorspace. If is set for example to
20603 7 it will display all 3 (if) available color components.
20608 No envelope, this is default.
20611 Instant envelope, minimum and maximum values presented in graph will be easily
20612 visible even with small @code{step} value.
20615 Hold minimum and maximum values presented in graph across time. This way you
20616 can still spot out of range values without constantly looking at waveforms.
20619 Peak and instant envelope combined together.
20625 No filtering, this is default.
20628 Luma and chroma combined together.
20631 Similar as above, but shows difference between blue and red chroma.
20634 Similar as above, but use different colors.
20637 Similar as above, but again with different colors.
20640 Displays only chroma.
20643 Displays actual color value on waveform.
20646 Similar as above, but with luma showing frequency of chroma values.
20650 Set which graticule to display.
20654 Do not display graticule.
20657 Display green graticule showing legal broadcast ranges.
20660 Display orange graticule showing legal broadcast ranges.
20663 Display invert graticule showing legal broadcast ranges.
20667 Set graticule opacity.
20670 Set graticule flags.
20674 Draw numbers above lines. By default enabled.
20677 Draw dots instead of lines.
20681 Set scale used for displaying graticule.
20688 Default is digital.
20691 Set background opacity.
20695 Set tint for output.
20696 Only used with lowpass filter and when display is not overlay and input
20697 pixel formats are not RGB.
20700 @section weave, doubleweave
20702 The @code{weave} takes a field-based video input and join
20703 each two sequential fields into single frame, producing a new double
20704 height clip with half the frame rate and half the frame count.
20706 The @code{doubleweave} works same as @code{weave} but without
20707 halving frame rate and frame count.
20709 It accepts the following option:
20713 Set first field. Available values are:
20717 Set the frame as top-field-first.
20720 Set the frame as bottom-field-first.
20724 @subsection Examples
20728 Interlace video using @ref{select} and @ref{separatefields} filter:
20730 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
20735 Apply the xBR high-quality magnification filter which is designed for pixel
20736 art. It follows a set of edge-detection rules, see
20737 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
20739 It accepts the following option:
20743 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
20744 @code{3xBR} and @code{4} for @code{4xBR}.
20745 Default is @code{3}.
20750 Apply cross fade from one input video stream to another input video stream.
20751 The cross fade is applied for specified duration.
20753 The filter accepts the following options:
20757 Set one of available transition effects:
20805 Default transition effect is fade.
20808 Set cross fade duration in seconds.
20809 Default duration is 1 second.
20812 Set cross fade start relative to first input stream in seconds.
20813 Default offset is 0.
20816 Set expression for custom transition effect.
20818 The expressions can use the following variables and functions:
20823 The coordinates of the current sample.
20827 The width and height of the image.
20830 Progress of transition effect.
20833 Currently processed plane.
20836 Return value of first input at current location and plane.
20839 Return value of second input at current location and plane.
20845 Return the value of the pixel at location (@var{x},@var{y}) of the
20846 first/second/third/fourth component of first input.
20852 Return the value of the pixel at location (@var{x},@var{y}) of the
20853 first/second/third/fourth component of second input.
20857 @subsection Examples
20861 Cross fade from one input video to another input video, with fade transition and duration of transition
20862 of 2 seconds starting at offset of 5 seconds:
20864 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
20869 Pick median pixels from several input videos.
20871 The filter accepts the following options:
20875 Set number of inputs.
20876 Default is 3. Allowed range is from 3 to 255.
20877 If number of inputs is even number, than result will be mean value between two median values.
20880 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
20883 Set median percentile. Default value is @code{0.5}.
20884 Default value of @code{0.5} will pick always median values, while @code{0} will pick
20885 minimum values, and @code{1} maximum values.
20889 Stack video inputs into custom layout.
20891 All streams must be of same pixel format.
20893 The filter accepts the following options:
20897 Set number of input streams. Default is 2.
20900 Specify layout of inputs.
20901 This option requires the desired layout configuration to be explicitly set by the user.
20902 This sets position of each video input in output. Each input
20903 is separated by '|'.
20904 The first number represents the column, and the second number represents the row.
20905 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
20906 where X is video input from which to take width or height.
20907 Multiple values can be used when separated by '+'. In such
20908 case values are summed together.
20910 Note that if inputs are of different sizes gaps may appear, as not all of
20911 the output video frame will be filled. Similarly, videos can overlap each
20912 other if their position doesn't leave enough space for the full frame of
20915 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
20916 a layout must be set by the user.
20919 If set to 1, force the output to terminate when the shortest input
20920 terminates. Default value is 0.
20923 If set to valid color, all unused pixels will be filled with that color.
20924 By default fill is set to none, so it is disabled.
20927 @subsection Examples
20931 Display 4 inputs into 2x2 grid.
20935 input1(0, 0) | input3(w0, 0)
20936 input2(0, h0) | input4(w0, h0)
20940 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
20943 Note that if inputs are of different sizes, gaps or overlaps may occur.
20946 Display 4 inputs into 1x4 grid.
20953 input4(0, h0+h1+h2)
20957 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
20960 Note that if inputs are of different widths, unused space will appear.
20963 Display 9 inputs into 3x3 grid.
20967 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
20968 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
20969 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
20973 xstack=inputs=9:layout=0_0|0_h0|0_h0+h1|w0_0|w0_h0|w0_h0+h1|w0+w3_0|w0+w3_h0|w0+w3_h0+h1
20976 Note that if inputs are of different sizes, gaps or overlaps may occur.
20979 Display 16 inputs into 4x4 grid.
20983 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
20984 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
20985 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
20986 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
20990 xstack=inputs=16:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2|w0_0|w0_h0|w0_h0+h1|w0_h0+h1+h2|w0+w4_0|
20991 w0+w4_h0|w0+w4_h0+h1|w0+w4_h0+h1+h2|w0+w4+w8_0|w0+w4+w8_h0|w0+w4+w8_h0+h1|w0+w4+w8_h0+h1+h2
20994 Note that if inputs are of different sizes, gaps or overlaps may occur.
21001 Deinterlace the input video ("yadif" means "yet another deinterlacing
21004 It accepts the following parameters:
21010 The interlacing mode to adopt. It accepts one of the following values:
21013 @item 0, send_frame
21014 Output one frame for each frame.
21015 @item 1, send_field
21016 Output one frame for each field.
21017 @item 2, send_frame_nospatial
21018 Like @code{send_frame}, but it skips the spatial interlacing check.
21019 @item 3, send_field_nospatial
21020 Like @code{send_field}, but it skips the spatial interlacing check.
21023 The default value is @code{send_frame}.
21026 The picture field parity assumed for the input interlaced video. It accepts one
21027 of the following values:
21031 Assume the top field is first.
21033 Assume the bottom field is first.
21035 Enable automatic detection of field parity.
21038 The default value is @code{auto}.
21039 If the interlacing is unknown or the decoder does not export this information,
21040 top field first will be assumed.
21043 Specify which frames to deinterlace. Accepts one of the following
21048 Deinterlace all frames.
21049 @item 1, interlaced
21050 Only deinterlace frames marked as interlaced.
21053 The default value is @code{all}.
21056 @section yadif_cuda
21058 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21059 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21062 It accepts the following parameters:
21068 The interlacing mode to adopt. It accepts one of the following values:
21071 @item 0, send_frame
21072 Output one frame for each frame.
21073 @item 1, send_field
21074 Output one frame for each field.
21075 @item 2, send_frame_nospatial
21076 Like @code{send_frame}, but it skips the spatial interlacing check.
21077 @item 3, send_field_nospatial
21078 Like @code{send_field}, but it skips the spatial interlacing check.
21081 The default value is @code{send_frame}.
21084 The picture field parity assumed for the input interlaced video. It accepts one
21085 of the following values:
21089 Assume the top field is first.
21091 Assume the bottom field is first.
21093 Enable automatic detection of field parity.
21096 The default value is @code{auto}.
21097 If the interlacing is unknown or the decoder does not export this information,
21098 top field first will be assumed.
21101 Specify which frames to deinterlace. Accepts one of the following
21106 Deinterlace all frames.
21107 @item 1, interlaced
21108 Only deinterlace frames marked as interlaced.
21111 The default value is @code{all}.
21116 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21117 The algorithm is described in
21118 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21120 It accepts the following parameters:
21124 Set the window radius. Default value is 3.
21127 Set which planes to filter. Default is only the first plane.
21130 Set blur strength. Default value is 128.
21133 @subsection Commands
21134 This filter supports same @ref{commands} as options.
21138 Apply Zoom & Pan effect.
21140 This filter accepts the following options:
21144 Set the zoom expression. Range is 1-10. Default is 1.
21148 Set the x and y expression. Default is 0.
21151 Set the duration expression in number of frames.
21152 This sets for how many number of frames effect will last for
21153 single input image.
21156 Set the output image size, default is 'hd720'.
21159 Set the output frame rate, default is '25'.
21162 Each expression can contain the following constants:
21181 Output frame count.
21184 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21186 @item out_time, time, ot
21187 The output timestamp expressed in seconds.
21191 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21192 for current input frame.
21196 'x' and 'y' of last output frame of previous input frame or 0 when there was
21197 not yet such frame (first input frame).
21200 Last calculated zoom from 'z' expression for current input frame.
21203 Last calculated zoom of last output frame of previous input frame.
21206 Number of output frames for current input frame. Calculated from 'd' expression
21207 for each input frame.
21210 number of output frames created for previous input frame
21213 Rational number: input width / input height
21216 sample aspect ratio
21219 display aspect ratio
21223 @subsection Examples
21227 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21229 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
21233 Zoom in up to 1.5x and pan always at center of picture:
21235 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21239 Same as above but without pausing:
21241 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21245 Zoom in 2x into center of picture only for the first second of the input video:
21247 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21254 Scale (resize) the input video, using the z.lib library:
21255 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21256 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21258 The zscale filter forces the output display aspect ratio to be the same
21259 as the input, by changing the output sample aspect ratio.
21261 If the input image format is different from the format requested by
21262 the next filter, the zscale filter will convert the input to the
21265 @subsection Options
21266 The filter accepts the following options.
21271 Set the output video dimension expression. Default value is the input
21274 If the @var{width} or @var{w} value is 0, the input width is used for
21275 the output. If the @var{height} or @var{h} value is 0, the input height
21276 is used for the output.
21278 If one and only one of the values is -n with n >= 1, the zscale filter
21279 will use a value that maintains the aspect ratio of the input image,
21280 calculated from the other specified dimension. After that it will,
21281 however, make sure that the calculated dimension is divisible by n and
21282 adjust the value if necessary.
21284 If both values are -n with n >= 1, the behavior will be identical to
21285 both values being set to 0 as previously detailed.
21287 See below for the list of accepted constants for use in the dimension
21291 Set the video size. For the syntax of this option, check the
21292 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21295 Set the dither type.
21297 Possible values are:
21302 @item error_diffusion
21308 Set the resize filter type.
21310 Possible values are:
21320 Default is bilinear.
21323 Set the color range.
21325 Possible values are:
21332 Default is same as input.
21335 Set the color primaries.
21337 Possible values are:
21347 Default is same as input.
21350 Set the transfer characteristics.
21352 Possible values are:
21366 Default is same as input.
21369 Set the colorspace matrix.
21371 Possible value are:
21382 Default is same as input.
21385 Set the input color range.
21387 Possible values are:
21394 Default is same as input.
21396 @item primariesin, pin
21397 Set the input color primaries.
21399 Possible values are:
21409 Default is same as input.
21411 @item transferin, tin
21412 Set the input transfer characteristics.
21414 Possible values are:
21425 Default is same as input.
21427 @item matrixin, min
21428 Set the input colorspace matrix.
21430 Possible value are:
21442 Set the output chroma location.
21444 Possible values are:
21455 @item chromalin, cin
21456 Set the input chroma location.
21458 Possible values are:
21470 Set the nominal peak luminance.
21473 The values of the @option{w} and @option{h} options are expressions
21474 containing the following constants:
21479 The input width and height
21483 These are the same as @var{in_w} and @var{in_h}.
21487 The output (scaled) width and height
21491 These are the same as @var{out_w} and @var{out_h}
21494 The same as @var{iw} / @var{ih}
21497 input sample aspect ratio
21500 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21504 horizontal and vertical input chroma subsample values. For example for the
21505 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21509 horizontal and vertical output chroma subsample values. For example for the
21510 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21513 @subsection Commands
21515 This filter supports the following commands:
21519 Set the output video dimension expression.
21520 The command accepts the same syntax of the corresponding option.
21522 If the specified expression is not valid, it is kept at its current
21526 @c man end VIDEO FILTERS
21528 @chapter OpenCL Video Filters
21529 @c man begin OPENCL VIDEO FILTERS
21531 Below is a description of the currently available OpenCL video filters.
21533 To enable compilation of these filters you need to configure FFmpeg with
21534 @code{--enable-opencl}.
21536 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21539 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21540 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21541 given device parameters.
21543 @item -filter_hw_device @var{name}
21544 Pass the hardware device called @var{name} to all filters in any filter graph.
21548 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21552 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21554 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21558 Since OpenCL filters are not able to access frame data in normal memory, all frame data needs to be uploaded(@ref{hwupload}) to hardware surfaces connected to the appropriate device before being used and then downloaded(@ref{hwdownload}) back to normal memory. Note that @ref{hwupload} will upload to a surface with the same layout as the software frame, so it may be necessary to add a @ref{format} filter immediately before to get the input into the right format and @ref{hwdownload} does not support all formats on the output - it may be necessary to insert an additional @ref{format} filter immediately following in the graph to get the output in a supported format.
21560 @section avgblur_opencl
21562 Apply average blur filter.
21564 The filter accepts the following options:
21568 Set horizontal radius size.
21569 Range is @code{[1, 1024]} and default value is @code{1}.
21572 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21575 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21578 @subsection Example
21582 Apply average blur filter with horizontal and vertical size of 3, setting each pixel of the output to the average value of the 7x7 region centered on it in the input. For pixels on the edges of the image, the region does not extend beyond the image boundaries, and so out-of-range coordinates are not used in the calculations.
21584 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21588 @section boxblur_opencl
21590 Apply a boxblur algorithm to the input video.
21592 It accepts the following parameters:
21596 @item luma_radius, lr
21597 @item luma_power, lp
21598 @item chroma_radius, cr
21599 @item chroma_power, cp
21600 @item alpha_radius, ar
21601 @item alpha_power, ap
21605 A description of the accepted options follows.
21608 @item luma_radius, lr
21609 @item chroma_radius, cr
21610 @item alpha_radius, ar
21611 Set an expression for the box radius in pixels used for blurring the
21612 corresponding input plane.
21614 The radius value must be a non-negative number, and must not be
21615 greater than the value of the expression @code{min(w,h)/2} for the
21616 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
21619 Default value for @option{luma_radius} is "2". If not specified,
21620 @option{chroma_radius} and @option{alpha_radius} default to the
21621 corresponding value set for @option{luma_radius}.
21623 The expressions can contain the following constants:
21627 The input width and height in pixels.
21631 The input chroma image width and height in pixels.
21635 The horizontal and vertical chroma subsample values. For example, for the
21636 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
21639 @item luma_power, lp
21640 @item chroma_power, cp
21641 @item alpha_power, ap
21642 Specify how many times the boxblur filter is applied to the
21643 corresponding plane.
21645 Default value for @option{luma_power} is 2. If not specified,
21646 @option{chroma_power} and @option{alpha_power} default to the
21647 corresponding value set for @option{luma_power}.
21649 A value of 0 will disable the effect.
21652 @subsection Examples
21654 Apply boxblur filter, setting each pixel of the output to the average value of box-radiuses @var{luma_radius}, @var{chroma_radius}, @var{alpha_radius} for each plane respectively. The filter will apply @var{luma_power}, @var{chroma_power}, @var{alpha_power} times onto the corresponding plane. For pixels on the edges of the image, the radius does not extend beyond the image boundaries, and so out-of-range coordinates are not used in the calculations.
21658 Apply a boxblur filter with the luma, chroma, and alpha radius
21659 set to 2 and luma, chroma, and alpha power set to 3. The filter will run 3 times with box-radius set to 2 for every plane of the image.
21661 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
21662 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
21666 Apply a boxblur filter with luma radius set to 2, luma_power to 1, chroma_radius to 4, chroma_power to 5, alpha_radius to 3 and alpha_power to 7.
21668 For the luma plane, a 2x2 box radius will be run once.
21670 For the chroma plane, a 4x4 box radius will be run 5 times.
21672 For the alpha plane, a 3x3 box radius will be run 7 times.
21674 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
21678 @section colorkey_opencl
21679 RGB colorspace color keying.
21681 The filter accepts the following options:
21685 The color which will be replaced with transparency.
21688 Similarity percentage with the key color.
21690 0.01 matches only the exact key color, while 1.0 matches everything.
21695 0.0 makes pixels either fully transparent, or not transparent at all.
21697 Higher values result in semi-transparent pixels, with a higher transparency
21698 the more similar the pixels color is to the key color.
21701 @subsection Examples
21705 Make every semi-green pixel in the input transparent with some slight blending:
21707 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
21711 @section convolution_opencl
21713 Apply convolution of 3x3, 5x5, 7x7 matrix.
21715 The filter accepts the following options:
21722 Set matrix for each plane.
21723 Matrix is sequence of 9, 25 or 49 signed numbers.
21724 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
21730 Set multiplier for calculated value for each plane.
21731 If unset or 0, it will be sum of all matrix elements.
21732 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
21738 Set bias for each plane. This value is added to the result of the multiplication.
21739 Useful for making the overall image brighter or darker.
21740 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
21744 @subsection Examples
21750 -i INPUT -vf "hwupload, convolution_opencl=0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0, hwdownload" OUTPUT
21756 -i INPUT -vf "hwupload, convolution_opencl=1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1/9:1/9:1/9:1/9, hwdownload" OUTPUT
21760 Apply edge enhance:
21762 -i INPUT -vf "hwupload, convolution_opencl=0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:5:1:1:1:0:128:128:128, hwdownload" OUTPUT
21768 -i INPUT -vf "hwupload, convolution_opencl=0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:5:5:5:1:0:128:128:128, hwdownload" OUTPUT
21772 Apply laplacian edge detector which includes diagonals:
21774 -i INPUT -vf "hwupload, convolution_opencl=1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:5:5:5:1:0:128:128:0, hwdownload" OUTPUT
21780 -i INPUT -vf "hwupload, convolution_opencl=-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2, hwdownload" OUTPUT
21784 @section erosion_opencl
21786 Apply erosion effect to the video.
21788 This filter replaces the pixel by the local(3x3) minimum.
21790 It accepts the following options:
21797 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21798 If @code{0}, plane will remain unchanged.
21801 Flag which specifies the pixel to refer to.
21802 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21804 Flags to local 3x3 coordinates region centered on @code{x}:
21813 @subsection Example
21817 Apply erosion filter with threshold0 set to 30, threshold1 set 40, threshold2 set to 50 and coordinates set to 231, setting each pixel of the output to the local minimum between pixels: 1, 2, 3, 6, 7, 8 of the 3x3 region centered on it in the input. If the difference between input pixel and local minimum is more then threshold of the corresponding plane, output pixel will be set to input pixel - threshold of corresponding plane.
21819 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21823 @section deshake_opencl
21824 Feature-point based video stabilization filter.
21826 The filter accepts the following options:
21830 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
21833 Whether or not additional debug info should be displayed, both in the processed output and in the console.
21835 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
21837 Viewing point matches in the output video is only supported for RGB input.
21839 Defaults to @code{0}.
21841 @item adaptive_crop
21842 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
21844 Defaults to @code{1}.
21846 @item refine_features
21847 Whether or not feature points should be refined at a sub-pixel level.
21849 This can be turned off for a slight performance gain at the cost of precision.
21851 Defaults to @code{1}.
21853 @item smooth_strength
21854 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
21856 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
21858 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
21860 Defaults to @code{0.0}.
21862 @item smooth_window_multiplier
21863 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
21865 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
21867 Acceptable values range from @code{0.1} to @code{10.0}.
21869 Larger values increase the amount of motion data available for determining how to smooth the camera path,
21870 potentially improving smoothness, but also increase latency and memory usage.
21872 Defaults to @code{2.0}.
21876 @subsection Examples
21880 Stabilize a video with a fixed, medium smoothing strength:
21882 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
21886 Stabilize a video with debugging (both in console and in rendered video):
21888 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
21892 @section dilation_opencl
21894 Apply dilation effect to the video.
21896 This filter replaces the pixel by the local(3x3) maximum.
21898 It accepts the following options:
21905 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21906 If @code{0}, plane will remain unchanged.
21909 Flag which specifies the pixel to refer to.
21910 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21912 Flags to local 3x3 coordinates region centered on @code{x}:
21921 @subsection Example
21925 Apply dilation filter with threshold0 set to 30, threshold1 set 40, threshold2 set to 50 and coordinates set to 231, setting each pixel of the output to the local maximum between pixels: 1, 2, 3, 6, 7, 8 of the 3x3 region centered on it in the input. If the difference between input pixel and local maximum is more then threshold of the corresponding plane, output pixel will be set to input pixel + threshold of corresponding plane.
21927 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21931 @section nlmeans_opencl
21933 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
21935 @section overlay_opencl
21937 Overlay one video on top of another.
21939 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
21940 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
21942 The filter accepts the following options:
21947 Set the x coordinate of the overlaid video on the main video.
21948 Default value is @code{0}.
21951 Set the y coordinate of the overlaid video on the main video.
21952 Default value is @code{0}.
21956 @subsection Examples
21960 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
21962 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21965 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
21967 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21972 @section pad_opencl
21974 Add paddings to the input image, and place the original input at the
21975 provided @var{x}, @var{y} coordinates.
21977 It accepts the following options:
21982 Specify an expression for the size of the output image with the
21983 paddings added. If the value for @var{width} or @var{height} is 0, the
21984 corresponding input size is used for the output.
21986 The @var{width} expression can reference the value set by the
21987 @var{height} expression, and vice versa.
21989 The default value of @var{width} and @var{height} is 0.
21993 Specify the offsets to place the input image at within the padded area,
21994 with respect to the top/left border of the output image.
21996 The @var{x} expression can reference the value set by the @var{y}
21997 expression, and vice versa.
21999 The default value of @var{x} and @var{y} is 0.
22001 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22002 so the input image is centered on the padded area.
22005 Specify the color of the padded area. For the syntax of this option,
22006 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22007 manual,ffmpeg-utils}.
22010 Pad to an aspect instead to a resolution.
22013 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22014 options are expressions containing the following constants:
22019 The input video width and height.
22023 These are the same as @var{in_w} and @var{in_h}.
22027 The output width and height (the size of the padded area), as
22028 specified by the @var{width} and @var{height} expressions.
22032 These are the same as @var{out_w} and @var{out_h}.
22036 The x and y offsets as specified by the @var{x} and @var{y}
22037 expressions, or NAN if not yet specified.
22040 same as @var{iw} / @var{ih}
22043 input sample aspect ratio
22046 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22049 @section prewitt_opencl
22051 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22053 The filter accepts the following option:
22057 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22060 Set value which will be multiplied with filtered result.
22061 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22064 Set value which will be added to filtered result.
22065 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22068 @subsection Example
22072 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22074 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22078 @anchor{program_opencl}
22079 @section program_opencl
22081 Filter video using an OpenCL program.
22086 OpenCL program source file.
22089 Kernel name in program.
22092 Number of inputs to the filter. Defaults to 1.
22095 Size of output frames. Defaults to the same as the first input.
22099 The @code{program_opencl} filter also supports the @ref{framesync} options.
22101 The program source file must contain a kernel function with the given name,
22102 which will be run once for each plane of the output. Each run on a plane
22103 gets enqueued as a separate 2D global NDRange with one work-item for each
22104 pixel to be generated. The global ID offset for each work-item is therefore
22105 the coordinates of a pixel in the destination image.
22107 The kernel function needs to take the following arguments:
22110 Destination image, @var{__write_only image2d_t}.
22112 This image will become the output; the kernel should write all of it.
22114 Frame index, @var{unsigned int}.
22116 This is a counter starting from zero and increasing by one for each frame.
22118 Source images, @var{__read_only image2d_t}.
22120 These are the most recent images on each input. The kernel may read from
22121 them to generate the output, but they can't be written to.
22128 Copy the input to the output (output must be the same size as the input).
22130 __kernel void copy(__write_only image2d_t destination,
22131 unsigned int index,
22132 __read_only image2d_t source)
22134 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22136 int2 location = (int2)(get_global_id(0), get_global_id(1));
22138 float4 value = read_imagef(source, sampler, location);
22140 write_imagef(destination, location, value);
22145 Apply a simple transformation, rotating the input by an amount increasing
22146 with the index counter. Pixel values are linearly interpolated by the
22147 sampler, and the output need not have the same dimensions as the input.
22149 __kernel void rotate_image(__write_only image2d_t dst,
22150 unsigned int index,
22151 __read_only image2d_t src)
22153 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22154 CLK_FILTER_LINEAR);
22156 float angle = (float)index / 100.0f;
22158 float2 dst_dim = convert_float2(get_image_dim(dst));
22159 float2 src_dim = convert_float2(get_image_dim(src));
22161 float2 dst_cen = dst_dim / 2.0f;
22162 float2 src_cen = src_dim / 2.0f;
22164 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22166 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22168 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22169 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22171 src_pos = src_pos * src_dim / dst_dim;
22173 float2 src_loc = src_pos + src_cen;
22175 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22176 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22177 write_imagef(dst, dst_loc, 0.5f);
22179 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22184 Blend two inputs together, with the amount of each input used varying
22185 with the index counter.
22187 __kernel void blend_images(__write_only image2d_t dst,
22188 unsigned int index,
22189 __read_only image2d_t src1,
22190 __read_only image2d_t src2)
22192 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22193 CLK_FILTER_LINEAR);
22195 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22197 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22198 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22199 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22201 float4 val1 = read_imagef(src1, sampler, src1_loc);
22202 float4 val2 = read_imagef(src2, sampler, src2_loc);
22204 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22210 @section roberts_opencl
22211 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22213 The filter accepts the following option:
22217 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22220 Set value which will be multiplied with filtered result.
22221 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22224 Set value which will be added to filtered result.
22225 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22228 @subsection Example
22232 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22234 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22238 @section sobel_opencl
22240 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22242 The filter accepts the following option:
22246 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22249 Set value which will be multiplied with filtered result.
22250 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22253 Set value which will be added to filtered result.
22254 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22257 @subsection Example
22261 Apply sobel operator with scale set to 2 and delta set to 10
22263 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22267 @section tonemap_opencl
22269 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22271 It accepts the following parameters:
22275 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22278 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22281 Apply desaturation for highlights that exceed this level of brightness. The
22282 higher the parameter, the more color information will be preserved. This
22283 setting helps prevent unnaturally blown-out colors for super-highlights, by
22284 (smoothly) turning into white instead. This makes images feel more natural,
22285 at the cost of reducing information about out-of-range colors.
22287 The default value is 0.5, and the algorithm here is a little different from
22288 the cpu version tonemap currently. A setting of 0.0 disables this option.
22291 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22292 is used to detect whether the scene has changed or not. If the distance between
22293 the current frame average brightness and the current running average exceeds
22294 a threshold value, we would re-calculate scene average and peak brightness.
22295 The default value is 0.2.
22298 Specify the output pixel format.
22300 Currently supported formats are:
22307 Set the output color range.
22309 Possible values are:
22315 Default is same as input.
22318 Set the output color primaries.
22320 Possible values are:
22326 Default is same as input.
22329 Set the output transfer characteristics.
22331 Possible values are:
22340 Set the output colorspace matrix.
22342 Possible value are:
22348 Default is same as input.
22352 @subsection Example
22356 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22358 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22362 @section unsharp_opencl
22364 Sharpen or blur the input video.
22366 It accepts the following parameters:
22369 @item luma_msize_x, lx
22370 Set the luma matrix horizontal size.
22371 Range is @code{[1, 23]} and default value is @code{5}.
22373 @item luma_msize_y, ly
22374 Set the luma matrix vertical size.
22375 Range is @code{[1, 23]} and default value is @code{5}.
22377 @item luma_amount, la
22378 Set the luma effect strength.
22379 Range is @code{[-10, 10]} and default value is @code{1.0}.
22381 Negative values will blur the input video, while positive values will
22382 sharpen it, a value of zero will disable the effect.
22384 @item chroma_msize_x, cx
22385 Set the chroma matrix horizontal size.
22386 Range is @code{[1, 23]} and default value is @code{5}.
22388 @item chroma_msize_y, cy
22389 Set the chroma matrix vertical size.
22390 Range is @code{[1, 23]} and default value is @code{5}.
22392 @item chroma_amount, ca
22393 Set the chroma effect strength.
22394 Range is @code{[-10, 10]} and default value is @code{0.0}.
22396 Negative values will blur the input video, while positive values will
22397 sharpen it, a value of zero will disable the effect.
22401 All parameters are optional and default to the equivalent of the
22402 string '5:5:1.0:5:5:0.0'.
22404 @subsection Examples
22408 Apply strong luma sharpen effect:
22410 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22414 Apply a strong blur of both luma and chroma parameters:
22416 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22420 @section xfade_opencl
22422 Cross fade two videos with custom transition effect by using OpenCL.
22424 It accepts the following options:
22428 Set one of possible transition effects.
22432 Select custom transition effect, the actual transition description
22433 will be picked from source and kernel options.
22445 Default transition is fade.
22449 OpenCL program source file for custom transition.
22452 Set name of kernel to use for custom transition from program source file.
22455 Set duration of video transition.
22458 Set time of start of transition relative to first video.
22461 The program source file must contain a kernel function with the given name,
22462 which will be run once for each plane of the output. Each run on a plane
22463 gets enqueued as a separate 2D global NDRange with one work-item for each
22464 pixel to be generated. The global ID offset for each work-item is therefore
22465 the coordinates of a pixel in the destination image.
22467 The kernel function needs to take the following arguments:
22470 Destination image, @var{__write_only image2d_t}.
22472 This image will become the output; the kernel should write all of it.
22475 First Source image, @var{__read_only image2d_t}.
22476 Second Source image, @var{__read_only image2d_t}.
22478 These are the most recent images on each input. The kernel may read from
22479 them to generate the output, but they can't be written to.
22482 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22489 Apply dots curtain transition effect:
22491 __kernel void blend_images(__write_only image2d_t dst,
22492 __read_only image2d_t src1,
22493 __read_only image2d_t src2,
22496 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22497 CLK_FILTER_LINEAR);
22498 int2 p = (int2)(get_global_id(0), get_global_id(1));
22499 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22500 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22503 float2 dots = (float2)(20.0, 20.0);
22504 float2 center = (float2)(0,0);
22507 float4 val1 = read_imagef(src1, sampler, p);
22508 float4 val2 = read_imagef(src2, sampler, p);
22509 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22511 write_imagef(dst, p, next ? val1 : val2);
22517 @c man end OPENCL VIDEO FILTERS
22519 @chapter VAAPI Video Filters
22520 @c man begin VAAPI VIDEO FILTERS
22522 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22524 To enable compilation of these filters you need to configure FFmpeg with
22525 @code{--enable-vaapi}.
22527 To use vaapi filters, you need to setup the vaapi device correctly. For more information, please read @url{https://trac.ffmpeg.org/wiki/Hardware/VAAPI}
22529 @section tonemap_vaapi
22531 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22532 It maps the dynamic range of HDR10 content to the SDR content.
22533 It currently only accepts HDR10 as input.
22535 It accepts the following parameters:
22539 Specify the output pixel format.
22541 Currently supported formats are:
22550 Set the output color primaries.
22552 Default is same as input.
22555 Set the output transfer characteristics.
22560 Set the output colorspace matrix.
22562 Default is same as input.
22566 @subsection Example
22570 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22572 tonemap_vaapi=format=p010:t=bt2020-10
22576 @c man end VAAPI VIDEO FILTERS
22578 @chapter Video Sources
22579 @c man begin VIDEO SOURCES
22581 Below is a description of the currently available video sources.
22585 Buffer video frames, and make them available to the filter chain.
22587 This source is mainly intended for a programmatic use, in particular
22588 through the interface defined in @file{libavfilter/buffersrc.h}.
22590 It accepts the following parameters:
22595 Specify the size (width and height) of the buffered video frames. For the
22596 syntax of this option, check the
22597 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22600 The input video width.
22603 The input video height.
22606 A string representing the pixel format of the buffered video frames.
22607 It may be a number corresponding to a pixel format, or a pixel format
22611 Specify the timebase assumed by the timestamps of the buffered frames.
22614 Specify the frame rate expected for the video stream.
22616 @item pixel_aspect, sar
22617 The sample (pixel) aspect ratio of the input video.
22620 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
22621 to the filtergraph description to specify swscale flags for automatically
22622 inserted scalers. See @ref{Filtergraph syntax}.
22624 @item hw_frames_ctx
22625 When using a hardware pixel format, this should be a reference to an
22626 AVHWFramesContext describing input frames.
22631 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
22634 will instruct the source to accept video frames with size 320x240 and
22635 with format "yuv410p", assuming 1/24 as the timestamps timebase and
22636 square pixels (1:1 sample aspect ratio).
22637 Since the pixel format with name "yuv410p" corresponds to the number 6
22638 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
22639 this example corresponds to:
22641 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
22644 Alternatively, the options can be specified as a flat string, but this
22645 syntax is deprecated:
22647 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
22651 Create a pattern generated by an elementary cellular automaton.
22653 The initial state of the cellular automaton can be defined through the
22654 @option{filename} and @option{pattern} options. If such options are
22655 not specified an initial state is created randomly.
22657 At each new frame a new row in the video is filled with the result of
22658 the cellular automaton next generation. The behavior when the whole
22659 frame is filled is defined by the @option{scroll} option.
22661 This source accepts the following options:
22665 Read the initial cellular automaton state, i.e. the starting row, from
22666 the specified file.
22667 In the file, each non-whitespace character is considered an alive
22668 cell, a newline will terminate the row, and further characters in the
22669 file will be ignored.
22672 Read the initial cellular automaton state, i.e. the starting row, from
22673 the specified string.
22675 Each non-whitespace character in the string is considered an alive
22676 cell, a newline will terminate the row, and further characters in the
22677 string will be ignored.
22680 Set the video rate, that is the number of frames generated per second.
22683 @item random_fill_ratio, ratio
22684 Set the random fill ratio for the initial cellular automaton row. It
22685 is a floating point number value ranging from 0 to 1, defaults to
22688 This option is ignored when a file or a pattern is specified.
22690 @item random_seed, seed
22691 Set the seed for filling randomly the initial row, must be an integer
22692 included between 0 and UINT32_MAX. If not specified, or if explicitly
22693 set to -1, the filter will try to use a good random seed on a best
22697 Set the cellular automaton rule, it is a number ranging from 0 to 255.
22698 Default value is 110.
22701 Set the size of the output video. For the syntax of this option, check the
22702 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22704 If @option{filename} or @option{pattern} is specified, the size is set
22705 by default to the width of the specified initial state row, and the
22706 height is set to @var{width} * PHI.
22708 If @option{size} is set, it must contain the width of the specified
22709 pattern string, and the specified pattern will be centered in the
22712 If a filename or a pattern string is not specified, the size value
22713 defaults to "320x518" (used for a randomly generated initial state).
22716 If set to 1, scroll the output upward when all the rows in the output
22717 have been already filled. If set to 0, the new generated row will be
22718 written over the top row just after the bottom row is filled.
22721 @item start_full, full
22722 If set to 1, completely fill the output with generated rows before
22723 outputting the first frame.
22724 This is the default behavior, for disabling set the value to 0.
22727 If set to 1, stitch the left and right row edges together.
22728 This is the default behavior, for disabling set the value to 0.
22731 @subsection Examples
22735 Read the initial state from @file{pattern}, and specify an output of
22738 cellauto=f=pattern:s=200x400
22742 Generate a random initial row with a width of 200 cells, with a fill
22745 cellauto=ratio=2/3:s=200x200
22749 Create a pattern generated by rule 18 starting by a single alive cell
22750 centered on an initial row with width 100:
22752 cellauto=p=@@:s=100x400:full=0:rule=18
22756 Specify a more elaborated initial pattern:
22758 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
22763 @anchor{coreimagesrc}
22764 @section coreimagesrc
22765 Video source generated on GPU using Apple's CoreImage API on OSX.
22767 This video source is a specialized version of the @ref{coreimage} video filter.
22768 Use a core image generator at the beginning of the applied filterchain to
22769 generate the content.
22771 The coreimagesrc video source accepts the following options:
22773 @item list_generators
22774 List all available generators along with all their respective options as well as
22775 possible minimum and maximum values along with the default values.
22777 list_generators=true
22781 Specify the size of the sourced video. For the syntax of this option, check the
22782 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22783 The default value is @code{320x240}.
22786 Specify the frame rate of the sourced video, as the number of frames
22787 generated per second. It has to be a string in the format
22788 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22789 number or a valid video frame rate abbreviation. The default value is
22793 Set the sample aspect ratio of the sourced video.
22796 Set the duration of the sourced video. See
22797 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22798 for the accepted syntax.
22800 If not specified, or the expressed duration is negative, the video is
22801 supposed to be generated forever.
22804 Additionally, all options of the @ref{coreimage} video filter are accepted.
22805 A complete filterchain can be used for further processing of the
22806 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
22807 and examples for details.
22809 @subsection Examples
22814 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
22815 given as complete and escaped command-line for Apple's standard bash shell:
22817 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
22819 This example is equivalent to the QRCode example of @ref{coreimage} without the
22820 need for a nullsrc video source.
22825 Generate several gradients.
22829 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22830 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22833 Set frame rate, expressed as number of frames per second. Default
22836 @item c0, c1, c2, c3, c4, c5, c6, c7
22837 Set 8 colors. Default values for colors is to pick random one.
22839 @item x0, y0, y0, y1
22840 Set gradient line source and destination points. If negative or out of range, random ones
22844 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
22847 Set seed for picking gradient line points.
22850 Set the duration of the sourced video. See
22851 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22852 for the accepted syntax.
22854 If not specified, or the expressed duration is negative, the video is
22855 supposed to be generated forever.
22858 Set speed of gradients rotation.
22862 @section mandelbrot
22864 Generate a Mandelbrot set fractal, and progressively zoom towards the
22865 point specified with @var{start_x} and @var{start_y}.
22867 This source accepts the following options:
22872 Set the terminal pts value. Default value is 400.
22875 Set the terminal scale value.
22876 Must be a floating point value. Default value is 0.3.
22879 Set the inner coloring mode, that is the algorithm used to draw the
22880 Mandelbrot fractal internal region.
22882 It shall assume one of the following values:
22887 Show time until convergence.
22889 Set color based on point closest to the origin of the iterations.
22894 Default value is @var{mincol}.
22897 Set the bailout value. Default value is 10.0.
22900 Set the maximum of iterations performed by the rendering
22901 algorithm. Default value is 7189.
22904 Set outer coloring mode.
22905 It shall assume one of following values:
22907 @item iteration_count
22908 Set iteration count mode.
22909 @item normalized_iteration_count
22910 set normalized iteration count mode.
22912 Default value is @var{normalized_iteration_count}.
22915 Set frame rate, expressed as number of frames per second. Default
22919 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22920 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22923 Set the initial scale value. Default value is 3.0.
22926 Set the initial x position. Must be a floating point value between
22927 -100 and 100. Default value is -0.743643887037158704752191506114774.
22930 Set the initial y position. Must be a floating point value between
22931 -100 and 100. Default value is -0.131825904205311970493132056385139.
22936 Generate various test patterns, as generated by the MPlayer test filter.
22938 The size of the generated video is fixed, and is 256x256.
22939 This source is useful in particular for testing encoding features.
22941 This source accepts the following options:
22946 Specify the frame rate of the sourced video, as the number of frames
22947 generated per second. It has to be a string in the format
22948 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22949 number or a valid video frame rate abbreviation. The default value is
22953 Set the duration of the sourced video. See
22954 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22955 for the accepted syntax.
22957 If not specified, or the expressed duration is negative, the video is
22958 supposed to be generated forever.
22962 Set the number or the name of the test to perform. Supported tests are:
22976 @item max_frames, m
22977 Set the maximum number of frames generated for each test, default value is 30.
22981 Default value is "all", which will cycle through the list of all tests.
22986 mptestsrc=t=dc_luma
22989 will generate a "dc_luma" test pattern.
22991 @section frei0r_src
22993 Provide a frei0r source.
22995 To enable compilation of this filter you need to install the frei0r
22996 header and configure FFmpeg with @code{--enable-frei0r}.
22998 This source accepts the following parameters:
23003 The size of the video to generate. For the syntax of this option, check the
23004 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23007 The framerate of the generated video. It may be a string of the form
23008 @var{num}/@var{den} or a frame rate abbreviation.
23011 The name to the frei0r source to load. For more information regarding frei0r and
23012 how to set the parameters, read the @ref{frei0r} section in the video filters
23015 @item filter_params
23016 A '|'-separated list of parameters to pass to the frei0r source.
23020 For example, to generate a frei0r partik0l source with size 200x200
23021 and frame rate 10 which is overlaid on the overlay filter main input:
23023 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23028 Generate a life pattern.
23030 This source is based on a generalization of John Conway's life game.
23032 The sourced input represents a life grid, each pixel represents a cell
23033 which can be in one of two possible states, alive or dead. Every cell
23034 interacts with its eight neighbours, which are the cells that are
23035 horizontally, vertically, or diagonally adjacent.
23037 At each interaction the grid evolves according to the adopted rule,
23038 which specifies the number of neighbor alive cells which will make a
23039 cell stay alive or born. The @option{rule} option allows one to specify
23042 This source accepts the following options:
23046 Set the file from which to read the initial grid state. In the file,
23047 each non-whitespace character is considered an alive cell, and newline
23048 is used to delimit the end of each row.
23050 If this option is not specified, the initial grid is generated
23054 Set the video rate, that is the number of frames generated per second.
23057 @item random_fill_ratio, ratio
23058 Set the random fill ratio for the initial random grid. It is a
23059 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23060 It is ignored when a file is specified.
23062 @item random_seed, seed
23063 Set the seed for filling the initial random grid, must be an integer
23064 included between 0 and UINT32_MAX. If not specified, or if explicitly
23065 set to -1, the filter will try to use a good random seed on a best
23071 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23072 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23073 @var{NS} specifies the number of alive neighbor cells which make a
23074 live cell stay alive, and @var{NB} the number of alive neighbor cells
23075 which make a dead cell to become alive (i.e. to "born").
23076 "s" and "b" can be used in place of "S" and "B", respectively.
23078 Alternatively a rule can be specified by an 18-bits integer. The 9
23079 high order bits are used to encode the next cell state if it is alive
23080 for each number of neighbor alive cells, the low order bits specify
23081 the rule for "borning" new cells. Higher order bits encode for an
23082 higher number of neighbor cells.
23083 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23084 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23086 Default value is "S23/B3", which is the original Conway's game of life
23087 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23088 cells, and will born a new cell if there are three alive cells around
23092 Set the size of the output video. For the syntax of this option, check the
23093 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23095 If @option{filename} is specified, the size is set by default to the
23096 same size of the input file. If @option{size} is set, it must contain
23097 the size specified in the input file, and the initial grid defined in
23098 that file is centered in the larger resulting area.
23100 If a filename is not specified, the size value defaults to "320x240"
23101 (used for a randomly generated initial grid).
23104 If set to 1, stitch the left and right grid edges together, and the
23105 top and bottom edges also. Defaults to 1.
23108 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23109 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23110 value from 0 to 255.
23113 Set the color of living (or new born) cells.
23116 Set the color of dead cells. If @option{mold} is set, this is the first color
23117 used to represent a dead cell.
23120 Set mold color, for definitely dead and moldy cells.
23122 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23123 ffmpeg-utils manual,ffmpeg-utils}.
23126 @subsection Examples
23130 Read a grid from @file{pattern}, and center it on a grid of size
23133 life=f=pattern:s=300x300
23137 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23139 life=ratio=2/3:s=200x200
23143 Specify a custom rule for evolving a randomly generated grid:
23149 Full example with slow death effect (mold) using @command{ffplay}:
23151 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23158 @anchor{haldclutsrc}
23161 @anchor{pal100bars}
23162 @anchor{rgbtestsrc}
23164 @anchor{smptehdbars}
23167 @anchor{yuvtestsrc}
23168 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23170 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23172 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23174 The @code{color} source provides an uniformly colored input.
23176 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23177 @ref{haldclut} filter.
23179 The @code{nullsrc} source returns unprocessed video frames. It is
23180 mainly useful to be employed in analysis / debugging tools, or as the
23181 source for filters which ignore the input data.
23183 The @code{pal75bars} source generates a color bars pattern, based on
23184 EBU PAL recommendations with 75% color levels.
23186 The @code{pal100bars} source generates a color bars pattern, based on
23187 EBU PAL recommendations with 100% color levels.
23189 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23190 detecting RGB vs BGR issues. You should see a red, green and blue
23191 stripe from top to bottom.
23193 The @code{smptebars} source generates a color bars pattern, based on
23194 the SMPTE Engineering Guideline EG 1-1990.
23196 The @code{smptehdbars} source generates a color bars pattern, based on
23197 the SMPTE RP 219-2002.
23199 The @code{testsrc} source generates a test video pattern, showing a
23200 color pattern, a scrolling gradient and a timestamp. This is mainly
23201 intended for testing purposes.
23203 The @code{testsrc2} source is similar to testsrc, but supports more
23204 pixel formats instead of just @code{rgb24}. This allows using it as an
23205 input for other tests without requiring a format conversion.
23207 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23208 see a y, cb and cr stripe from top to bottom.
23210 The sources accept the following parameters:
23215 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23216 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23217 pixels to be used as identity matrix for 3D lookup tables. Each component is
23218 coded on a @code{1/(N*N)} scale.
23221 Specify the color of the source, only available in the @code{color}
23222 source. For the syntax of this option, check the
23223 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23226 Specify the size of the sourced video. For the syntax of this option, check the
23227 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23228 The default value is @code{320x240}.
23230 This option is not available with the @code{allrgb}, @code{allyuv}, and
23231 @code{haldclutsrc} filters.
23234 Specify the frame rate of the sourced video, as the number of frames
23235 generated per second. It has to be a string in the format
23236 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23237 number or a valid video frame rate abbreviation. The default value is
23241 Set the duration of the sourced video. See
23242 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23243 for the accepted syntax.
23245 If not specified, or the expressed duration is negative, the video is
23246 supposed to be generated forever.
23248 Since the frame rate is used as time base, all frames including the last one
23249 will have their full duration. If the specified duration is not a multiple
23250 of the frame duration, it will be rounded up.
23253 Set the sample aspect ratio of the sourced video.
23256 Specify the alpha (opacity) of the background, only available in the
23257 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23258 255 (fully opaque, the default).
23261 Set the number of decimals to show in the timestamp, only available in the
23262 @code{testsrc} source.
23264 The displayed timestamp value will correspond to the original
23265 timestamp value multiplied by the power of 10 of the specified
23266 value. Default value is 0.
23269 @subsection Examples
23273 Generate a video with a duration of 5.3 seconds, with size
23274 176x144 and a frame rate of 10 frames per second:
23276 testsrc=duration=5.3:size=qcif:rate=10
23280 The following graph description will generate a red source
23281 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23284 color=c=red@@0.2:s=qcif:r=10
23288 If the input content is to be ignored, @code{nullsrc} can be used. The
23289 following command generates noise in the luminance plane by employing
23290 the @code{geq} filter:
23292 nullsrc=s=256x256, geq=random(1)*255:128:128
23296 @subsection Commands
23298 The @code{color} source supports the following commands:
23302 Set the color of the created image. Accepts the same syntax of the
23303 corresponding @option{color} option.
23308 Generate video using an OpenCL program.
23313 OpenCL program source file.
23316 Kernel name in program.
23319 Size of frames to generate. This must be set.
23322 Pixel format to use for the generated frames. This must be set.
23325 Number of frames generated every second. Default value is '25'.
23329 For details of how the program loading works, see the @ref{program_opencl}
23336 Generate a colour ramp by setting pixel values from the position of the pixel
23337 in the output image. (Note that this will work with all pixel formats, but
23338 the generated output will not be the same.)
23340 __kernel void ramp(__write_only image2d_t dst,
23341 unsigned int index)
23343 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23346 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23348 write_imagef(dst, loc, val);
23353 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23355 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23356 unsigned int index)
23358 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23360 float4 value = 0.0f;
23361 int x = loc.x + index;
23362 int y = loc.y + index;
23363 while (x > 0 || y > 0) {
23364 if (x % 3 == 1 && y % 3 == 1) {
23372 write_imagef(dst, loc, value);
23378 @section sierpinski
23380 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23382 This source accepts the following options:
23386 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23387 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23390 Set frame rate, expressed as number of frames per second. Default
23394 Set seed which is used for random panning.
23397 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23400 Set fractal type, can be default @code{carpet} or @code{triangle}.
23403 @c man end VIDEO SOURCES
23405 @chapter Video Sinks
23406 @c man begin VIDEO SINKS
23408 Below is a description of the currently available video sinks.
23410 @section buffersink
23412 Buffer video frames, and make them available to the end of the filter
23415 This sink is mainly intended for programmatic use, in particular
23416 through the interface defined in @file{libavfilter/buffersink.h}
23417 or the options system.
23419 It accepts a pointer to an AVBufferSinkContext structure, which
23420 defines the incoming buffers' formats, to be passed as the opaque
23421 parameter to @code{avfilter_init_filter} for initialization.
23425 Null video sink: do absolutely nothing with the input video. It is
23426 mainly useful as a template and for use in analysis / debugging
23429 @c man end VIDEO SINKS
23431 @chapter Multimedia Filters
23432 @c man begin MULTIMEDIA FILTERS
23434 Below is a description of the currently available multimedia filters.
23438 Convert input audio to a video output, displaying the audio bit scope.
23440 The filter accepts the following options:
23444 Set frame rate, expressed as number of frames per second. Default
23448 Specify the video size for the output. For the syntax of this option, check the
23449 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23450 Default value is @code{1024x256}.
23453 Specify list of colors separated by space or by '|' which will be used to
23454 draw channels. Unrecognized or missing colors will be replaced
23458 @section adrawgraph
23459 Draw a graph using input audio metadata.
23461 See @ref{drawgraph}
23463 @section agraphmonitor
23465 See @ref{graphmonitor}.
23467 @section ahistogram
23469 Convert input audio to a video output, displaying the volume histogram.
23471 The filter accepts the following options:
23475 Specify how histogram is calculated.
23477 It accepts the following values:
23480 Use single histogram for all channels.
23482 Use separate histogram for each channel.
23484 Default is @code{single}.
23487 Set frame rate, expressed as number of frames per second. Default
23491 Specify the video size for the output. For the syntax of this option, check the
23492 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23493 Default value is @code{hd720}.
23498 It accepts the following values:
23509 reverse logarithmic
23511 Default is @code{log}.
23514 Set amplitude scale.
23516 It accepts the following values:
23523 Default is @code{log}.
23526 Set how much frames to accumulate in histogram.
23527 Default is 1. Setting this to -1 accumulates all frames.
23530 Set histogram ratio of window height.
23533 Set sonogram sliding.
23535 It accepts the following values:
23538 replace old rows with new ones.
23540 scroll from top to bottom.
23542 Default is @code{replace}.
23545 @section aphasemeter
23547 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23548 representing mean phase of current audio frame. A video output can also be produced and is
23549 enabled by default. The audio is passed through as first output.
23551 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23552 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23553 and @code{1} means channels are in phase.
23555 The filter accepts the following options, all related to its video output:
23559 Set the output frame rate. Default value is @code{25}.
23562 Set the video size for the output. For the syntax of this option, check the
23563 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23564 Default value is @code{800x400}.
23569 Specify the red, green, blue contrast. Default values are @code{2},
23570 @code{7} and @code{1}.
23571 Allowed range is @code{[0, 255]}.
23574 Set color which will be used for drawing median phase. If color is
23575 @code{none} which is default, no median phase value will be drawn.
23578 Enable video output. Default is enabled.
23581 @subsection phasing detection
23583 The filter also detects out of phase and mono sequences in stereo streams.
23584 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
23586 The filter accepts the following options for this detection:
23590 Enable mono and out of phase detection. Default is disabled.
23593 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
23594 Allowed range is @code{[0, 1]}.
23597 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
23598 Allowed range is @code{[90, 180]}.
23601 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
23604 @subsection Examples
23608 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
23610 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
23614 @section avectorscope
23616 Convert input audio to a video output, representing the audio vector
23619 The filter is used to measure the difference between channels of stereo
23620 audio stream. A monaural signal, consisting of identical left and right
23621 signal, results in straight vertical line. Any stereo separation is visible
23622 as a deviation from this line, creating a Lissajous figure.
23623 If the straight (or deviation from it) but horizontal line appears this
23624 indicates that the left and right channels are out of phase.
23626 The filter accepts the following options:
23630 Set the vectorscope mode.
23632 Available values are:
23635 Lissajous rotated by 45 degrees.
23638 Same as above but not rotated.
23641 Shape resembling half of circle.
23644 Default value is @samp{lissajous}.
23647 Set the video size for the output. For the syntax of this option, check the
23648 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23649 Default value is @code{400x400}.
23652 Set the output frame rate. Default value is @code{25}.
23658 Specify the red, green, blue and alpha contrast. Default values are @code{40},
23659 @code{160}, @code{80} and @code{255}.
23660 Allowed range is @code{[0, 255]}.
23666 Specify the red, green, blue and alpha fade. Default values are @code{15},
23667 @code{10}, @code{5} and @code{5}.
23668 Allowed range is @code{[0, 255]}.
23671 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
23672 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
23675 Set the vectorscope drawing mode.
23677 Available values are:
23680 Draw dot for each sample.
23683 Draw line between previous and current sample.
23686 Default value is @samp{dot}.
23689 Specify amplitude scale of audio samples.
23691 Available values are:
23707 Swap left channel axis with right channel axis.
23717 Mirror only x axis.
23720 Mirror only y axis.
23728 @subsection Examples
23732 Complete example using @command{ffplay}:
23734 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
23735 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
23739 @section bench, abench
23741 Benchmark part of a filtergraph.
23743 The filter accepts the following options:
23747 Start or stop a timer.
23749 Available values are:
23752 Get the current time, set it as frame metadata (using the key
23753 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
23756 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
23757 the input frame metadata to get the time difference. Time difference, average,
23758 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
23759 @code{min}) are then printed. The timestamps are expressed in seconds.
23763 @subsection Examples
23767 Benchmark @ref{selectivecolor} filter:
23769 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
23775 Concatenate audio and video streams, joining them together one after the
23778 The filter works on segments of synchronized video and audio streams. All
23779 segments must have the same number of streams of each type, and that will
23780 also be the number of streams at output.
23782 The filter accepts the following options:
23787 Set the number of segments. Default is 2.
23790 Set the number of output video streams, that is also the number of video
23791 streams in each segment. Default is 1.
23794 Set the number of output audio streams, that is also the number of audio
23795 streams in each segment. Default is 0.
23798 Activate unsafe mode: do not fail if segments have a different format.
23802 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
23803 @var{a} audio outputs.
23805 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
23806 segment, in the same order as the outputs, then the inputs for the second
23809 Related streams do not always have exactly the same duration, for various
23810 reasons including codec frame size or sloppy authoring. For that reason,
23811 related synchronized streams (e.g. a video and its audio track) should be
23812 concatenated at once. The concat filter will use the duration of the longest
23813 stream in each segment (except the last one), and if necessary pad shorter
23814 audio streams with silence.
23816 For this filter to work correctly, all segments must start at timestamp 0.
23818 All corresponding streams must have the same parameters in all segments; the
23819 filtering system will automatically select a common pixel format for video
23820 streams, and a common sample format, sample rate and channel layout for
23821 audio streams, but other settings, such as resolution, must be converted
23822 explicitly by the user.
23824 Different frame rates are acceptable but will result in variable frame rate
23825 at output; be sure to configure the output file to handle it.
23827 @subsection Examples
23831 Concatenate an opening, an episode and an ending, all in bilingual version
23832 (video in stream 0, audio in streams 1 and 2):
23834 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
23835 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
23836 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
23837 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
23841 Concatenate two parts, handling audio and video separately, using the
23842 (a)movie sources, and adjusting the resolution:
23844 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
23845 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
23846 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
23848 Note that a desync will happen at the stitch if the audio and video streams
23849 do not have exactly the same duration in the first file.
23853 @subsection Commands
23855 This filter supports the following commands:
23858 Close the current segment and step to the next one
23864 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
23865 level. By default, it logs a message at a frequency of 10Hz with the
23866 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
23867 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
23869 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
23870 sample format is double-precision floating point. The input stream will be converted to
23871 this specification, if needed. Users may need to insert aformat and/or aresample filters
23872 after this filter to obtain the original parameters.
23874 The filter also has a video output (see the @var{video} option) with a real
23875 time graph to observe the loudness evolution. The graphic contains the logged
23876 message mentioned above, so it is not printed anymore when this option is set,
23877 unless the verbose logging is set. The main graphing area contains the
23878 short-term loudness (3 seconds of analysis), and the gauge on the right is for
23879 the momentary loudness (400 milliseconds), but can optionally be configured
23880 to instead display short-term loudness (see @var{gauge}).
23882 The green area marks a +/- 1LU target range around the target loudness
23883 (-23LUFS by default, unless modified through @var{target}).
23885 More information about the Loudness Recommendation EBU R128 on
23886 @url{http://tech.ebu.ch/loudness}.
23888 The filter accepts the following options:
23893 Activate the video output. The audio stream is passed unchanged whether this
23894 option is set or no. The video stream will be the first output stream if
23895 activated. Default is @code{0}.
23898 Set the video size. This option is for video only. For the syntax of this
23900 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23901 Default and minimum resolution is @code{640x480}.
23904 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
23905 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
23906 other integer value between this range is allowed.
23909 Set metadata injection. If set to @code{1}, the audio input will be segmented
23910 into 100ms output frames, each of them containing various loudness information
23911 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
23913 Default is @code{0}.
23916 Force the frame logging level.
23918 Available values are:
23921 information logging level
23923 verbose logging level
23926 By default, the logging level is set to @var{info}. If the @option{video} or
23927 the @option{metadata} options are set, it switches to @var{verbose}.
23932 Available modes can be cumulated (the option is a @code{flag} type). Possible
23936 Disable any peak mode (default).
23938 Enable sample-peak mode.
23940 Simple peak mode looking for the higher sample value. It logs a message
23941 for sample-peak (identified by @code{SPK}).
23943 Enable true-peak mode.
23945 If enabled, the peak lookup is done on an over-sampled version of the input
23946 stream for better peak accuracy. It logs a message for true-peak.
23947 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
23948 This mode requires a build with @code{libswresample}.
23952 Treat mono input files as "dual mono". If a mono file is intended for playback
23953 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
23954 If set to @code{true}, this option will compensate for this effect.
23955 Multi-channel input files are not affected by this option.
23958 Set a specific pan law to be used for the measurement of dual mono files.
23959 This parameter is optional, and has a default value of -3.01dB.
23962 Set a specific target level (in LUFS) used as relative zero in the visualization.
23963 This parameter is optional and has a default value of -23LUFS as specified
23964 by EBU R128. However, material published online may prefer a level of -16LUFS
23965 (e.g. for use with podcasts or video platforms).
23968 Set the value displayed by the gauge. Valid values are @code{momentary} and s
23969 @code{shortterm}. By default the momentary value will be used, but in certain
23970 scenarios it may be more useful to observe the short term value instead (e.g.
23974 Sets the display scale for the loudness. Valid parameters are @code{absolute}
23975 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
23976 video output, not the summary or continuous log output.
23979 @subsection Examples
23983 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
23985 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
23989 Run an analysis with @command{ffmpeg}:
23991 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
23995 @section interleave, ainterleave
23997 Temporally interleave frames from several inputs.
23999 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24001 These filters read frames from several inputs and send the oldest
24002 queued frame to the output.
24004 Input streams must have well defined, monotonically increasing frame
24007 In order to submit one frame to output, these filters need to enqueue
24008 at least one frame for each input, so they cannot work in case one
24009 input is not yet terminated and will not receive incoming frames.
24011 For example consider the case when one input is a @code{select} filter
24012 which always drops input frames. The @code{interleave} filter will keep
24013 reading from that input, but it will never be able to send new frames
24014 to output until the input sends an end-of-stream signal.
24016 Also, depending on inputs synchronization, the filters will drop
24017 frames in case one input receives more frames than the other ones, and
24018 the queue is already filled.
24020 These filters accept the following options:
24024 Set the number of different inputs, it is 2 by default.
24027 How to determine the end-of-stream.
24031 The duration of the longest input. (default)
24034 The duration of the shortest input.
24037 The duration of the first input.
24042 @subsection Examples
24046 Interleave frames belonging to different streams using @command{ffmpeg}:
24048 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24052 Add flickering blur effect:
24054 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24058 @section metadata, ametadata
24060 Manipulate frame metadata.
24062 This filter accepts the following options:
24066 Set mode of operation of the filter.
24068 Can be one of the following:
24072 If both @code{value} and @code{key} is set, select frames
24073 which have such metadata. If only @code{key} is set, select
24074 every frame that has such key in metadata.
24077 Add new metadata @code{key} and @code{value}. If key is already available
24081 Modify value of already present key.
24084 If @code{value} is set, delete only keys that have such value.
24085 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24089 Print key and its value if metadata was found. If @code{key} is not set print all
24090 metadata values available in frame.
24094 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24097 Set metadata value which will be used. This option is mandatory for
24098 @code{modify} and @code{add} mode.
24101 Which function to use when comparing metadata value and @code{value}.
24103 Can be one of following:
24107 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24110 Values are interpreted as strings, returns true if metadata value starts with
24111 the @code{value} option string.
24114 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24117 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24120 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24123 Values are interpreted as floats, returns true if expression from option @code{expr}
24127 Values are interpreted as strings, returns true if metadata value ends with
24128 the @code{value} option string.
24132 Set expression which is used when @code{function} is set to @code{expr}.
24133 The expression is evaluated through the eval API and can contain the following
24138 Float representation of @code{value} from metadata key.
24141 Float representation of @code{value} as supplied by user in @code{value} option.
24145 If specified in @code{print} mode, output is written to the named file. Instead of
24146 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24147 for standard output. If @code{file} option is not set, output is written to the log
24148 with AV_LOG_INFO loglevel.
24151 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24155 @subsection Examples
24159 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24162 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24165 Print silencedetect output to file @file{metadata.txt}.
24167 silencedetect,ametadata=mode=print:file=metadata.txt
24170 Direct all metadata to a pipe with file descriptor 4.
24172 metadata=mode=print:file='pipe\:4'
24176 @section perms, aperms
24178 Set read/write permissions for the output frames.
24180 These filters are mainly aimed at developers to test direct path in the
24181 following filter in the filtergraph.
24183 The filters accept the following options:
24187 Select the permissions mode.
24189 It accepts the following values:
24192 Do nothing. This is the default.
24194 Set all the output frames read-only.
24196 Set all the output frames directly writable.
24198 Make the frame read-only if writable, and writable if read-only.
24200 Set each output frame read-only or writable randomly.
24204 Set the seed for the @var{random} mode, must be an integer included between
24205 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24206 @code{-1}, the filter will try to use a good random seed on a best effort
24210 Note: in case of auto-inserted filter between the permission filter and the
24211 following one, the permission might not be received as expected in that
24212 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24213 perms/aperms filter can avoid this problem.
24215 @section realtime, arealtime
24217 Slow down filtering to match real time approximately.
24219 These filters will pause the filtering for a variable amount of time to
24220 match the output rate with the input timestamps.
24221 They are similar to the @option{re} option to @code{ffmpeg}.
24223 They accept the following options:
24227 Time limit for the pauses. Any pause longer than that will be considered
24228 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24230 Speed factor for processing. The value must be a float larger than zero.
24231 Values larger than 1.0 will result in faster than realtime processing,
24232 smaller will slow processing down. The @var{limit} is automatically adapted
24233 accordingly. Default is 1.0.
24235 A processing speed faster than what is possible without these filters cannot
24240 @section select, aselect
24242 Select frames to pass in output.
24244 This filter accepts the following options:
24249 Set expression, which is evaluated for each input frame.
24251 If the expression is evaluated to zero, the frame is discarded.
24253 If the evaluation result is negative or NaN, the frame is sent to the
24254 first output; otherwise it is sent to the output with index
24255 @code{ceil(val)-1}, assuming that the input index starts from 0.
24257 For example a value of @code{1.2} corresponds to the output with index
24258 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24261 Set the number of outputs. The output to which to send the selected
24262 frame is based on the result of the evaluation. Default value is 1.
24265 The expression can contain the following constants:
24269 The (sequential) number of the filtered frame, starting from 0.
24272 The (sequential) number of the selected frame, starting from 0.
24274 @item prev_selected_n
24275 The sequential number of the last selected frame. It's NAN if undefined.
24278 The timebase of the input timestamps.
24281 The PTS (Presentation TimeStamp) of the filtered video frame,
24282 expressed in @var{TB} units. It's NAN if undefined.
24285 The PTS of the filtered video frame,
24286 expressed in seconds. It's NAN if undefined.
24289 The PTS of the previously filtered video frame. It's NAN if undefined.
24291 @item prev_selected_pts
24292 The PTS of the last previously filtered video frame. It's NAN if undefined.
24294 @item prev_selected_t
24295 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24298 The PTS of the first video frame in the video. It's NAN if undefined.
24301 The time of the first video frame in the video. It's NAN if undefined.
24303 @item pict_type @emph{(video only)}
24304 The type of the filtered frame. It can assume one of the following
24316 @item interlace_type @emph{(video only)}
24317 The frame interlace type. It can assume one of the following values:
24320 The frame is progressive (not interlaced).
24322 The frame is top-field-first.
24324 The frame is bottom-field-first.
24327 @item consumed_sample_n @emph{(audio only)}
24328 the number of selected samples before the current frame
24330 @item samples_n @emph{(audio only)}
24331 the number of samples in the current frame
24333 @item sample_rate @emph{(audio only)}
24334 the input sample rate
24337 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24340 the position in the file of the filtered frame, -1 if the information
24341 is not available (e.g. for synthetic video)
24343 @item scene @emph{(video only)}
24344 value between 0 and 1 to indicate a new scene; a low value reflects a low
24345 probability for the current frame to introduce a new scene, while a higher
24346 value means the current frame is more likely to be one (see the example below)
24348 @item concatdec_select
24349 The concat demuxer can select only part of a concat input file by setting an
24350 inpoint and an outpoint, but the output packets may not be entirely contained
24351 in the selected interval. By using this variable, it is possible to skip frames
24352 generated by the concat demuxer which are not exactly contained in the selected
24355 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24356 and the @var{lavf.concat.duration} packet metadata values which are also
24357 present in the decoded frames.
24359 The @var{concatdec_select} variable is -1 if the frame pts is at least
24360 start_time and either the duration metadata is missing or the frame pts is less
24361 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24364 That basically means that an input frame is selected if its pts is within the
24365 interval set by the concat demuxer.
24369 The default value of the select expression is "1".
24371 @subsection Examples
24375 Select all frames in input:
24380 The example above is the same as:
24392 Select only I-frames:
24394 select='eq(pict_type\,I)'
24398 Select one frame every 100:
24400 select='not(mod(n\,100))'
24404 Select only frames contained in the 10-20 time interval:
24406 select=between(t\,10\,20)
24410 Select only I-frames contained in the 10-20 time interval:
24412 select=between(t\,10\,20)*eq(pict_type\,I)
24416 Select frames with a minimum distance of 10 seconds:
24418 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24422 Use aselect to select only audio frames with samples number > 100:
24424 aselect='gt(samples_n\,100)'
24428 Create a mosaic of the first scenes:
24430 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24433 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24437 Send even and odd frames to separate outputs, and compose them:
24439 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24443 Select useful frames from an ffconcat file which is using inpoints and
24444 outpoints but where the source files are not intra frame only.
24446 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24450 @section sendcmd, asendcmd
24452 Send commands to filters in the filtergraph.
24454 These filters read commands to be sent to other filters in the
24457 @code{sendcmd} must be inserted between two video filters,
24458 @code{asendcmd} must be inserted between two audio filters, but apart
24459 from that they act the same way.
24461 The specification of commands can be provided in the filter arguments
24462 with the @var{commands} option, or in a file specified by the
24463 @var{filename} option.
24465 These filters accept the following options:
24468 Set the commands to be read and sent to the other filters.
24470 Set the filename of the commands to be read and sent to the other
24474 @subsection Commands syntax
24476 A commands description consists of a sequence of interval
24477 specifications, comprising a list of commands to be executed when a
24478 particular event related to that interval occurs. The occurring event
24479 is typically the current frame time entering or leaving a given time
24482 An interval is specified by the following syntax:
24484 @var{START}[-@var{END}] @var{COMMANDS};
24487 The time interval is specified by the @var{START} and @var{END} times.
24488 @var{END} is optional and defaults to the maximum time.
24490 The current frame time is considered within the specified interval if
24491 it is included in the interval [@var{START}, @var{END}), that is when
24492 the time is greater or equal to @var{START} and is lesser than
24495 @var{COMMANDS} consists of a sequence of one or more command
24496 specifications, separated by ",", relating to that interval. The
24497 syntax of a command specification is given by:
24499 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24502 @var{FLAGS} is optional and specifies the type of events relating to
24503 the time interval which enable sending the specified command, and must
24504 be a non-null sequence of identifier flags separated by "+" or "|" and
24505 enclosed between "[" and "]".
24507 The following flags are recognized:
24510 The command is sent when the current frame timestamp enters the
24511 specified interval. In other words, the command is sent when the
24512 previous frame timestamp was not in the given interval, and the
24516 The command is sent when the current frame timestamp leaves the
24517 specified interval. In other words, the command is sent when the
24518 previous frame timestamp was in the given interval, and the
24522 The command @var{ARG} is interpreted as expression and result of
24523 expression is passed as @var{ARG}.
24525 The expression is evaluated through the eval API and can contain the following
24530 Original position in the file of the frame, or undefined if undefined
24531 for the current frame.
24534 The presentation timestamp in input.
24537 The count of the input frame for video or audio, starting from 0.
24540 The time in seconds of the current frame.
24543 The start time in seconds of the current command interval.
24546 The end time in seconds of the current command interval.
24549 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24554 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24557 @var{TARGET} specifies the target of the command, usually the name of
24558 the filter class or a specific filter instance name.
24560 @var{COMMAND} specifies the name of the command for the target filter.
24562 @var{ARG} is optional and specifies the optional list of argument for
24563 the given @var{COMMAND}.
24565 Between one interval specification and another, whitespaces, or
24566 sequences of characters starting with @code{#} until the end of line,
24567 are ignored and can be used to annotate comments.
24569 A simplified BNF description of the commands specification syntax
24572 @var{COMMAND_FLAG} ::= "enter" | "leave"
24573 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24574 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24575 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24576 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24577 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24580 @subsection Examples
24584 Specify audio tempo change at second 4:
24586 asendcmd=c='4.0 atempo tempo 1.5',atempo
24590 Target a specific filter instance:
24592 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
24596 Specify a list of drawtext and hue commands in a file.
24598 # show text in the interval 5-10
24599 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
24600 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
24602 # desaturate the image in the interval 15-20
24603 15.0-20.0 [enter] hue s 0,
24604 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
24606 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
24608 # apply an exponential saturation fade-out effect, starting from time 25
24609 25 [enter] hue s exp(25-t)
24612 A filtergraph allowing to read and process the above command list
24613 stored in a file @file{test.cmd}, can be specified with:
24615 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
24620 @section setpts, asetpts
24622 Change the PTS (presentation timestamp) of the input frames.
24624 @code{setpts} works on video frames, @code{asetpts} on audio frames.
24626 This filter accepts the following options:
24631 The expression which is evaluated for each frame to construct its timestamp.
24635 The expression is evaluated through the eval API and can contain the following
24639 @item FRAME_RATE, FR
24640 frame rate, only defined for constant frame-rate video
24643 The presentation timestamp in input
24646 The count of the input frame for video or the number of consumed samples,
24647 not including the current frame for audio, starting from 0.
24649 @item NB_CONSUMED_SAMPLES
24650 The number of consumed samples, not including the current frame (only
24653 @item NB_SAMPLES, S
24654 The number of samples in the current frame (only audio)
24656 @item SAMPLE_RATE, SR
24657 The audio sample rate.
24660 The PTS of the first frame.
24663 the time in seconds of the first frame
24666 State whether the current frame is interlaced.
24669 the time in seconds of the current frame
24672 original position in the file of the frame, or undefined if undefined
24673 for the current frame
24676 The previous input PTS.
24679 previous input time in seconds
24682 The previous output PTS.
24685 previous output time in seconds
24688 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
24692 The wallclock (RTC) time at the start of the movie in microseconds.
24695 The timebase of the input timestamps.
24699 @subsection Examples
24703 Start counting PTS from zero
24705 setpts=PTS-STARTPTS
24709 Apply fast motion effect:
24715 Apply slow motion effect:
24721 Set fixed rate of 25 frames per second:
24727 Set fixed rate 25 fps with some jitter:
24729 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
24733 Apply an offset of 10 seconds to the input PTS:
24739 Generate timestamps from a "live source" and rebase onto the current timebase:
24741 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
24745 Generate timestamps by counting samples:
24754 Force color range for the output video frame.
24756 The @code{setrange} filter marks the color range property for the
24757 output frames. It does not change the input frame, but only sets the
24758 corresponding property, which affects how the frame is treated by
24761 The filter accepts the following options:
24766 Available values are:
24770 Keep the same color range property.
24772 @item unspecified, unknown
24773 Set the color range as unspecified.
24775 @item limited, tv, mpeg
24776 Set the color range as limited.
24778 @item full, pc, jpeg
24779 Set the color range as full.
24783 @section settb, asettb
24785 Set the timebase to use for the output frames timestamps.
24786 It is mainly useful for testing timebase configuration.
24788 It accepts the following parameters:
24793 The expression which is evaluated into the output timebase.
24797 The value for @option{tb} is an arithmetic expression representing a
24798 rational. The expression can contain the constants "AVTB" (the default
24799 timebase), "intb" (the input timebase) and "sr" (the sample rate,
24800 audio only). Default value is "intb".
24802 @subsection Examples
24806 Set the timebase to 1/25:
24812 Set the timebase to 1/10:
24818 Set the timebase to 1001/1000:
24824 Set the timebase to 2*intb:
24830 Set the default timebase value:
24837 Convert input audio to a video output representing frequency spectrum
24838 logarithmically using Brown-Puckette constant Q transform algorithm with
24839 direct frequency domain coefficient calculation (but the transform itself
24840 is not really constant Q, instead the Q factor is actually variable/clamped),
24841 with musical tone scale, from E0 to D#10.
24843 The filter accepts the following options:
24847 Specify the video size for the output. It must be even. For the syntax of this option,
24848 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24849 Default value is @code{1920x1080}.
24852 Set the output frame rate. Default value is @code{25}.
24855 Set the bargraph height. It must be even. Default value is @code{-1} which
24856 computes the bargraph height automatically.
24859 Set the axis height. It must be even. Default value is @code{-1} which computes
24860 the axis height automatically.
24863 Set the sonogram height. It must be even. Default value is @code{-1} which
24864 computes the sonogram height automatically.
24867 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
24868 instead. Default value is @code{1}.
24870 @item sono_v, volume
24871 Specify the sonogram volume expression. It can contain variables:
24874 the @var{bar_v} evaluated expression
24875 @item frequency, freq, f
24876 the frequency where it is evaluated
24877 @item timeclamp, tc
24878 the value of @var{timeclamp} option
24882 @item a_weighting(f)
24883 A-weighting of equal loudness
24884 @item b_weighting(f)
24885 B-weighting of equal loudness
24886 @item c_weighting(f)
24887 C-weighting of equal loudness.
24889 Default value is @code{16}.
24891 @item bar_v, volume2
24892 Specify the bargraph volume expression. It can contain variables:
24895 the @var{sono_v} evaluated expression
24896 @item frequency, freq, f
24897 the frequency where it is evaluated
24898 @item timeclamp, tc
24899 the value of @var{timeclamp} option
24903 @item a_weighting(f)
24904 A-weighting of equal loudness
24905 @item b_weighting(f)
24906 B-weighting of equal loudness
24907 @item c_weighting(f)
24908 C-weighting of equal loudness.
24910 Default value is @code{sono_v}.
24912 @item sono_g, gamma
24913 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
24914 higher gamma makes the spectrum having more range. Default value is @code{3}.
24915 Acceptable range is @code{[1, 7]}.
24917 @item bar_g, gamma2
24918 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
24922 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
24923 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
24925 @item timeclamp, tc
24926 Specify the transform timeclamp. At low frequency, there is trade-off between
24927 accuracy in time domain and frequency domain. If timeclamp is lower,
24928 event in time domain is represented more accurately (such as fast bass drum),
24929 otherwise event in frequency domain is represented more accurately
24930 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
24933 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
24934 limits future samples by applying asymmetric windowing in time domain, useful
24935 when low latency is required. Accepted range is @code{[0, 1]}.
24938 Specify the transform base frequency. Default value is @code{20.01523126408007475},
24939 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
24942 Specify the transform end frequency. Default value is @code{20495.59681441799654},
24943 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
24946 This option is deprecated and ignored.
24949 Specify the transform length in time domain. Use this option to control accuracy
24950 trade-off between time domain and frequency domain at every frequency sample.
24951 It can contain variables:
24953 @item frequency, freq, f
24954 the frequency where it is evaluated
24955 @item timeclamp, tc
24956 the value of @var{timeclamp} option.
24958 Default value is @code{384*tc/(384+tc*f)}.
24961 Specify the transform count for every video frame. Default value is @code{6}.
24962 Acceptable range is @code{[1, 30]}.
24965 Specify the transform count for every single pixel. Default value is @code{0},
24966 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
24969 Specify font file for use with freetype to draw the axis. If not specified,
24970 use embedded font. Note that drawing with font file or embedded font is not
24971 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
24975 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
24976 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
24980 Specify font color expression. This is arithmetic expression that should return
24981 integer value 0xRRGGBB. It can contain variables:
24983 @item frequency, freq, f
24984 the frequency where it is evaluated
24985 @item timeclamp, tc
24986 the value of @var{timeclamp} option
24991 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
24992 @item r(x), g(x), b(x)
24993 red, green, and blue value of intensity x.
24995 Default value is @code{st(0, (midi(f)-59.5)/12);
24996 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
24997 r(1-ld(1)) + b(ld(1))}.
25000 Specify image file to draw the axis. This option override @var{fontfile} and
25001 @var{fontcolor} option.
25004 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25005 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25006 Default value is @code{1}.
25009 Set colorspace. The accepted values are:
25012 Unspecified (default)
25021 BT.470BG or BT.601-6 625
25024 SMPTE-170M or BT.601-6 525
25030 BT.2020 with non-constant luminance
25035 Set spectrogram color scheme. This is list of floating point values with format
25036 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25037 The default is @code{1|0.5|0|0|0.5|1}.
25041 @subsection Examples
25045 Playing audio while showing the spectrum:
25047 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25051 Same as above, but with frame rate 30 fps:
25053 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25057 Playing at 1280x720:
25059 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25063 Disable sonogram display:
25069 A1 and its harmonics: A1, A2, (near)E3, A3:
25071 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),
25072 asplit[a][out1]; [a] showcqt [out0]'
25076 Same as above, but with more accuracy in frequency domain:
25078 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),
25079 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25085 bar_v=10:sono_v=bar_v*a_weighting(f)
25089 Custom gamma, now spectrum is linear to the amplitude.
25095 Custom tlength equation:
25097 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)))'
25101 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25103 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25107 Custom font using fontconfig:
25109 font='Courier New,Monospace,mono|bold'
25113 Custom frequency range with custom axis using image file:
25115 axisfile=myaxis.png:basefreq=40:endfreq=10000
25121 Convert input audio to video output representing the audio power spectrum.
25122 Audio amplitude is on Y-axis while frequency is on X-axis.
25124 The filter accepts the following options:
25128 Specify size of video. For the syntax of this option, check the
25129 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25130 Default is @code{1024x512}.
25134 This set how each frequency bin will be represented.
25136 It accepts the following values:
25142 Default is @code{bar}.
25145 Set amplitude scale.
25147 It accepts the following values:
25161 Default is @code{log}.
25164 Set frequency scale.
25166 It accepts the following values:
25175 Reverse logarithmic scale.
25177 Default is @code{lin}.
25180 Set window size. Allowed range is from 16 to 65536.
25182 Default is @code{2048}
25185 Set windowing function.
25187 It accepts the following values:
25210 Default is @code{hanning}.
25213 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25214 which means optimal overlap for selected window function will be picked.
25217 Set time averaging. Setting this to 0 will display current maximal peaks.
25218 Default is @code{1}, which means time averaging is disabled.
25221 Specify list of colors separated by space or by '|' which will be used to
25222 draw channel frequencies. Unrecognized or missing colors will be replaced
25226 Set channel display mode.
25228 It accepts the following values:
25233 Default is @code{combined}.
25236 Set minimum amplitude used in @code{log} amplitude scaler.
25239 Set data display mode.
25241 It accepts the following values:
25247 Default is @code{magnitude}.
25250 @section showspatial
25252 Convert stereo input audio to a video output, representing the spatial relationship
25253 between two channels.
25255 The filter accepts the following options:
25259 Specify the video size for the output. For the syntax of this option, check the
25260 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25261 Default value is @code{512x512}.
25264 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25267 Set window function.
25269 It accepts the following values:
25294 Default value is @code{hann}.
25297 Set ratio of overlap window. Default value is @code{0.5}.
25298 When value is @code{1} overlap is set to recommended size for specific
25299 window function currently used.
25302 @anchor{showspectrum}
25303 @section showspectrum
25305 Convert input audio to a video output, representing the audio frequency
25308 The filter accepts the following options:
25312 Specify the video size for the output. For the syntax of this option, check the
25313 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25314 Default value is @code{640x512}.
25317 Specify how the spectrum should slide along the window.
25319 It accepts the following values:
25322 the samples start again on the left when they reach the right
25324 the samples scroll from right to left
25326 frames are only produced when the samples reach the right
25328 the samples scroll from left to right
25331 Default value is @code{replace}.
25334 Specify display mode.
25336 It accepts the following values:
25339 all channels are displayed in the same row
25341 all channels are displayed in separate rows
25344 Default value is @samp{combined}.
25347 Specify display color mode.
25349 It accepts the following values:
25352 each channel is displayed in a separate color
25354 each channel is displayed using the same color scheme
25356 each channel is displayed using the rainbow color scheme
25358 each channel is displayed using the moreland color scheme
25360 each channel is displayed using the nebulae color scheme
25362 each channel is displayed using the fire color scheme
25364 each channel is displayed using the fiery color scheme
25366 each channel is displayed using the fruit color scheme
25368 each channel is displayed using the cool color scheme
25370 each channel is displayed using the magma color scheme
25372 each channel is displayed using the green color scheme
25374 each channel is displayed using the viridis color scheme
25376 each channel is displayed using the plasma color scheme
25378 each channel is displayed using the cividis color scheme
25380 each channel is displayed using the terrain color scheme
25383 Default value is @samp{channel}.
25386 Specify scale used for calculating intensity color values.
25388 It accepts the following values:
25393 square root, default
25404 Default value is @samp{sqrt}.
25407 Specify frequency scale.
25409 It accepts the following values:
25417 Default value is @samp{lin}.
25420 Set saturation modifier for displayed colors. Negative values provide
25421 alternative color scheme. @code{0} is no saturation at all.
25422 Saturation must be in [-10.0, 10.0] range.
25423 Default value is @code{1}.
25426 Set window function.
25428 It accepts the following values:
25453 Default value is @code{hann}.
25456 Set orientation of time vs frequency axis. Can be @code{vertical} or
25457 @code{horizontal}. Default is @code{vertical}.
25460 Set ratio of overlap window. Default value is @code{0}.
25461 When value is @code{1} overlap is set to recommended size for specific
25462 window function currently used.
25465 Set scale gain for calculating intensity color values.
25466 Default value is @code{1}.
25469 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25472 Set color rotation, must be in [-1.0, 1.0] range.
25473 Default value is @code{0}.
25476 Set start frequency from which to display spectrogram. Default is @code{0}.
25479 Set stop frequency to which to display spectrogram. Default is @code{0}.
25482 Set upper frame rate limit. Default is @code{auto}, unlimited.
25485 Draw time and frequency axes and legends. Default is disabled.
25488 The usage is very similar to the showwaves filter; see the examples in that
25491 @subsection Examples
25495 Large window with logarithmic color scaling:
25497 showspectrum=s=1280x480:scale=log
25501 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25503 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25504 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25508 @section showspectrumpic
25510 Convert input audio to a single video frame, representing the audio frequency
25513 The filter accepts the following options:
25517 Specify the video size for the output. For the syntax of this option, check the
25518 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25519 Default value is @code{4096x2048}.
25522 Specify display mode.
25524 It accepts the following values:
25527 all channels are displayed in the same row
25529 all channels are displayed in separate rows
25531 Default value is @samp{combined}.
25534 Specify display color mode.
25536 It accepts the following values:
25539 each channel is displayed in a separate color
25541 each channel is displayed using the same color scheme
25543 each channel is displayed using the rainbow color scheme
25545 each channel is displayed using the moreland color scheme
25547 each channel is displayed using the nebulae color scheme
25549 each channel is displayed using the fire color scheme
25551 each channel is displayed using the fiery color scheme
25553 each channel is displayed using the fruit color scheme
25555 each channel is displayed using the cool color scheme
25557 each channel is displayed using the magma color scheme
25559 each channel is displayed using the green color scheme
25561 each channel is displayed using the viridis color scheme
25563 each channel is displayed using the plasma color scheme
25565 each channel is displayed using the cividis color scheme
25567 each channel is displayed using the terrain color scheme
25569 Default value is @samp{intensity}.
25572 Specify scale used for calculating intensity color values.
25574 It accepts the following values:
25579 square root, default
25589 Default value is @samp{log}.
25592 Specify frequency scale.
25594 It accepts the following values:
25602 Default value is @samp{lin}.
25605 Set saturation modifier for displayed colors. Negative values provide
25606 alternative color scheme. @code{0} is no saturation at all.
25607 Saturation must be in [-10.0, 10.0] range.
25608 Default value is @code{1}.
25611 Set window function.
25613 It accepts the following values:
25637 Default value is @code{hann}.
25640 Set orientation of time vs frequency axis. Can be @code{vertical} or
25641 @code{horizontal}. Default is @code{vertical}.
25644 Set scale gain for calculating intensity color values.
25645 Default value is @code{1}.
25648 Draw time and frequency axes and legends. Default is enabled.
25651 Set color rotation, must be in [-1.0, 1.0] range.
25652 Default value is @code{0}.
25655 Set start frequency from which to display spectrogram. Default is @code{0}.
25658 Set stop frequency to which to display spectrogram. Default is @code{0}.
25661 @subsection Examples
25665 Extract an audio spectrogram of a whole audio track
25666 in a 1024x1024 picture using @command{ffmpeg}:
25668 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
25672 @section showvolume
25674 Convert input audio volume to a video output.
25676 The filter accepts the following options:
25683 Set border width, allowed range is [0, 5]. Default is 1.
25686 Set channel width, allowed range is [80, 8192]. Default is 400.
25689 Set channel height, allowed range is [1, 900]. Default is 20.
25692 Set fade, allowed range is [0, 1]. Default is 0.95.
25695 Set volume color expression.
25697 The expression can use the following variables:
25701 Current max volume of channel in dB.
25707 Current channel number, starting from 0.
25711 If set, displays channel names. Default is enabled.
25714 If set, displays volume values. Default is enabled.
25717 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
25718 default is @code{h}.
25721 Set step size, allowed range is [0, 5]. Default is 0, which means
25725 Set background opacity, allowed range is [0, 1]. Default is 0.
25728 Set metering mode, can be peak: @code{p} or rms: @code{r},
25729 default is @code{p}.
25732 Set display scale, can be linear: @code{lin} or log: @code{log},
25733 default is @code{lin}.
25737 If set to > 0., display a line for the max level
25738 in the previous seconds.
25739 default is disabled: @code{0.}
25742 The color of the max line. Use when @code{dm} option is set to > 0.
25743 default is: @code{orange}
25748 Convert input audio to a video output, representing the samples waves.
25750 The filter accepts the following options:
25754 Specify the video size for the output. For the syntax of this option, check the
25755 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25756 Default value is @code{600x240}.
25761 Available values are:
25764 Draw a point for each sample.
25767 Draw a vertical line for each sample.
25770 Draw a point for each sample and a line between them.
25773 Draw a centered vertical line for each sample.
25776 Default value is @code{point}.
25779 Set the number of samples which are printed on the same column. A
25780 larger value will decrease the frame rate. Must be a positive
25781 integer. This option can be set only if the value for @var{rate}
25782 is not explicitly specified.
25785 Set the (approximate) output frame rate. This is done by setting the
25786 option @var{n}. Default value is "25".
25788 @item split_channels
25789 Set if channels should be drawn separately or overlap. Default value is 0.
25792 Set colors separated by '|' which are going to be used for drawing of each channel.
25795 Set amplitude scale.
25797 Available values are:
25815 Set the draw mode. This is mostly useful to set for high @var{n}.
25817 Available values are:
25820 Scale pixel values for each drawn sample.
25823 Draw every sample directly.
25826 Default value is @code{scale}.
25829 @subsection Examples
25833 Output the input file audio and the corresponding video representation
25836 amovie=a.mp3,asplit[out0],showwaves[out1]
25840 Create a synthetic signal and show it with showwaves, forcing a
25841 frame rate of 30 frames per second:
25843 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
25847 @section showwavespic
25849 Convert input audio to a single video frame, representing the samples waves.
25851 The filter accepts the following options:
25855 Specify the video size for the output. For the syntax of this option, check the
25856 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25857 Default value is @code{600x240}.
25859 @item split_channels
25860 Set if channels should be drawn separately or overlap. Default value is 0.
25863 Set colors separated by '|' which are going to be used for drawing of each channel.
25866 Set amplitude scale.
25868 Available values are:
25888 Available values are:
25891 Scale pixel values for each drawn sample.
25894 Draw every sample directly.
25897 Default value is @code{scale}.
25900 Set the filter mode.
25902 Available values are:
25905 Use average samples values for each drawn sample.
25908 Use peak samples values for each drawn sample.
25911 Default value is @code{average}.
25914 @subsection Examples
25918 Extract a channel split representation of the wave form of a whole audio track
25919 in a 1024x800 picture using @command{ffmpeg}:
25921 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
25925 @section sidedata, asidedata
25927 Delete frame side data, or select frames based on it.
25929 This filter accepts the following options:
25933 Set mode of operation of the filter.
25935 Can be one of the following:
25939 Select every frame with side data of @code{type}.
25942 Delete side data of @code{type}. If @code{type} is not set, delete all side
25948 Set side data type used with all modes. Must be set for @code{select} mode. For
25949 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
25950 in @file{libavutil/frame.h}. For example, to choose
25951 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
25955 @section spectrumsynth
25957 Synthesize audio from 2 input video spectrums, first input stream represents
25958 magnitude across time and second represents phase across time.
25959 The filter will transform from frequency domain as displayed in videos back
25960 to time domain as presented in audio output.
25962 This filter is primarily created for reversing processed @ref{showspectrum}
25963 filter outputs, but can synthesize sound from other spectrograms too.
25964 But in such case results are going to be poor if the phase data is not
25965 available, because in such cases phase data need to be recreated, usually
25966 it's just recreated from random noise.
25967 For best results use gray only output (@code{channel} color mode in
25968 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
25969 @code{lin} scale for phase video. To produce phase, for 2nd video, use
25970 @code{data} option. Inputs videos should generally use @code{fullframe}
25971 slide mode as that saves resources needed for decoding video.
25973 The filter accepts the following options:
25977 Specify sample rate of output audio, the sample rate of audio from which
25978 spectrum was generated may differ.
25981 Set number of channels represented in input video spectrums.
25984 Set scale which was used when generating magnitude input spectrum.
25985 Can be @code{lin} or @code{log}. Default is @code{log}.
25988 Set slide which was used when generating inputs spectrums.
25989 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
25990 Default is @code{fullframe}.
25993 Set window function used for resynthesis.
25996 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25997 which means optimal overlap for selected window function will be picked.
26000 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26001 Default is @code{vertical}.
26004 @subsection Examples
26008 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26009 then resynthesize videos back to audio with spectrumsynth:
26011 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
26012 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
26013 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26017 @section split, asplit
26019 Split input into several identical outputs.
26021 @code{asplit} works with audio input, @code{split} with video.
26023 The filter accepts a single parameter which specifies the number of outputs. If
26024 unspecified, it defaults to 2.
26026 @subsection Examples
26030 Create two separate outputs from the same input:
26032 [in] split [out0][out1]
26036 To create 3 or more outputs, you need to specify the number of
26039 [in] asplit=3 [out0][out1][out2]
26043 Create two separate outputs from the same input, one cropped and
26046 [in] split [splitout1][splitout2];
26047 [splitout1] crop=100:100:0:0 [cropout];
26048 [splitout2] pad=200:200:100:100 [padout];
26052 Create 5 copies of the input audio with @command{ffmpeg}:
26054 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26060 Receive commands sent through a libzmq client, and forward them to
26061 filters in the filtergraph.
26063 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26064 must be inserted between two video filters, @code{azmq} between two
26065 audio filters. Both are capable to send messages to any filter type.
26067 To enable these filters you need to install the libzmq library and
26068 headers and configure FFmpeg with @code{--enable-libzmq}.
26070 For more information about libzmq see:
26071 @url{http://www.zeromq.org/}
26073 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26074 receives messages sent through a network interface defined by the
26075 @option{bind_address} (or the abbreviation "@option{b}") option.
26076 Default value of this option is @file{tcp://localhost:5555}. You may
26077 want to alter this value to your needs, but do not forget to escape any
26078 ':' signs (see @ref{filtergraph escaping}).
26080 The received message must be in the form:
26082 @var{TARGET} @var{COMMAND} [@var{ARG}]
26085 @var{TARGET} specifies the target of the command, usually the name of
26086 the filter class or a specific filter instance name. The default
26087 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26088 but you can override this by using the @samp{filter_name@@id} syntax
26089 (see @ref{Filtergraph syntax}).
26091 @var{COMMAND} specifies the name of the command for the target filter.
26093 @var{ARG} is optional and specifies the optional argument list for the
26094 given @var{COMMAND}.
26096 Upon reception, the message is processed and the corresponding command
26097 is injected into the filtergraph. Depending on the result, the filter
26098 will send a reply to the client, adopting the format:
26100 @var{ERROR_CODE} @var{ERROR_REASON}
26104 @var{MESSAGE} is optional.
26106 @subsection Examples
26108 Look at @file{tools/zmqsend} for an example of a zmq client which can
26109 be used to send commands processed by these filters.
26111 Consider the following filtergraph generated by @command{ffplay}.
26112 In this example the last overlay filter has an instance name. All other
26113 filters will have default instance names.
26116 ffplay -dumpgraph 1 -f lavfi "
26117 color=s=100x100:c=red [l];
26118 color=s=100x100:c=blue [r];
26119 nullsrc=s=200x100, zmq [bg];
26120 [bg][l] overlay [bg+l];
26121 [bg+l][r] overlay@@my=x=100 "
26124 To change the color of the left side of the video, the following
26125 command can be used:
26127 echo Parsed_color_0 c yellow | tools/zmqsend
26130 To change the right side:
26132 echo Parsed_color_1 c pink | tools/zmqsend
26135 To change the position of the right side:
26137 echo overlay@@my x 150 | tools/zmqsend
26141 @c man end MULTIMEDIA FILTERS
26143 @chapter Multimedia Sources
26144 @c man begin MULTIMEDIA SOURCES
26146 Below is a description of the currently available multimedia sources.
26150 This is the same as @ref{movie} source, except it selects an audio
26156 Read audio and/or video stream(s) from a movie container.
26158 It accepts the following parameters:
26162 The name of the resource to read (not necessarily a file; it can also be a
26163 device or a stream accessed through some protocol).
26165 @item format_name, f
26166 Specifies the format assumed for the movie to read, and can be either
26167 the name of a container or an input device. If not specified, the
26168 format is guessed from @var{movie_name} or by probing.
26170 @item seek_point, sp
26171 Specifies the seek point in seconds. The frames will be output
26172 starting from this seek point. The parameter is evaluated with
26173 @code{av_strtod}, so the numerical value may be suffixed by an IS
26174 postfix. The default value is "0".
26177 Specifies the streams to read. Several streams can be specified,
26178 separated by "+". The source will then have as many outputs, in the
26179 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26180 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26181 respectively the default (best suited) video and audio stream. Default
26182 is "dv", or "da" if the filter is called as "amovie".
26184 @item stream_index, si
26185 Specifies the index of the video stream to read. If the value is -1,
26186 the most suitable video stream will be automatically selected. The default
26187 value is "-1". Deprecated. If the filter is called "amovie", it will select
26188 audio instead of video.
26191 Specifies how many times to read the stream in sequence.
26192 If the value is 0, the stream will be looped infinitely.
26193 Default value is "1".
26195 Note that when the movie is looped the source timestamps are not
26196 changed, so it will generate non monotonically increasing timestamps.
26198 @item discontinuity
26199 Specifies the time difference between frames above which the point is
26200 considered a timestamp discontinuity which is removed by adjusting the later
26204 It allows overlaying a second video on top of the main input of
26205 a filtergraph, as shown in this graph:
26207 input -----------> deltapts0 --> overlay --> output
26210 movie --> scale--> deltapts1 -------+
26212 @subsection Examples
26216 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26217 on top of the input labelled "in":
26219 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26220 [in] setpts=PTS-STARTPTS [main];
26221 [main][over] overlay=16:16 [out]
26225 Read from a video4linux2 device, and overlay it on top of the input
26228 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26229 [in] setpts=PTS-STARTPTS [main];
26230 [main][over] overlay=16:16 [out]
26234 Read the first video stream and the audio stream with id 0x81 from
26235 dvd.vob; the video is connected to the pad named "video" and the audio is
26236 connected to the pad named "audio":
26238 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26242 @subsection Commands
26244 Both movie and amovie support the following commands:
26247 Perform seek using "av_seek_frame".
26248 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26251 @var{stream_index}: If stream_index is -1, a default
26252 stream is selected, and @var{timestamp} is automatically converted
26253 from AV_TIME_BASE units to the stream specific time_base.
26255 @var{timestamp}: Timestamp in AVStream.time_base units
26256 or, if no stream is specified, in AV_TIME_BASE units.
26258 @var{flags}: Flags which select direction and seeking mode.
26262 Get movie duration in AV_TIME_BASE units.
26266 @c man end MULTIMEDIA SOURCES