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}.
1483 Apply an arbitrary Infinite Impulse Response filter.
1485 It accepts the following parameters:
1489 Set B/numerator/zeros/reflection coefficients.
1492 Set A/denominator/poles/ladder coefficients.
1504 Set coefficients format.
1508 lattice-ladder function
1510 analog transfer function
1512 digital transfer function
1514 Z-plane zeros/poles, cartesian (default)
1516 Z-plane zeros/poles, polar radians
1518 Z-plane zeros/poles, polar degrees
1524 Set type of processing.
1536 Set filtering precision.
1540 double-precision floating-point (default)
1542 single-precision floating-point
1550 Normalize filter coefficients, by default is enabled.
1551 Enabling it will normalize magnitude response at DC to 0dB.
1554 How much to use filtered signal in output. Default is 1.
1555 Range is between 0 and 1.
1558 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1559 By default it is disabled.
1562 Set for which IR channel to display frequency response. By default is first channel
1563 displayed. This option is used only when @var{response} is enabled.
1566 Set video stream size. This option is used only when @var{response} is enabled.
1569 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1572 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1573 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1576 Different coefficients and gains can be provided for every channel, in such case
1577 use '|' to separate coefficients or gains. Last provided coefficients will be
1578 used for all remaining channels.
1580 @subsection Examples
1584 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1586 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
1590 Same as above but in @code{zp} format:
1592 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
1596 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1598 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1604 The limiter prevents an input signal from rising over a desired threshold.
1605 This limiter uses lookahead technology to prevent your signal from distorting.
1606 It means that there is a small delay after the signal is processed. Keep in mind
1607 that the delay it produces is the attack time you set.
1609 The filter accepts the following options:
1613 Set input gain. Default is 1.
1616 Set output gain. Default is 1.
1619 Don't let signals above this level pass the limiter. Default is 1.
1622 The limiter will reach its attenuation level in this amount of time in
1623 milliseconds. Default is 5 milliseconds.
1626 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1627 Default is 50 milliseconds.
1630 When gain reduction is always needed ASC takes care of releasing to an
1631 average reduction level rather than reaching a reduction of 0 in the release
1635 Select how much the release time is affected by ASC, 0 means nearly no changes
1636 in release time while 1 produces higher release times.
1639 Auto level output signal. Default is enabled.
1640 This normalizes audio back to 0dB if enabled.
1643 Depending on picked setting it is recommended to upsample input 2x or 4x times
1644 with @ref{aresample} before applying this filter.
1648 Apply a two-pole all-pass filter with central frequency (in Hz)
1649 @var{frequency}, and filter-width @var{width}.
1650 An all-pass filter changes the audio's frequency to phase relationship
1651 without changing its frequency to amplitude relationship.
1653 The filter accepts the following options:
1657 Set frequency in Hz.
1660 Set method to specify band-width of filter.
1675 Specify the band-width of a filter in width_type units.
1678 How much to use filtered signal in output. Default is 1.
1679 Range is between 0 and 1.
1682 Specify which channels to filter, by default all available are filtered.
1685 Normalize biquad coefficients, by default is disabled.
1686 Enabling it will normalize magnitude response at DC to 0dB.
1689 Set the filter order, can be 1 or 2. Default is 2.
1692 Set transform type of IIR filter.
1701 @subsection Commands
1703 This filter supports the following commands:
1706 Change allpass frequency.
1707 Syntax for the command is : "@var{frequency}"
1710 Change allpass width_type.
1711 Syntax for the command is : "@var{width_type}"
1714 Change allpass width.
1715 Syntax for the command is : "@var{width}"
1719 Syntax for the command is : "@var{mix}"
1726 The filter accepts the following options:
1730 Set the number of loops. Setting this value to -1 will result in infinite loops.
1734 Set maximal number of samples. Default is 0.
1737 Set first sample of loop. Default is 0.
1743 Merge two or more audio streams into a single multi-channel stream.
1745 The filter accepts the following options:
1750 Set the number of inputs. Default is 2.
1754 If the channel layouts of the inputs are disjoint, and therefore compatible,
1755 the channel layout of the output will be set accordingly and the channels
1756 will be reordered as necessary. If the channel layouts of the inputs are not
1757 disjoint, the output will have all the channels of the first input then all
1758 the channels of the second input, in that order, and the channel layout of
1759 the output will be the default value corresponding to the total number of
1762 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1763 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1764 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1765 first input, b1 is the first channel of the second input).
1767 On the other hand, if both input are in stereo, the output channels will be
1768 in the default order: a1, a2, b1, b2, and the channel layout will be
1769 arbitrarily set to 4.0, which may or may not be the expected value.
1771 All inputs must have the same sample rate, and format.
1773 If inputs do not have the same duration, the output will stop with the
1776 @subsection Examples
1780 Merge two mono files into a stereo stream:
1782 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1786 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1788 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
1794 Mixes multiple audio inputs into a single output.
1796 Note that this filter only supports float samples (the @var{amerge}
1797 and @var{pan} audio filters support many formats). If the @var{amix}
1798 input has integer samples then @ref{aresample} will be automatically
1799 inserted to perform the conversion to float samples.
1803 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1805 will mix 3 input audio streams to a single output with the same duration as the
1806 first input and a dropout transition time of 3 seconds.
1808 It accepts the following parameters:
1812 The number of inputs. If unspecified, it defaults to 2.
1815 How to determine the end-of-stream.
1819 The duration of the longest input. (default)
1822 The duration of the shortest input.
1825 The duration of the first input.
1829 @item dropout_transition
1830 The transition time, in seconds, for volume renormalization when an input
1831 stream ends. The default value is 2 seconds.
1834 Specify weight of each input audio stream as sequence.
1835 Each weight is separated by space. By default all inputs have same weight.
1838 @subsection Commands
1840 This filter supports the following commands:
1843 Syntax is same as option with same name.
1848 Multiply first audio stream with second audio stream and store result
1849 in output audio stream. Multiplication is done by multiplying each
1850 sample from first stream with sample at same position from second stream.
1852 With this element-wise multiplication one can create amplitude fades and
1853 amplitude modulations.
1855 @section anequalizer
1857 High-order parametric multiband equalizer for each channel.
1859 It accepts the following parameters:
1863 This option string is in format:
1864 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1865 Each equalizer band is separated by '|'.
1869 Set channel number to which equalization will be applied.
1870 If input doesn't have that channel the entry is ignored.
1873 Set central frequency for band.
1874 If input doesn't have that frequency the entry is ignored.
1877 Set band width in Hertz.
1880 Set band gain in dB.
1883 Set filter type for band, optional, can be:
1887 Butterworth, this is default.
1898 With this option activated frequency response of anequalizer is displayed
1902 Set video stream size. Only useful if curves option is activated.
1905 Set max gain that will be displayed. Only useful if curves option is activated.
1906 Setting this to a reasonable value makes it possible to display gain which is derived from
1907 neighbour bands which are too close to each other and thus produce higher gain
1908 when both are activated.
1911 Set frequency scale used to draw frequency response in video output.
1912 Can be linear or logarithmic. Default is logarithmic.
1915 Set color for each channel curve which is going to be displayed in video stream.
1916 This is list of color names separated by space or by '|'.
1917 Unrecognised or missing colors will be replaced by white color.
1920 @subsection Examples
1924 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1925 for first 2 channels using Chebyshev type 1 filter:
1927 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1931 @subsection Commands
1933 This filter supports the following commands:
1936 Alter existing filter parameters.
1937 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1939 @var{fN} is existing filter number, starting from 0, if no such filter is available
1941 @var{freq} set new frequency parameter.
1942 @var{width} set new width parameter in Hertz.
1943 @var{gain} set new gain parameter in dB.
1945 Full filter invocation with asendcmd may look like this:
1946 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1951 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1953 Each sample is adjusted by looking for other samples with similar contexts. This
1954 context similarity is defined by comparing their surrounding patches of size
1955 @option{p}. Patches are searched in an area of @option{r} around the sample.
1957 The filter accepts the following options:
1961 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1964 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1965 Default value is 2 milliseconds.
1968 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1969 Default value is 6 milliseconds.
1972 Set the output mode.
1974 It accepts the following values:
1977 Pass input unchanged.
1980 Pass noise filtered out.
1985 Default value is @var{o}.
1989 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
1992 @subsection Commands
1994 This filter supports the all above options as @ref{commands}.
1997 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
1999 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2000 relate to producing the least mean square of the error signal (difference between the desired,
2001 2nd input audio stream and the actual signal, the 1st input audio stream).
2003 A description of the accepted options follows.
2016 Set the filter leakage.
2019 It accepts the following values:
2028 Pass filtered samples.
2031 Pass difference between desired and filtered samples.
2033 Default value is @var{o}.
2037 @subsection Examples
2041 One of many usages of this filter is noise reduction, input audio is filtered
2042 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2044 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2048 @subsection Commands
2050 This filter supports the same commands as options, excluding option @code{order}.
2054 Pass the audio source unchanged to the output.
2058 Pad the end of an audio stream with silence.
2060 This can be used together with @command{ffmpeg} @option{-shortest} to
2061 extend audio streams to the same length as the video stream.
2063 A description of the accepted options follows.
2067 Set silence packet size. Default value is 4096.
2070 Set the number of samples of silence to add to the end. After the
2071 value is reached, the stream is terminated. This option is mutually
2072 exclusive with @option{whole_len}.
2075 Set the minimum total number of samples in the output audio stream. If
2076 the value is longer than the input audio length, silence is added to
2077 the end, until the value is reached. This option is mutually exclusive
2078 with @option{pad_len}.
2081 Specify the duration of samples of silence to add. See
2082 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2083 for the accepted syntax. Used only if set to non-zero value.
2086 Specify the minimum total duration in the output audio stream. See
2087 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2088 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2089 the input audio length, silence is added to the end, until the value is reached.
2090 This option is mutually exclusive with @option{pad_dur}
2093 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2094 nor @option{whole_dur} option is set, the filter will add silence to the end of
2095 the input stream indefinitely.
2097 @subsection Examples
2101 Add 1024 samples of silence to the end of the input:
2107 Make sure the audio output will contain at least 10000 samples, pad
2108 the input with silence if required:
2110 apad=whole_len=10000
2114 Use @command{ffmpeg} to pad the audio input with silence, so that the
2115 video stream will always result the shortest and will be converted
2116 until the end in the output file when using the @option{shortest}
2119 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2124 Add a phasing effect to the input audio.
2126 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2127 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2129 A description of the accepted parameters follows.
2133 Set input gain. Default is 0.4.
2136 Set output gain. Default is 0.74
2139 Set delay in milliseconds. Default is 3.0.
2142 Set decay. Default is 0.4.
2145 Set modulation speed in Hz. Default is 0.5.
2148 Set modulation type. Default is triangular.
2150 It accepts the following values:
2157 @section aphaseshift
2158 Apply phase shift to input audio samples.
2160 The filter accepts the following options:
2164 Specify phase shift. Allowed range is from -1.0 to 1.0.
2165 Default value is 0.0.
2168 @subsection Commands
2170 This filter supports the above option as @ref{commands}.
2174 Audio pulsator is something between an autopanner and a tremolo.
2175 But it can produce funny stereo effects as well. Pulsator changes the volume
2176 of the left and right channel based on a LFO (low frequency oscillator) with
2177 different waveforms and shifted phases.
2178 This filter have the ability to define an offset between left and right
2179 channel. An offset of 0 means that both LFO shapes match each other.
2180 The left and right channel are altered equally - a conventional tremolo.
2181 An offset of 50% means that the shape of the right channel is exactly shifted
2182 in phase (or moved backwards about half of the frequency) - pulsator acts as
2183 an autopanner. At 1 both curves match again. Every setting in between moves the
2184 phase shift gapless between all stages and produces some "bypassing" sounds with
2185 sine and triangle waveforms. The more you set the offset near 1 (starting from
2186 the 0.5) the faster the signal passes from the left to the right speaker.
2188 The filter accepts the following options:
2192 Set input gain. By default it is 1. Range is [0.015625 - 64].
2195 Set output gain. By default it is 1. Range is [0.015625 - 64].
2198 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2199 sawup or sawdown. Default is sine.
2202 Set modulation. Define how much of original signal is affected by the LFO.
2205 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2208 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2211 Set pulse width. Default is 1. Allowed range is [0 - 2].
2214 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2217 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2221 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2225 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2226 if timing is set to hz.
2232 Resample the input audio to the specified parameters, using the
2233 libswresample library. If none are specified then the filter will
2234 automatically convert between its input and output.
2236 This filter is also able to stretch/squeeze the audio data to make it match
2237 the timestamps or to inject silence / cut out audio to make it match the
2238 timestamps, do a combination of both or do neither.
2240 The filter accepts the syntax
2241 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2242 expresses a sample rate and @var{resampler_options} is a list of
2243 @var{key}=@var{value} pairs, separated by ":". See the
2244 @ref{Resampler Options,,"Resampler Options" section in the
2245 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2246 for the complete list of supported options.
2248 @subsection Examples
2252 Resample the input audio to 44100Hz:
2258 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2259 samples per second compensation:
2261 aresample=async=1000
2267 Reverse an audio clip.
2269 Warning: This filter requires memory to buffer the entire clip, so trimming
2272 @subsection Examples
2276 Take the first 5 seconds of a clip, and reverse it.
2278 atrim=end=5,areverse
2284 Reduce noise from speech using Recurrent Neural Networks.
2286 This filter accepts the following options:
2290 Set train model file to load. This option is always required.
2293 @section asetnsamples
2295 Set the number of samples per each output audio frame.
2297 The last output packet may contain a different number of samples, as
2298 the filter will flush all the remaining samples when the input audio
2301 The filter accepts the following options:
2305 @item nb_out_samples, n
2306 Set the number of frames per each output audio frame. The number is
2307 intended as the number of samples @emph{per each channel}.
2308 Default value is 1024.
2311 If set to 1, the filter will pad the last audio frame with zeroes, so
2312 that the last frame will contain the same number of samples as the
2313 previous ones. Default value is 1.
2316 For example, to set the number of per-frame samples to 1234 and
2317 disable padding for the last frame, use:
2319 asetnsamples=n=1234:p=0
2324 Set the sample rate without altering the PCM data.
2325 This will result in a change of speed and pitch.
2327 The filter accepts the following options:
2330 @item sample_rate, r
2331 Set the output sample rate. Default is 44100 Hz.
2336 Show a line containing various information for each input audio frame.
2337 The input audio is not modified.
2339 The shown line contains a sequence of key/value pairs of the form
2340 @var{key}:@var{value}.
2342 The following values are shown in the output:
2346 The (sequential) number of the input frame, starting from 0.
2349 The presentation timestamp of the input frame, in time base units; the time base
2350 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2353 The presentation timestamp of the input frame in seconds.
2356 position of the frame in the input stream, -1 if this information in
2357 unavailable and/or meaningless (for example in case of synthetic audio)
2366 The sample rate for the audio frame.
2369 The number of samples (per channel) in the frame.
2372 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2373 audio, the data is treated as if all the planes were concatenated.
2375 @item plane_checksums
2376 A list of Adler-32 checksums for each data plane.
2380 Apply audio soft clipping.
2382 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2383 along a smooth curve, rather than the abrupt shape of hard-clipping.
2385 This filter accepts the following options:
2389 Set type of soft-clipping.
2391 It accepts the following values:
2405 Set additional parameter which controls sigmoid function.
2408 Set oversampling factor.
2411 @subsection Commands
2413 This filter supports the all above options as @ref{commands}.
2416 Automatic Speech Recognition
2418 This filter uses PocketSphinx for speech recognition. To enable
2419 compilation of this filter, you need to configure FFmpeg with
2420 @code{--enable-pocketsphinx}.
2422 It accepts the following options:
2426 Set sampling rate of input audio. Defaults is @code{16000}.
2427 This need to match speech models, otherwise one will get poor results.
2430 Set dictionary containing acoustic model files.
2433 Set pronunciation dictionary.
2436 Set language model file.
2439 Set language model set.
2442 Set which language model to use.
2445 Set output for log messages.
2448 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2453 Display time domain statistical information about the audio channels.
2454 Statistics are calculated and displayed for each audio channel and,
2455 where applicable, an overall figure is also given.
2457 It accepts the following option:
2460 Short window length in seconds, used for peak and trough RMS measurement.
2461 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2465 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2466 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2469 Available keys for each channel are:
2515 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2516 this @code{lavfi.astats.Overall.Peak_count}.
2518 For description what each key means read below.
2521 Set number of frame after which stats are going to be recalculated.
2522 Default is disabled.
2524 @item measure_perchannel
2525 Select the entries which need to be measured per channel. The metadata keys can
2526 be used as flags, default is @option{all} which measures everything.
2527 @option{none} disables all per channel measurement.
2529 @item measure_overall
2530 Select the entries which need to be measured overall. The metadata keys can
2531 be used as flags, default is @option{all} which measures everything.
2532 @option{none} disables all overall measurement.
2536 A description of each shown parameter follows:
2540 Mean amplitude displacement from zero.
2543 Minimal sample level.
2546 Maximal sample level.
2548 @item Min difference
2549 Minimal difference between two consecutive samples.
2551 @item Max difference
2552 Maximal difference between two consecutive samples.
2554 @item Mean difference
2555 Mean difference between two consecutive samples.
2556 The average of each difference between two consecutive samples.
2558 @item RMS difference
2559 Root Mean Square difference between two consecutive samples.
2563 Standard peak and RMS level measured in dBFS.
2567 Peak and trough values for RMS level measured over a short window.
2570 Standard ratio of peak to RMS level (note: not in dB).
2573 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2574 (i.e. either @var{Min level} or @var{Max level}).
2577 Number of occasions (not the number of samples) that the signal attained either
2578 @var{Min level} or @var{Max level}.
2580 @item Noise floor dB
2581 Minimum local peak measured in dBFS over a short window.
2583 @item Noise floor count
2584 Number of occasions (not the number of samples) that the signal attained
2588 Overall bit depth of audio. Number of bits used for each sample.
2591 Measured dynamic range of audio in dB.
2593 @item Zero crossings
2594 Number of points where the waveform crosses the zero level axis.
2596 @item Zero crossings rate
2597 Rate of Zero crossings and number of audio samples.
2601 Boost subwoofer frequencies.
2603 The filter accepts the following options:
2607 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2608 Default value is 0.5.
2611 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2612 Default value is 0.8.
2615 Set delay line decay gain value. Allowed range is from 0 to 1.
2616 Default value is 0.7.
2619 Set delay line feedback gain value. Allowed range is from 0 to 1.
2620 Default value is 0.5.
2623 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2624 Default value is 100.
2627 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2628 Default value is 0.5.
2631 Set delay. Allowed range is from 1 to 100.
2632 Default value is 20.
2635 @subsection Commands
2637 This filter supports the all above options as @ref{commands}.
2640 Cut super frequencies.
2642 The filter accepts the following options:
2646 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2647 Default value is 20000.
2650 @subsection Commands
2652 This filter supports the all above options as @ref{commands}.
2658 The filter accepts exactly one parameter, the audio tempo. If not
2659 specified then the filter will assume nominal 1.0 tempo. Tempo must
2660 be in the [0.5, 100.0] range.
2662 Note that tempo greater than 2 will skip some samples rather than
2663 blend them in. If for any reason this is a concern it is always
2664 possible to daisy-chain several instances of atempo to achieve the
2665 desired product tempo.
2667 @subsection Examples
2671 Slow down audio to 80% tempo:
2677 To speed up audio to 300% tempo:
2683 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2685 atempo=sqrt(3),atempo=sqrt(3)
2689 @subsection Commands
2691 This filter supports the following commands:
2694 Change filter tempo scale factor.
2695 Syntax for the command is : "@var{tempo}"
2700 Trim the input so that the output contains one continuous subpart of the input.
2702 It accepts the following parameters:
2705 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2706 sample with the timestamp @var{start} will be the first sample in the output.
2709 Specify time of the first audio sample that will be dropped, i.e. the
2710 audio sample immediately preceding the one with the timestamp @var{end} will be
2711 the last sample in the output.
2714 Same as @var{start}, except this option sets the start timestamp in samples
2718 Same as @var{end}, except this option sets the end timestamp in samples instead
2722 The maximum duration of the output in seconds.
2725 The number of the first sample that should be output.
2728 The number of the first sample that should be dropped.
2731 @option{start}, @option{end}, and @option{duration} are expressed as time
2732 duration specifications; see
2733 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2735 Note that the first two sets of the start/end options and the @option{duration}
2736 option look at the frame timestamp, while the _sample options simply count the
2737 samples that pass through the filter. So start/end_pts and start/end_sample will
2738 give different results when the timestamps are wrong, inexact or do not start at
2739 zero. Also note that this filter does not modify the timestamps. If you wish
2740 to have the output timestamps start at zero, insert the asetpts filter after the
2743 If multiple start or end options are set, this filter tries to be greedy and
2744 keep all samples that match at least one of the specified constraints. To keep
2745 only the part that matches all the constraints at once, chain multiple atrim
2748 The defaults are such that all the input is kept. So it is possible to set e.g.
2749 just the end values to keep everything before the specified time.
2754 Drop everything except the second minute of input:
2756 ffmpeg -i INPUT -af atrim=60:120
2760 Keep only the first 1000 samples:
2762 ffmpeg -i INPUT -af atrim=end_sample=1000
2767 @section axcorrelate
2768 Calculate normalized cross-correlation between two input audio streams.
2770 Resulted samples are always between -1 and 1 inclusive.
2771 If result is 1 it means two input samples are highly correlated in that selected segment.
2772 Result 0 means they are not correlated at all.
2773 If result is -1 it means two input samples are out of phase, which means they cancel each
2776 The filter accepts the following options:
2780 Set size of segment over which cross-correlation is calculated.
2781 Default is 256. Allowed range is from 2 to 131072.
2784 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2785 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2786 are always zero and thus need much less calculations to make.
2787 This is generally not true, but is valid for typical audio streams.
2790 @subsection Examples
2794 Calculate correlation between channels in stereo audio stream:
2796 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2802 Apply a two-pole Butterworth band-pass filter with central
2803 frequency @var{frequency}, and (3dB-point) band-width width.
2804 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2805 instead of the default: constant 0dB peak gain.
2806 The filter roll off at 6dB per octave (20dB per decade).
2808 The filter accepts the following options:
2812 Set the filter's central frequency. Default is @code{3000}.
2815 Constant skirt gain if set to 1. Defaults to 0.
2818 Set method to specify band-width of filter.
2833 Specify the band-width of a filter in width_type units.
2836 How much to use filtered signal in output. Default is 1.
2837 Range is between 0 and 1.
2840 Specify which channels to filter, by default all available are filtered.
2843 Normalize biquad coefficients, by default is disabled.
2844 Enabling it will normalize magnitude response at DC to 0dB.
2847 Set transform type of IIR filter.
2856 @subsection Commands
2858 This filter supports the following commands:
2861 Change bandpass frequency.
2862 Syntax for the command is : "@var{frequency}"
2865 Change bandpass width_type.
2866 Syntax for the command is : "@var{width_type}"
2869 Change bandpass width.
2870 Syntax for the command is : "@var{width}"
2873 Change bandpass mix.
2874 Syntax for the command is : "@var{mix}"
2879 Apply a two-pole Butterworth band-reject filter with central
2880 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2881 The filter roll off at 6dB per octave (20dB per decade).
2883 The filter accepts the following options:
2887 Set the filter's central frequency. Default is @code{3000}.
2890 Set method to specify band-width of filter.
2905 Specify the band-width of a filter in width_type units.
2908 How much to use filtered signal in output. Default is 1.
2909 Range is between 0 and 1.
2912 Specify which channels to filter, by default all available are filtered.
2915 Normalize biquad coefficients, by default is disabled.
2916 Enabling it will normalize magnitude response at DC to 0dB.
2919 Set transform type of IIR filter.
2928 @subsection Commands
2930 This filter supports the following commands:
2933 Change bandreject frequency.
2934 Syntax for the command is : "@var{frequency}"
2937 Change bandreject width_type.
2938 Syntax for the command is : "@var{width_type}"
2941 Change bandreject width.
2942 Syntax for the command is : "@var{width}"
2945 Change bandreject mix.
2946 Syntax for the command is : "@var{mix}"
2949 @section bass, lowshelf
2951 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2952 shelving filter with a response similar to that of a standard
2953 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2955 The filter accepts the following options:
2959 Give the gain at 0 Hz. Its useful range is about -20
2960 (for a large cut) to +20 (for a large boost).
2961 Beware of clipping when using a positive gain.
2964 Set the filter's central frequency and so can be used
2965 to extend or reduce the frequency range to be boosted or cut.
2966 The default value is @code{100} Hz.
2969 Set method to specify band-width of filter.
2984 Determine how steep is the filter's shelf transition.
2987 How much to use filtered signal in output. Default is 1.
2988 Range is between 0 and 1.
2991 Specify which channels to filter, by default all available are filtered.
2994 Normalize biquad coefficients, by default is disabled.
2995 Enabling it will normalize magnitude response at DC to 0dB.
2998 Set transform type of IIR filter.
3007 @subsection Commands
3009 This filter supports the following commands:
3012 Change bass frequency.
3013 Syntax for the command is : "@var{frequency}"
3016 Change bass width_type.
3017 Syntax for the command is : "@var{width_type}"
3021 Syntax for the command is : "@var{width}"
3025 Syntax for the command is : "@var{gain}"
3029 Syntax for the command is : "@var{mix}"
3034 Apply a biquad IIR filter with the given coefficients.
3035 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3036 are the numerator and denominator coefficients respectively.
3037 and @var{channels}, @var{c} specify which channels to filter, by default all
3038 available are filtered.
3040 @subsection Commands
3042 This filter supports the following commands:
3050 Change biquad parameter.
3051 Syntax for the command is : "@var{value}"
3054 How much to use filtered signal in output. Default is 1.
3055 Range is between 0 and 1.
3058 Specify which channels to filter, by default all available are filtered.
3061 Normalize biquad coefficients, by default is disabled.
3062 Enabling it will normalize magnitude response at DC to 0dB.
3065 Set transform type of IIR filter.
3075 Bauer stereo to binaural transformation, which improves headphone listening of
3076 stereo audio records.
3078 To enable compilation of this filter you need to configure FFmpeg with
3079 @code{--enable-libbs2b}.
3081 It accepts the following parameters:
3085 Pre-defined crossfeed level.
3089 Default level (fcut=700, feed=50).
3092 Chu Moy circuit (fcut=700, feed=60).
3095 Jan Meier circuit (fcut=650, feed=95).
3100 Cut frequency (in Hz).
3109 Remap input channels to new locations.
3111 It accepts the following parameters:
3114 Map channels from input to output. The argument is a '|'-separated list of
3115 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3116 @var{in_channel} form. @var{in_channel} can be either the name of the input
3117 channel (e.g. FL for front left) or its index in the input channel layout.
3118 @var{out_channel} is the name of the output channel or its index in the output
3119 channel layout. If @var{out_channel} is not given then it is implicitly an
3120 index, starting with zero and increasing by one for each mapping.
3122 @item channel_layout
3123 The channel layout of the output stream.
3126 If no mapping is present, the filter will implicitly map input channels to
3127 output channels, preserving indices.
3129 @subsection Examples
3133 For example, assuming a 5.1+downmix input MOV file,
3135 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3137 will create an output WAV file tagged as stereo from the downmix channels of
3141 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3143 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3147 @section channelsplit
3149 Split each channel from an input audio stream into a separate output stream.
3151 It accepts the following parameters:
3153 @item channel_layout
3154 The channel layout of the input stream. The default is "stereo".
3156 A channel layout describing the channels to be extracted as separate output streams
3157 or "all" to extract each input channel as a separate stream. The default is "all".
3159 Choosing channels not present in channel layout in the input will result in an error.
3162 @subsection Examples
3166 For example, assuming a stereo input MP3 file,
3168 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3170 will create an output Matroska file with two audio streams, one containing only
3171 the left channel and the other the right channel.
3174 Split a 5.1 WAV file into per-channel files:
3176 ffmpeg -i in.wav -filter_complex
3177 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3178 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3179 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3184 Extract only LFE from a 5.1 WAV file:
3186 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3187 -map '[LFE]' lfe.wav
3192 Add a chorus effect to the audio.
3194 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3196 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3197 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3198 The modulation depth defines the range the modulated delay is played before or after
3199 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3200 sound tuned around the original one, like in a chorus where some vocals are slightly
3203 It accepts the following parameters:
3206 Set input gain. Default is 0.4.
3209 Set output gain. Default is 0.4.
3212 Set delays. A typical delay is around 40ms to 60ms.
3224 @subsection Examples
3230 chorus=0.7:0.9:55:0.4:0.25:2
3236 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3240 Fuller sounding chorus with three delays:
3242 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
3247 Compress or expand the audio's dynamic range.
3249 It accepts the following parameters:
3255 A list of times in seconds for each channel over which the instantaneous level
3256 of the input signal is averaged to determine its volume. @var{attacks} refers to
3257 increase of volume and @var{decays} refers to decrease of volume. For most
3258 situations, the attack time (response to the audio getting louder) should be
3259 shorter than the decay time, because the human ear is more sensitive to sudden
3260 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3261 a typical value for decay is 0.8 seconds.
3262 If specified number of attacks & decays is lower than number of channels, the last
3263 set attack/decay will be used for all remaining channels.
3266 A list of points for the transfer function, specified in dB relative to the
3267 maximum possible signal amplitude. Each key points list must be defined using
3268 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3269 @code{x0/y0 x1/y1 x2/y2 ....}
3271 The input values must be in strictly increasing order but the transfer function
3272 does not have to be monotonically rising. The point @code{0/0} is assumed but
3273 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3274 function are @code{-70/-70|-60/-20|1/0}.
3277 Set the curve radius in dB for all joints. It defaults to 0.01.
3280 Set the additional gain in dB to be applied at all points on the transfer
3281 function. This allows for easy adjustment of the overall gain.
3285 Set an initial volume, in dB, to be assumed for each channel when filtering
3286 starts. This permits the user to supply a nominal level initially, so that, for
3287 example, a very large gain is not applied to initial signal levels before the
3288 companding has begun to operate. A typical value for audio which is initially
3289 quiet is -90 dB. It defaults to 0.
3292 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3293 delayed before being fed to the volume adjuster. Specifying a delay
3294 approximately equal to the attack/decay times allows the filter to effectively
3295 operate in predictive rather than reactive mode. It defaults to 0.
3299 @subsection Examples
3303 Make music with both quiet and loud passages suitable for listening to in a
3306 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3309 Another example for audio with whisper and explosion parts:
3311 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3315 A noise gate for when the noise is at a lower level than the signal:
3317 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3321 Here is another noise gate, this time for when the noise is at a higher level
3322 than the signal (making it, in some ways, similar to squelch):
3324 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3328 2:1 compression starting at -6dB:
3330 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3334 2:1 compression starting at -9dB:
3336 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3340 2:1 compression starting at -12dB:
3342 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3346 2:1 compression starting at -18dB:
3348 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3352 3:1 compression starting at -15dB:
3354 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3360 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3366 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
3370 Hard limiter at -6dB:
3372 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3376 Hard limiter at -12dB:
3378 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3382 Hard noise gate at -35 dB:
3384 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3390 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3394 @section compensationdelay
3396 Compensation Delay Line is a metric based delay to compensate differing
3397 positions of microphones or speakers.
3399 For example, you have recorded guitar with two microphones placed in
3400 different locations. Because the front of sound wave has fixed speed in
3401 normal conditions, the phasing of microphones can vary and depends on
3402 their location and interposition. The best sound mix can be achieved when
3403 these microphones are in phase (synchronized). Note that a distance of
3404 ~30 cm between microphones makes one microphone capture the signal in
3405 antiphase to the other microphone. That makes the final mix sound moody.
3406 This filter helps to solve phasing problems by adding different delays
3407 to each microphone track and make them synchronized.
3409 The best result can be reached when you take one track as base and
3410 synchronize other tracks one by one with it.
3411 Remember that synchronization/delay tolerance depends on sample rate, too.
3412 Higher sample rates will give more tolerance.
3414 The filter accepts the following parameters:
3418 Set millimeters distance. This is compensation distance for fine tuning.
3422 Set cm distance. This is compensation distance for tightening distance setup.
3426 Set meters distance. This is compensation distance for hard distance setup.
3430 Set dry amount. Amount of unprocessed (dry) signal.
3434 Set wet amount. Amount of processed (wet) signal.
3438 Set temperature in degrees Celsius. This is the temperature of the environment.
3443 Apply headphone crossfeed filter.
3445 Crossfeed is the process of blending the left and right channels of stereo
3447 It is mainly used to reduce extreme stereo separation of low frequencies.
3449 The intent is to produce more speaker like sound to the listener.
3451 The filter accepts the following options:
3455 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3456 This sets gain of low shelf filter for side part of stereo image.
3457 Default is -6dB. Max allowed is -30db when strength is set to 1.
3460 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3461 This sets cut off frequency of low shelf filter. Default is cut off near
3462 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3465 Set curve slope of low shelf filter. Default is 0.5.
3466 Allowed range is from 0.01 to 1.
3469 Set input gain. Default is 0.9.
3472 Set output gain. Default is 1.
3475 @subsection Commands
3477 This filter supports the all above options as @ref{commands}.
3479 @section crystalizer
3480 Simple algorithm to expand audio dynamic range.
3482 The filter accepts the following options:
3486 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3487 (unchanged sound) to 10.0 (maximum effect).
3490 Enable clipping. By default is enabled.
3493 @subsection Commands
3495 This filter supports the all above options as @ref{commands}.
3498 Apply a DC shift to the audio.
3500 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3501 in the recording chain) from the audio. The effect of a DC offset is reduced
3502 headroom and hence volume. The @ref{astats} filter can be used to determine if
3503 a signal has a DC offset.
3507 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3511 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3512 used to prevent clipping.
3517 Apply de-essing to the audio samples.
3521 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3525 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3529 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3533 Set the output mode.
3535 It accepts the following values:
3538 Pass input unchanged.
3541 Pass ess filtered out.
3546 Default value is @var{o}.
3552 Measure audio dynamic range.
3554 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3555 is found in transition material. And anything less that 8 have very poor dynamics
3556 and is very compressed.
3558 The filter accepts the following options:
3562 Set window length in seconds used to split audio into segments of equal length.
3563 Default is 3 seconds.
3567 Dynamic Audio Normalizer.
3569 This filter applies a certain amount of gain to the input audio in order
3570 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3571 contrast to more "simple" normalization algorithms, the Dynamic Audio
3572 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3573 This allows for applying extra gain to the "quiet" sections of the audio
3574 while avoiding distortions or clipping the "loud" sections. In other words:
3575 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3576 sections, in the sense that the volume of each section is brought to the
3577 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3578 this goal *without* applying "dynamic range compressing". It will retain 100%
3579 of the dynamic range *within* each section of the audio file.
3583 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3584 Default is 500 milliseconds.
3585 The Dynamic Audio Normalizer processes the input audio in small chunks,
3586 referred to as frames. This is required, because a peak magnitude has no
3587 meaning for just a single sample value. Instead, we need to determine the
3588 peak magnitude for a contiguous sequence of sample values. While a "standard"
3589 normalizer would simply use the peak magnitude of the complete file, the
3590 Dynamic Audio Normalizer determines the peak magnitude individually for each
3591 frame. The length of a frame is specified in milliseconds. By default, the
3592 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3593 been found to give good results with most files.
3594 Note that the exact frame length, in number of samples, will be determined
3595 automatically, based on the sampling rate of the individual input audio file.
3598 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3599 number. Default is 31.
3600 Probably the most important parameter of the Dynamic Audio Normalizer is the
3601 @code{window size} of the Gaussian smoothing filter. The filter's window size
3602 is specified in frames, centered around the current frame. For the sake of
3603 simplicity, this must be an odd number. Consequently, the default value of 31
3604 takes into account the current frame, as well as the 15 preceding frames and
3605 the 15 subsequent frames. Using a larger window results in a stronger
3606 smoothing effect and thus in less gain variation, i.e. slower gain
3607 adaptation. Conversely, using a smaller window results in a weaker smoothing
3608 effect and thus in more gain variation, i.e. faster gain adaptation.
3609 In other words, the more you increase this value, the more the Dynamic Audio
3610 Normalizer will behave like a "traditional" normalization filter. On the
3611 contrary, the more you decrease this value, the more the Dynamic Audio
3612 Normalizer will behave like a dynamic range compressor.
3615 Set the target peak value. This specifies the highest permissible magnitude
3616 level for the normalized audio input. This filter will try to approach the
3617 target peak magnitude as closely as possible, but at the same time it also
3618 makes sure that the normalized signal will never exceed the peak magnitude.
3619 A frame's maximum local gain factor is imposed directly by the target peak
3620 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3621 It is not recommended to go above this value.
3624 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3625 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3626 factor for each input frame, i.e. the maximum gain factor that does not
3627 result in clipping or distortion. The maximum gain factor is determined by
3628 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3629 additionally bounds the frame's maximum gain factor by a predetermined
3630 (global) maximum gain factor. This is done in order to avoid excessive gain
3631 factors in "silent" or almost silent frames. By default, the maximum gain
3632 factor is 10.0, For most inputs the default value should be sufficient and
3633 it usually is not recommended to increase this value. Though, for input
3634 with an extremely low overall volume level, it may be necessary to allow even
3635 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3636 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3637 Instead, a "sigmoid" threshold function will be applied. This way, the
3638 gain factors will smoothly approach the threshold value, but never exceed that
3642 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3643 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3644 This means that the maximum local gain factor for each frame is defined
3645 (only) by the frame's highest magnitude sample. This way, the samples can
3646 be amplified as much as possible without exceeding the maximum signal
3647 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3648 Normalizer can also take into account the frame's root mean square,
3649 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3650 determine the power of a time-varying signal. It is therefore considered
3651 that the RMS is a better approximation of the "perceived loudness" than
3652 just looking at the signal's peak magnitude. Consequently, by adjusting all
3653 frames to a constant RMS value, a uniform "perceived loudness" can be
3654 established. If a target RMS value has been specified, a frame's local gain
3655 factor is defined as the factor that would result in exactly that RMS value.
3656 Note, however, that the maximum local gain factor is still restricted by the
3657 frame's highest magnitude sample, in order to prevent clipping.
3660 Enable channels coupling. By default is enabled.
3661 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3662 amount. This means the same gain factor will be applied to all channels, i.e.
3663 the maximum possible gain factor is determined by the "loudest" channel.
3664 However, in some recordings, it may happen that the volume of the different
3665 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3666 In this case, this option can be used to disable the channel coupling. This way,
3667 the gain factor will be determined independently for each channel, depending
3668 only on the individual channel's highest magnitude sample. This allows for
3669 harmonizing the volume of the different channels.
3672 Enable DC bias correction. By default is disabled.
3673 An audio signal (in the time domain) is a sequence of sample values.
3674 In the Dynamic Audio Normalizer these sample values are represented in the
3675 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3676 audio signal, or "waveform", should be centered around the zero point.
3677 That means if we calculate the mean value of all samples in a file, or in a
3678 single frame, then the result should be 0.0 or at least very close to that
3679 value. If, however, there is a significant deviation of the mean value from
3680 0.0, in either positive or negative direction, this is referred to as a
3681 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3682 Audio Normalizer provides optional DC bias correction.
3683 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3684 the mean value, or "DC correction" offset, of each input frame and subtract
3685 that value from all of the frame's sample values which ensures those samples
3686 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3687 boundaries, the DC correction offset values will be interpolated smoothly
3688 between neighbouring frames.
3690 @item altboundary, b
3691 Enable alternative boundary mode. By default is disabled.
3692 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3693 around each frame. This includes the preceding frames as well as the
3694 subsequent frames. However, for the "boundary" frames, located at the very
3695 beginning and at the very end of the audio file, not all neighbouring
3696 frames are available. In particular, for the first few frames in the audio
3697 file, the preceding frames are not known. And, similarly, for the last few
3698 frames in the audio file, the subsequent frames are not known. Thus, the
3699 question arises which gain factors should be assumed for the missing frames
3700 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3701 to deal with this situation. The default boundary mode assumes a gain factor
3702 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3703 "fade out" at the beginning and at the end of the input, respectively.
3706 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3707 By default, the Dynamic Audio Normalizer does not apply "traditional"
3708 compression. This means that signal peaks will not be pruned and thus the
3709 full dynamic range will be retained within each local neighbourhood. However,
3710 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3711 normalization algorithm with a more "traditional" compression.
3712 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3713 (thresholding) function. If (and only if) the compression feature is enabled,
3714 all input frames will be processed by a soft knee thresholding function prior
3715 to the actual normalization process. Put simply, the thresholding function is
3716 going to prune all samples whose magnitude exceeds a certain threshold value.
3717 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3718 value. Instead, the threshold value will be adjusted for each individual
3720 In general, smaller parameters result in stronger compression, and vice versa.
3721 Values below 3.0 are not recommended, because audible distortion may appear.
3724 Set the target threshold value. This specifies the lowest permissible
3725 magnitude level for the audio input which will be normalized.
3726 If input frame volume is above this value frame will be normalized.
3727 Otherwise frame may not be normalized at all. The default value is set
3728 to 0, which means all input frames will be normalized.
3729 This option is mostly useful if digital noise is not wanted to be amplified.
3732 @subsection Commands
3734 This filter supports the all above options as @ref{commands}.
3738 Make audio easier to listen to on headphones.
3740 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3741 so that when listened to on headphones the stereo image is moved from
3742 inside your head (standard for headphones) to outside and in front of
3743 the listener (standard for speakers).
3749 Apply a two-pole peaking equalisation (EQ) filter. With this
3750 filter, the signal-level at and around a selected frequency can
3751 be increased or decreased, whilst (unlike bandpass and bandreject
3752 filters) that at all other frequencies is unchanged.
3754 In order to produce complex equalisation curves, this filter can
3755 be given several times, each with a different central frequency.
3757 The filter accepts the following options:
3761 Set the filter's central frequency in Hz.
3764 Set method to specify band-width of filter.
3779 Specify the band-width of a filter in width_type units.
3782 Set the required gain or attenuation in dB.
3783 Beware of clipping when using a positive gain.
3786 How much to use filtered signal in output. Default is 1.
3787 Range is between 0 and 1.
3790 Specify which channels to filter, by default all available are filtered.
3793 Normalize biquad coefficients, by default is disabled.
3794 Enabling it will normalize magnitude response at DC to 0dB.
3797 Set transform type of IIR filter.
3806 @subsection Examples
3809 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3811 equalizer=f=1000:t=h:width=200:g=-10
3815 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3817 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3821 @subsection Commands
3823 This filter supports the following commands:
3826 Change equalizer frequency.
3827 Syntax for the command is : "@var{frequency}"
3830 Change equalizer width_type.
3831 Syntax for the command is : "@var{width_type}"
3834 Change equalizer width.
3835 Syntax for the command is : "@var{width}"
3838 Change equalizer gain.
3839 Syntax for the command is : "@var{gain}"
3842 Change equalizer mix.
3843 Syntax for the command is : "@var{mix}"
3846 @section extrastereo
3848 Linearly increases the difference between left and right channels which
3849 adds some sort of "live" effect to playback.
3851 The filter accepts the following options:
3855 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3856 (average of both channels), with 1.0 sound will be unchanged, with
3857 -1.0 left and right channels will be swapped.
3860 Enable clipping. By default is enabled.
3863 @subsection Commands
3865 This filter supports the all above options as @ref{commands}.
3867 @section firequalizer
3868 Apply FIR Equalization using arbitrary frequency response.
3870 The filter accepts the following option:
3874 Set gain curve equation (in dB). The expression can contain variables:
3877 the evaluated frequency
3881 channel number, set to 0 when multichannels evaluation is disabled
3883 channel id, see libavutil/channel_layout.h, set to the first channel id when
3884 multichannels evaluation is disabled
3888 channel_layout, see libavutil/channel_layout.h
3893 @item gain_interpolate(f)
3894 interpolate gain on frequency f based on gain_entry
3895 @item cubic_interpolate(f)
3896 same as gain_interpolate, but smoother
3898 This option is also available as command. Default is @code{gain_interpolate(f)}.
3901 Set gain entry for gain_interpolate function. The expression can
3905 store gain entry at frequency f with value g
3907 This option is also available as command.
3910 Set filter delay in seconds. Higher value means more accurate.
3911 Default is @code{0.01}.
3914 Set filter accuracy in Hz. Lower value means more accurate.
3915 Default is @code{5}.
3918 Set window function. Acceptable values are:
3921 rectangular window, useful when gain curve is already smooth
3923 hann window (default)
3929 3-terms continuous 1st derivative nuttall window
3931 minimum 3-terms discontinuous nuttall window
3933 4-terms continuous 1st derivative nuttall window
3935 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3937 blackman-harris window
3943 If enabled, use fixed number of audio samples. This improves speed when
3944 filtering with large delay. Default is disabled.
3947 Enable multichannels evaluation on gain. Default is disabled.
3950 Enable zero phase mode by subtracting timestamp to compensate delay.
3951 Default is disabled.
3954 Set scale used by gain. Acceptable values are:
3957 linear frequency, linear gain
3959 linear frequency, logarithmic (in dB) gain (default)
3961 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3963 logarithmic frequency, logarithmic gain
3967 Set file for dumping, suitable for gnuplot.
3970 Set scale for dumpfile. Acceptable values are same with scale option.
3974 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3975 Default is disabled.
3978 Enable minimum phase impulse response. Default is disabled.
3981 @subsection Examples
3986 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3989 lowpass at 1000 Hz with gain_entry:
3991 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3994 custom equalization:
3996 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3999 higher delay with zero phase to compensate delay:
4001 firequalizer=delay=0.1:fixed=on:zero_phase=on
4004 lowpass on left channel, highpass on right channel:
4006 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4007 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4012 Apply a flanging effect to the audio.
4014 The filter accepts the following options:
4018 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4021 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4024 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4028 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4029 Default value is 71.
4032 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4035 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4036 Default value is @var{sinusoidal}.
4039 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4040 Default value is 25.
4043 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4044 Default is @var{linear}.
4048 Apply Haas effect to audio.
4050 Note that this makes most sense to apply on mono signals.
4051 With this filter applied to mono signals it give some directionality and
4052 stretches its stereo image.
4054 The filter accepts the following options:
4058 Set input level. By default is @var{1}, or 0dB
4061 Set output level. By default is @var{1}, or 0dB.
4064 Set gain applied to side part of signal. By default is @var{1}.
4067 Set kind of middle source. Can be one of the following:
4077 Pick middle part signal of stereo image.
4080 Pick side part signal of stereo image.
4084 Change middle phase. By default is disabled.
4087 Set left channel delay. By default is @var{2.05} milliseconds.
4090 Set left channel balance. By default is @var{-1}.
4093 Set left channel gain. By default is @var{1}.
4096 Change left phase. By default is disabled.
4099 Set right channel delay. By defaults is @var{2.12} milliseconds.
4102 Set right channel balance. By default is @var{1}.
4105 Set right channel gain. By default is @var{1}.
4108 Change right phase. By default is enabled.
4113 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4114 embedded HDCD codes is expanded into a 20-bit PCM stream.
4116 The filter supports the Peak Extend and Low-level Gain Adjustment features
4117 of HDCD, and detects the Transient Filter flag.
4120 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4123 When using the filter with wav, note the default encoding for wav is 16-bit,
4124 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4125 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4127 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4128 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4131 The filter accepts the following options:
4134 @item disable_autoconvert
4135 Disable any automatic format conversion or resampling in the filter graph.
4137 @item process_stereo
4138 Process the stereo channels together. If target_gain does not match between
4139 channels, consider it invalid and use the last valid target_gain.
4142 Set the code detect timer period in ms.
4145 Always extend peaks above -3dBFS even if PE isn't signaled.
4148 Replace audio with a solid tone and adjust the amplitude to signal some
4149 specific aspect of the decoding process. The output file can be loaded in
4150 an audio editor alongside the original to aid analysis.
4152 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4159 Gain adjustment level at each sample
4161 Samples where peak extend occurs
4163 Samples where the code detect timer is active
4165 Samples where the target gain does not match between channels
4171 Apply head-related transfer functions (HRTFs) to create virtual
4172 loudspeakers around the user for binaural listening via headphones.
4173 The HRIRs are provided via additional streams, for each channel
4174 one stereo input stream is needed.
4176 The filter accepts the following options:
4180 Set mapping of input streams for convolution.
4181 The argument is a '|'-separated list of channel names in order as they
4182 are given as additional stream inputs for filter.
4183 This also specify number of input streams. Number of input streams
4184 must be not less than number of channels in first stream plus one.
4187 Set gain applied to audio. Value is in dB. Default is 0.
4190 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4191 processing audio in time domain which is slow.
4192 @var{freq} is processing audio in frequency domain which is fast.
4193 Default is @var{freq}.
4196 Set custom gain for LFE channels. Value is in dB. Default is 0.
4199 Set size of frame in number of samples which will be processed at once.
4200 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4203 Set format of hrir stream.
4204 Default value is @var{stereo}. Alternative value is @var{multich}.
4205 If value is set to @var{stereo}, number of additional streams should
4206 be greater or equal to number of input channels in first input stream.
4207 Also each additional stream should have stereo number of channels.
4208 If value is set to @var{multich}, number of additional streams should
4209 be exactly one. Also number of input channels of additional stream
4210 should be equal or greater than twice number of channels of first input
4214 @subsection Examples
4218 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4219 each amovie filter use stereo file with IR coefficients as input.
4220 The files give coefficients for each position of virtual loudspeaker:
4223 -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"
4228 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4229 but now in @var{multich} @var{hrir} format.
4231 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"
4238 Apply a high-pass filter with 3dB point frequency.
4239 The filter can be either single-pole, or double-pole (the default).
4240 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4242 The filter accepts the following options:
4246 Set frequency in Hz. Default is 3000.
4249 Set number of poles. Default is 2.
4252 Set method to specify band-width of filter.
4267 Specify the band-width of a filter in width_type units.
4268 Applies only to double-pole filter.
4269 The default is 0.707q and gives a Butterworth response.
4272 How much to use filtered signal in output. Default is 1.
4273 Range is between 0 and 1.
4276 Specify which channels to filter, by default all available are filtered.
4279 Normalize biquad coefficients, by default is disabled.
4280 Enabling it will normalize magnitude response at DC to 0dB.
4283 Set transform type of IIR filter.
4292 @subsection Commands
4294 This filter supports the following commands:
4297 Change highpass frequency.
4298 Syntax for the command is : "@var{frequency}"
4301 Change highpass width_type.
4302 Syntax for the command is : "@var{width_type}"
4305 Change highpass width.
4306 Syntax for the command is : "@var{width}"
4309 Change highpass mix.
4310 Syntax for the command is : "@var{mix}"
4315 Join multiple input streams into one multi-channel stream.
4317 It accepts the following parameters:
4321 The number of input streams. It defaults to 2.
4323 @item channel_layout
4324 The desired output channel layout. It defaults to stereo.
4327 Map channels from inputs to output. The argument is a '|'-separated list of
4328 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4329 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4330 can be either the name of the input channel (e.g. FL for front left) or its
4331 index in the specified input stream. @var{out_channel} is the name of the output
4335 The filter will attempt to guess the mappings when they are not specified
4336 explicitly. It does so by first trying to find an unused matching input channel
4337 and if that fails it picks the first unused input channel.
4339 Join 3 inputs (with properly set channel layouts):
4341 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4344 Build a 5.1 output from 6 single-channel streams:
4346 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4347 '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'
4353 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4355 To enable compilation of this filter you need to configure FFmpeg with
4356 @code{--enable-ladspa}.
4360 Specifies the name of LADSPA plugin library to load. If the environment
4361 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4362 each one of the directories specified by the colon separated list in
4363 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4364 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4365 @file{/usr/lib/ladspa/}.
4368 Specifies the plugin within the library. Some libraries contain only
4369 one plugin, but others contain many of them. If this is not set filter
4370 will list all available plugins within the specified library.
4373 Set the '|' separated list of controls which are zero or more floating point
4374 values that determine the behavior of the loaded plugin (for example delay,
4376 Controls need to be defined using the following syntax:
4377 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4378 @var{valuei} is the value set on the @var{i}-th control.
4379 Alternatively they can be also defined using the following syntax:
4380 @var{value0}|@var{value1}|@var{value2}|..., where
4381 @var{valuei} is the value set on the @var{i}-th control.
4382 If @option{controls} is set to @code{help}, all available controls and
4383 their valid ranges are printed.
4385 @item sample_rate, s
4386 Specify the sample rate, default to 44100. Only used if plugin have
4390 Set the number of samples per channel per each output frame, default
4391 is 1024. Only used if plugin have zero inputs.
4394 Set the minimum duration of the sourced audio. See
4395 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4396 for the accepted syntax.
4397 Note that the resulting duration may be greater than the specified duration,
4398 as the generated audio is always cut at the end of a complete frame.
4399 If not specified, or the expressed duration is negative, the audio is
4400 supposed to be generated forever.
4401 Only used if plugin have zero inputs.
4404 Enable latency compensation, by default is disabled.
4405 Only used if plugin have inputs.
4408 @subsection Examples
4412 List all available plugins within amp (LADSPA example plugin) library:
4418 List all available controls and their valid ranges for @code{vcf_notch}
4419 plugin from @code{VCF} library:
4421 ladspa=f=vcf:p=vcf_notch:c=help
4425 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4428 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4432 Add reverberation to the audio using TAP-plugins
4433 (Tom's Audio Processing plugins):
4435 ladspa=file=tap_reverb:tap_reverb
4439 Generate white noise, with 0.2 amplitude:
4441 ladspa=file=cmt:noise_source_white:c=c0=.2
4445 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4446 @code{C* Audio Plugin Suite} (CAPS) library:
4448 ladspa=file=caps:Click:c=c1=20'
4452 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4454 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4458 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4459 @code{SWH Plugins} collection:
4461 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4465 Attenuate low frequencies using Multiband EQ from Steve Harris
4466 @code{SWH Plugins} collection:
4468 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4472 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4475 ladspa=caps:Narrower
4479 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4481 ladspa=caps:White:.2
4485 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4487 ladspa=caps:Fractal:c=c1=1
4491 Dynamic volume normalization using @code{VLevel} plugin:
4493 ladspa=vlevel-ladspa:vlevel_mono
4497 @subsection Commands
4499 This filter supports the following commands:
4502 Modify the @var{N}-th control value.
4504 If the specified value is not valid, it is ignored and prior one is kept.
4509 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4510 Support for both single pass (livestreams, files) and double pass (files) modes.
4511 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4512 detect true peaks, the audio stream will be upsampled to 192 kHz.
4513 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4515 The filter accepts the following options:
4519 Set integrated loudness target.
4520 Range is -70.0 - -5.0. Default value is -24.0.
4523 Set loudness range target.
4524 Range is 1.0 - 20.0. Default value is 7.0.
4527 Set maximum true peak.
4528 Range is -9.0 - +0.0. Default value is -2.0.
4530 @item measured_I, measured_i
4531 Measured IL of input file.
4532 Range is -99.0 - +0.0.
4534 @item measured_LRA, measured_lra
4535 Measured LRA of input file.
4536 Range is 0.0 - 99.0.
4538 @item measured_TP, measured_tp
4539 Measured true peak of input file.
4540 Range is -99.0 - +99.0.
4542 @item measured_thresh
4543 Measured threshold of input file.
4544 Range is -99.0 - +0.0.
4547 Set offset gain. Gain is applied before the true-peak limiter.
4548 Range is -99.0 - +99.0. Default is +0.0.
4551 Normalize by linearly scaling the source audio.
4552 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4553 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4554 be lower than source LRA and the change in integrated loudness shouldn't
4555 result in a true peak which exceeds the target TP. If any of these
4556 conditions aren't met, normalization mode will revert to @var{dynamic}.
4557 Options are @code{true} or @code{false}. Default is @code{true}.
4560 Treat mono input files as "dual-mono". If a mono file is intended for playback
4561 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4562 If set to @code{true}, this option will compensate for this effect.
4563 Multi-channel input files are not affected by this option.
4564 Options are true or false. Default is false.
4567 Set print format for stats. Options are summary, json, or none.
4568 Default value is none.
4573 Apply a low-pass filter with 3dB point frequency.
4574 The filter can be either single-pole or double-pole (the default).
4575 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4577 The filter accepts the following options:
4581 Set frequency in Hz. Default is 500.
4584 Set number of poles. Default is 2.
4587 Set method to specify band-width of filter.
4602 Specify the band-width of a filter in width_type units.
4603 Applies only to double-pole filter.
4604 The default is 0.707q and gives a Butterworth response.
4607 How much to use filtered signal in output. Default is 1.
4608 Range is between 0 and 1.
4611 Specify which channels to filter, by default all available are filtered.
4614 Normalize biquad coefficients, by default is disabled.
4615 Enabling it will normalize magnitude response at DC to 0dB.
4618 Set transform type of IIR filter.
4627 @subsection Examples
4630 Lowpass only LFE channel, it LFE is not present it does nothing:
4636 @subsection Commands
4638 This filter supports the following commands:
4641 Change lowpass frequency.
4642 Syntax for the command is : "@var{frequency}"
4645 Change lowpass width_type.
4646 Syntax for the command is : "@var{width_type}"
4649 Change lowpass width.
4650 Syntax for the command is : "@var{width}"
4654 Syntax for the command is : "@var{mix}"
4659 Load a LV2 (LADSPA Version 2) plugin.
4661 To enable compilation of this filter you need to configure FFmpeg with
4662 @code{--enable-lv2}.
4666 Specifies the plugin URI. You may need to escape ':'.
4669 Set the '|' separated list of controls which are zero or more floating point
4670 values that determine the behavior of the loaded plugin (for example delay,
4672 If @option{controls} is set to @code{help}, all available controls and
4673 their valid ranges are printed.
4675 @item sample_rate, s
4676 Specify the sample rate, default to 44100. Only used if plugin have
4680 Set the number of samples per channel per each output frame, default
4681 is 1024. Only used if plugin have zero inputs.
4684 Set the minimum duration of the sourced audio. See
4685 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4686 for the accepted syntax.
4687 Note that the resulting duration may be greater than the specified duration,
4688 as the generated audio is always cut at the end of a complete frame.
4689 If not specified, or the expressed duration is negative, the audio is
4690 supposed to be generated forever.
4691 Only used if plugin have zero inputs.
4694 @subsection Examples
4698 Apply bass enhancer plugin from Calf:
4700 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4704 Apply vinyl plugin from Calf:
4706 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4710 Apply bit crusher plugin from ArtyFX:
4712 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4717 Multiband Compress or expand the audio's dynamic range.
4719 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4720 This is akin to the crossover of a loudspeaker, and results in flat frequency
4721 response when absent compander action.
4723 It accepts the following parameters:
4727 This option syntax is:
4728 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4729 For explanation of each item refer to compand filter documentation.
4735 Mix channels with specific gain levels. The filter accepts the output
4736 channel layout followed by a set of channels definitions.
4738 This filter is also designed to efficiently remap the channels of an audio
4741 The filter accepts parameters of the form:
4742 "@var{l}|@var{outdef}|@var{outdef}|..."
4746 output channel layout or number of channels
4749 output channel specification, of the form:
4750 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4753 output channel to define, either a channel name (FL, FR, etc.) or a channel
4754 number (c0, c1, etc.)
4757 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4760 input channel to use, see out_name for details; it is not possible to mix
4761 named and numbered input channels
4764 If the `=' in a channel specification is replaced by `<', then the gains for
4765 that specification will be renormalized so that the total is 1, thus
4766 avoiding clipping noise.
4768 @subsection Mixing examples
4770 For example, if you want to down-mix from stereo to mono, but with a bigger
4771 factor for the left channel:
4773 pan=1c|c0=0.9*c0+0.1*c1
4776 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4777 7-channels surround:
4779 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4782 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4783 that should be preferred (see "-ac" option) unless you have very specific
4786 @subsection Remapping examples
4788 The channel remapping will be effective if, and only if:
4791 @item gain coefficients are zeroes or ones,
4792 @item only one input per channel output,
4795 If all these conditions are satisfied, the filter will notify the user ("Pure
4796 channel mapping detected"), and use an optimized and lossless method to do the
4799 For example, if you have a 5.1 source and want a stereo audio stream by
4800 dropping the extra channels:
4802 pan="stereo| c0=FL | c1=FR"
4805 Given the same source, you can also switch front left and front right channels
4806 and keep the input channel layout:
4808 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4811 If the input is a stereo audio stream, you can mute the front left channel (and
4812 still keep the stereo channel layout) with:
4817 Still with a stereo audio stream input, you can copy the right channel in both
4818 front left and right:
4820 pan="stereo| c0=FR | c1=FR"
4825 ReplayGain scanner filter. This filter takes an audio stream as an input and
4826 outputs it unchanged.
4827 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4831 Convert the audio sample format, sample rate and channel layout. It is
4832 not meant to be used directly.
4835 Apply time-stretching and pitch-shifting with librubberband.
4837 To enable compilation of this filter, you need to configure FFmpeg with
4838 @code{--enable-librubberband}.
4840 The filter accepts the following options:
4844 Set tempo scale factor.
4847 Set pitch scale factor.
4850 Set transients detector.
4851 Possible values are:
4860 Possible values are:
4869 Possible values are:
4876 Set processing window size.
4877 Possible values are:
4886 Possible values are:
4893 Enable formant preservation when shift pitching.
4894 Possible values are:
4902 Possible values are:
4911 Possible values are:
4918 @subsection Commands
4920 This filter supports the following commands:
4923 Change filter tempo scale factor.
4924 Syntax for the command is : "@var{tempo}"
4927 Change filter pitch scale factor.
4928 Syntax for the command is : "@var{pitch}"
4931 @section sidechaincompress
4933 This filter acts like normal compressor but has the ability to compress
4934 detected signal using second input signal.
4935 It needs two input streams and returns one output stream.
4936 First input stream will be processed depending on second stream signal.
4937 The filtered signal then can be filtered with other filters in later stages of
4938 processing. See @ref{pan} and @ref{amerge} filter.
4940 The filter accepts the following options:
4944 Set input gain. Default is 1. Range is between 0.015625 and 64.
4947 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4948 Default is @code{downward}.
4951 If a signal of second stream raises above this level it will affect the gain
4952 reduction of first stream.
4953 By default is 0.125. Range is between 0.00097563 and 1.
4956 Set a ratio about which the signal is reduced. 1:2 means that if the level
4957 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4958 Default is 2. Range is between 1 and 20.
4961 Amount of milliseconds the signal has to rise above the threshold before gain
4962 reduction starts. Default is 20. Range is between 0.01 and 2000.
4965 Amount of milliseconds the signal has to fall below the threshold before
4966 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4969 Set the amount by how much signal will be amplified after processing.
4970 Default is 1. Range is from 1 to 64.
4973 Curve the sharp knee around the threshold to enter gain reduction more softly.
4974 Default is 2.82843. Range is between 1 and 8.
4977 Choose if the @code{average} level between all channels of side-chain stream
4978 or the louder(@code{maximum}) channel of side-chain stream affects the
4979 reduction. Default is @code{average}.
4982 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4983 of @code{rms}. Default is @code{rms} which is mainly smoother.
4986 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4989 How much to use compressed signal in output. Default is 1.
4990 Range is between 0 and 1.
4993 @subsection Commands
4995 This filter supports the all above options as @ref{commands}.
4997 @subsection Examples
5001 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5002 depending on the signal of 2nd input and later compressed signal to be
5003 merged with 2nd input:
5005 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5009 @section sidechaingate
5011 A sidechain gate acts like a normal (wideband) gate but has the ability to
5012 filter the detected signal before sending it to the gain reduction stage.
5013 Normally a gate uses the full range signal to detect a level above the
5015 For example: If you cut all lower frequencies from your sidechain signal
5016 the gate will decrease the volume of your track only if not enough highs
5017 appear. With this technique you are able to reduce the resonation of a
5018 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5020 It needs two input streams and returns one output stream.
5021 First input stream will be processed depending on second stream signal.
5023 The filter accepts the following options:
5027 Set input level before filtering.
5028 Default is 1. Allowed range is from 0.015625 to 64.
5031 Set the mode of operation. Can be @code{upward} or @code{downward}.
5032 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5033 will be amplified, expanding dynamic range in upward direction.
5034 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5037 Set the level of gain reduction when the signal is below the threshold.
5038 Default is 0.06125. Allowed range is from 0 to 1.
5039 Setting this to 0 disables reduction and then filter behaves like expander.
5042 If a signal rises above this level the gain reduction is released.
5043 Default is 0.125. Allowed range is from 0 to 1.
5046 Set a ratio about which the signal is reduced.
5047 Default is 2. Allowed range is from 1 to 9000.
5050 Amount of milliseconds the signal has to rise above the threshold before gain
5052 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5055 Amount of milliseconds the signal has to fall below the threshold before the
5056 reduction is increased again. Default is 250 milliseconds.
5057 Allowed range is from 0.01 to 9000.
5060 Set amount of amplification of signal after processing.
5061 Default is 1. Allowed range is from 1 to 64.
5064 Curve the sharp knee around the threshold to enter gain reduction more softly.
5065 Default is 2.828427125. Allowed range is from 1 to 8.
5068 Choose if exact signal should be taken for detection or an RMS like one.
5069 Default is rms. Can be peak or rms.
5072 Choose if the average level between all channels or the louder channel affects
5074 Default is average. Can be average or maximum.
5077 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5080 @section silencedetect
5082 Detect silence in an audio stream.
5084 This filter logs a message when it detects that the input audio volume is less
5085 or equal to a noise tolerance value for a duration greater or equal to the
5086 minimum detected noise duration.
5088 The printed times and duration are expressed in seconds. The
5089 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5090 is set on the first frame whose timestamp equals or exceeds the detection
5091 duration and it contains the timestamp of the first frame of the silence.
5093 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5094 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5095 keys are set on the first frame after the silence. If @option{mono} is
5096 enabled, and each channel is evaluated separately, the @code{.X}
5097 suffixed keys are used, and @code{X} corresponds to the channel number.
5099 The filter accepts the following options:
5103 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5104 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5107 Set silence duration until notification (default is 2 seconds). See
5108 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5109 for the accepted syntax.
5112 Process each channel separately, instead of combined. By default is disabled.
5115 @subsection Examples
5119 Detect 5 seconds of silence with -50dB noise tolerance:
5121 silencedetect=n=-50dB:d=5
5125 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5126 tolerance in @file{silence.mp3}:
5128 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5132 @section silenceremove
5134 Remove silence from the beginning, middle or end of the audio.
5136 The filter accepts the following options:
5140 This value is used to indicate if audio should be trimmed at beginning of
5141 the audio. A value of zero indicates no silence should be trimmed from the
5142 beginning. When specifying a non-zero value, it trims audio up until it
5143 finds non-silence. Normally, when trimming silence from beginning of audio
5144 the @var{start_periods} will be @code{1} but it can be increased to higher
5145 values to trim all audio up to specific count of non-silence periods.
5146 Default value is @code{0}.
5148 @item start_duration
5149 Specify the amount of time that non-silence must be detected before it stops
5150 trimming audio. By increasing the duration, bursts of noises can be treated
5151 as silence and trimmed off. Default value is @code{0}.
5153 @item start_threshold
5154 This indicates what sample value should be treated as silence. For digital
5155 audio, a value of @code{0} may be fine but for audio recorded from analog,
5156 you may wish to increase the value to account for background noise.
5157 Can be specified in dB (in case "dB" is appended to the specified value)
5158 or amplitude ratio. Default value is @code{0}.
5161 Specify max duration of silence at beginning that will be kept after
5162 trimming. Default is 0, which is equal to trimming all samples detected
5166 Specify mode of detection of silence end in start of multi-channel audio.
5167 Can be @var{any} or @var{all}. Default is @var{any}.
5168 With @var{any}, any sample that is detected as non-silence will cause
5169 stopped trimming of silence.
5170 With @var{all}, only if all channels are detected as non-silence will cause
5171 stopped trimming of silence.
5174 Set the count for trimming silence from the end of audio.
5175 To remove silence from the middle of a file, specify a @var{stop_periods}
5176 that is negative. This value is then treated as a positive value and is
5177 used to indicate the effect should restart processing as specified by
5178 @var{start_periods}, making it suitable for removing periods of silence
5179 in the middle of the audio.
5180 Default value is @code{0}.
5183 Specify a duration of silence that must exist before audio is not copied any
5184 more. By specifying a higher duration, silence that is wanted can be left in
5186 Default value is @code{0}.
5188 @item stop_threshold
5189 This is the same as @option{start_threshold} but for trimming silence from
5191 Can be specified in dB (in case "dB" is appended to the specified value)
5192 or amplitude ratio. Default value is @code{0}.
5195 Specify max duration of silence at end that will be kept after
5196 trimming. Default is 0, which is equal to trimming all samples detected
5200 Specify mode of detection of silence start in end of multi-channel audio.
5201 Can be @var{any} or @var{all}. Default is @var{any}.
5202 With @var{any}, any sample that is detected as non-silence will cause
5203 stopped trimming of silence.
5204 With @var{all}, only if all channels are detected as non-silence will cause
5205 stopped trimming of silence.
5208 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5209 and works better with digital silence which is exactly 0.
5210 Default value is @code{rms}.
5213 Set duration in number of seconds used to calculate size of window in number
5214 of samples for detecting silence.
5215 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5218 @subsection Examples
5222 The following example shows how this filter can be used to start a recording
5223 that does not contain the delay at the start which usually occurs between
5224 pressing the record button and the start of the performance:
5226 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5230 Trim all silence encountered from beginning to end where there is more than 1
5231 second of silence in audio:
5233 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5237 Trim all digital silence samples, using peak detection, from beginning to end
5238 where there is more than 0 samples of digital silence in audio and digital
5239 silence is detected in all channels at same positions in stream:
5241 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5247 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5248 loudspeakers around the user for binaural listening via headphones (audio
5249 formats up to 9 channels supported).
5250 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5251 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5252 Austrian Academy of Sciences.
5254 To enable compilation of this filter you need to configure FFmpeg with
5255 @code{--enable-libmysofa}.
5257 The filter accepts the following options:
5261 Set the SOFA file used for rendering.
5264 Set gain applied to audio. Value is in dB. Default is 0.
5267 Set rotation of virtual loudspeakers in deg. Default is 0.
5270 Set elevation of virtual speakers in deg. Default is 0.
5273 Set distance in meters between loudspeakers and the listener with near-field
5274 HRTFs. Default is 1.
5277 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5278 processing audio in time domain which is slow.
5279 @var{freq} is processing audio in frequency domain which is fast.
5280 Default is @var{freq}.
5283 Set custom positions of virtual loudspeakers. Syntax for this option is:
5284 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5285 Each virtual loudspeaker is described with short channel name following with
5286 azimuth and elevation in degrees.
5287 Each virtual loudspeaker description is separated by '|'.
5288 For example to override front left and front right channel positions use:
5289 'speakers=FL 45 15|FR 345 15'.
5290 Descriptions with unrecognised channel names are ignored.
5293 Set custom gain for LFE channels. Value is in dB. Default is 0.
5296 Set custom frame size in number of samples. Default is 1024.
5297 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5298 is set to @var{freq}.
5301 Should all IRs be normalized upon importing SOFA file.
5302 By default is enabled.
5305 Should nearest IRs be interpolated with neighbor IRs if exact position
5306 does not match. By default is disabled.
5309 Minphase all IRs upon loading of SOFA file. By default is disabled.
5312 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5315 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5318 @subsection Examples
5322 Using ClubFritz6 sofa file:
5324 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5328 Using ClubFritz12 sofa file and bigger radius with small rotation:
5330 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5334 Similar as above but with custom speaker positions for front left, front right, back left and back right
5335 and also with custom gain:
5337 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5344 This filter expands or compresses each half-cycle of audio samples
5345 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5346 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5348 The filter accepts the following options:
5352 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5353 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5356 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5357 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5358 would be such that local peak value reaches target peak value but never to surpass it and that
5359 ratio between new and previous peak value does not surpass this option value.
5361 @item compression, c
5362 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5363 This option controls maximum local half-cycle of samples compression. This option is used
5364 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5365 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5366 that peak's half-cycle will be compressed by current compression factor.
5369 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5370 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5371 Any half-cycle samples with their local peak value below or same as this option value will be
5372 compressed by current compression factor, otherwise, if greater than threshold value they will be
5373 expanded with expansion factor so that it could reach peak target value but never surpass it.
5376 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5377 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5378 each new half-cycle until it reaches @option{expansion} value.
5379 Setting this options too high may lead to distortions.
5382 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5383 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5384 each new half-cycle until it reaches @option{compression} value.
5387 Specify which channels to filter, by default all available channels are filtered.
5390 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5391 option. When enabled any half-cycle of samples with their local peak value below or same as
5392 @option{threshold} option will be expanded otherwise it will be compressed.
5395 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5396 When disabled each filtered channel gain calculation is independent, otherwise when this option
5397 is enabled the minimum of all possible gains for each filtered channel is used.
5400 @subsection Commands
5402 This filter supports the all above options as @ref{commands}.
5404 @section stereotools
5406 This filter has some handy utilities to manage stereo signals, for converting
5407 M/S stereo recordings to L/R signal while having control over the parameters
5408 or spreading the stereo image of master track.
5410 The filter accepts the following options:
5414 Set input level before filtering for both channels. Defaults is 1.
5415 Allowed range is from 0.015625 to 64.
5418 Set output level after filtering for both channels. Defaults is 1.
5419 Allowed range is from 0.015625 to 64.
5422 Set input balance between both channels. Default is 0.
5423 Allowed range is from -1 to 1.
5426 Set output balance between both channels. Default is 0.
5427 Allowed range is from -1 to 1.
5430 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5431 clipping. Disabled by default.
5434 Mute the left channel. Disabled by default.
5437 Mute the right channel. Disabled by default.
5440 Change the phase of the left channel. Disabled by default.
5443 Change the phase of the right channel. Disabled by default.
5446 Set stereo mode. Available values are:
5450 Left/Right to Left/Right, this is default.
5453 Left/Right to Mid/Side.
5456 Mid/Side to Left/Right.
5459 Left/Right to Left/Left.
5462 Left/Right to Right/Right.
5465 Left/Right to Left + Right.
5468 Left/Right to Right/Left.
5471 Mid/Side to Left/Left.
5474 Mid/Side to Right/Right.
5478 Set level of side signal. Default is 1.
5479 Allowed range is from 0.015625 to 64.
5482 Set balance of side signal. Default is 0.
5483 Allowed range is from -1 to 1.
5486 Set level of the middle signal. Default is 1.
5487 Allowed range is from 0.015625 to 64.
5490 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5493 Set stereo base between mono and inversed channels. Default is 0.
5494 Allowed range is from -1 to 1.
5497 Set delay in milliseconds how much to delay left from right channel and
5498 vice versa. Default is 0. Allowed range is from -20 to 20.
5501 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5504 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5506 @item bmode_in, bmode_out
5507 Set balance mode for balance_in/balance_out option.
5509 Can be one of the following:
5513 Classic balance mode. Attenuate one channel at time.
5514 Gain is raised up to 1.
5517 Similar as classic mode above but gain is raised up to 2.
5520 Equal power distribution, from -6dB to +6dB range.
5524 @subsection Examples
5528 Apply karaoke like effect:
5530 stereotools=mlev=0.015625
5534 Convert M/S signal to L/R:
5536 "stereotools=mode=ms>lr"
5540 @section stereowiden
5542 This filter enhance the stereo effect by suppressing signal common to both
5543 channels and by delaying the signal of left into right and vice versa,
5544 thereby widening the stereo effect.
5546 The filter accepts the following options:
5550 Time in milliseconds of the delay of left signal into right and vice versa.
5551 Default is 20 milliseconds.
5554 Amount of gain in delayed signal into right and vice versa. Gives a delay
5555 effect of left signal in right output and vice versa which gives widening
5556 effect. Default is 0.3.
5559 Cross feed of left into right with inverted phase. This helps in suppressing
5560 the mono. If the value is 1 it will cancel all the signal common to both
5561 channels. Default is 0.3.
5564 Set level of input signal of original channel. Default is 0.8.
5567 @subsection Commands
5569 This filter supports the all above options except @code{delay} as @ref{commands}.
5571 @section superequalizer
5572 Apply 18 band equalizer.
5574 The filter accepts the following options:
5581 Set 131Hz band gain.
5583 Set 185Hz band gain.
5585 Set 262Hz band gain.
5587 Set 370Hz band gain.
5589 Set 523Hz band gain.
5591 Set 740Hz band gain.
5593 Set 1047Hz band gain.
5595 Set 1480Hz band gain.
5597 Set 2093Hz band gain.
5599 Set 2960Hz band gain.
5601 Set 4186Hz band gain.
5603 Set 5920Hz band gain.
5605 Set 8372Hz band gain.
5607 Set 11840Hz band gain.
5609 Set 16744Hz band gain.
5611 Set 20000Hz band gain.
5615 Apply audio surround upmix filter.
5617 This filter allows to produce multichannel output from audio stream.
5619 The filter accepts the following options:
5623 Set output channel layout. By default, this is @var{5.1}.
5625 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5626 for the required syntax.
5629 Set input channel layout. By default, this is @var{stereo}.
5631 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5632 for the required syntax.
5635 Set input volume level. By default, this is @var{1}.
5638 Set output volume level. By default, this is @var{1}.
5641 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5644 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5647 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5650 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5651 In @var{add} mode, LFE channel is created from input audio and added to output.
5652 In @var{sub} mode, LFE channel is created from input audio and added to output but
5653 also all non-LFE output channels are subtracted with output LFE channel.
5656 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5657 Default is @var{90}.
5660 Set front center input volume. By default, this is @var{1}.
5663 Set front center output volume. By default, this is @var{1}.
5666 Set front left input volume. By default, this is @var{1}.
5669 Set front left output volume. By default, this is @var{1}.
5672 Set front right input volume. By default, this is @var{1}.
5675 Set front right output volume. By default, this is @var{1}.
5678 Set side left input volume. By default, this is @var{1}.
5681 Set side left output volume. By default, this is @var{1}.
5684 Set side right input volume. By default, this is @var{1}.
5687 Set side right output volume. By default, this is @var{1}.
5690 Set back left input volume. By default, this is @var{1}.
5693 Set back left output volume. By default, this is @var{1}.
5696 Set back right input volume. By default, this is @var{1}.
5699 Set back right output volume. By default, this is @var{1}.
5702 Set back center input volume. By default, this is @var{1}.
5705 Set back center output volume. By default, this is @var{1}.
5708 Set LFE input volume. By default, this is @var{1}.
5711 Set LFE output volume. By default, this is @var{1}.
5714 Set spread usage of stereo image across X axis for all channels.
5717 Set spread usage of stereo image across Y axis for all channels.
5719 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5720 Set spread usage of stereo image across X axis for each channel.
5722 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5723 Set spread usage of stereo image across Y axis for each channel.
5726 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5729 Set window function.
5731 It accepts the following values:
5754 Default is @code{hann}.
5757 Set window overlap. If set to 1, the recommended overlap for selected
5758 window function will be picked. Default is @code{0.5}.
5761 @section treble, highshelf
5763 Boost or cut treble (upper) frequencies of the audio using a two-pole
5764 shelving filter with a response similar to that of a standard
5765 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5767 The filter accepts the following options:
5771 Give the gain at whichever is the lower of ~22 kHz and the
5772 Nyquist frequency. Its useful range is about -20 (for a large cut)
5773 to +20 (for a large boost). Beware of clipping when using a positive gain.
5776 Set the filter's central frequency and so can be used
5777 to extend or reduce the frequency range to be boosted or cut.
5778 The default value is @code{3000} Hz.
5781 Set method to specify band-width of filter.
5796 Determine how steep is the filter's shelf transition.
5799 How much to use filtered signal in output. Default is 1.
5800 Range is between 0 and 1.
5803 Specify which channels to filter, by default all available are filtered.
5806 Normalize biquad coefficients, by default is disabled.
5807 Enabling it will normalize magnitude response at DC to 0dB.
5810 Set transform type of IIR filter.
5819 @subsection Commands
5821 This filter supports the following commands:
5824 Change treble frequency.
5825 Syntax for the command is : "@var{frequency}"
5828 Change treble width_type.
5829 Syntax for the command is : "@var{width_type}"
5832 Change treble width.
5833 Syntax for the command is : "@var{width}"
5837 Syntax for the command is : "@var{gain}"
5841 Syntax for the command is : "@var{mix}"
5846 Sinusoidal amplitude modulation.
5848 The filter accepts the following options:
5852 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5853 (20 Hz or lower) will result in a tremolo effect.
5854 This filter may also be used as a ring modulator by specifying
5855 a modulation frequency higher than 20 Hz.
5856 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5859 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5860 Default value is 0.5.
5865 Sinusoidal phase modulation.
5867 The filter accepts the following options:
5871 Modulation frequency in Hertz.
5872 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5875 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5876 Default value is 0.5.
5881 Adjust the input audio volume.
5883 It accepts the following parameters:
5887 Set audio volume expression.
5889 Output values are clipped to the maximum value.
5891 The output audio volume is given by the relation:
5893 @var{output_volume} = @var{volume} * @var{input_volume}
5896 The default value for @var{volume} is "1.0".
5899 This parameter represents the mathematical precision.
5901 It determines which input sample formats will be allowed, which affects the
5902 precision of the volume scaling.
5906 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5908 32-bit floating-point; this limits input sample format to FLT. (default)
5910 64-bit floating-point; this limits input sample format to DBL.
5914 Choose the behaviour on encountering ReplayGain side data in input frames.
5918 Remove ReplayGain side data, ignoring its contents (the default).
5921 Ignore ReplayGain side data, but leave it in the frame.
5924 Prefer the track gain, if present.
5927 Prefer the album gain, if present.
5930 @item replaygain_preamp
5931 Pre-amplification gain in dB to apply to the selected replaygain gain.
5933 Default value for @var{replaygain_preamp} is 0.0.
5935 @item replaygain_noclip
5936 Prevent clipping by limiting the gain applied.
5938 Default value for @var{replaygain_noclip} is 1.
5941 Set when the volume expression is evaluated.
5943 It accepts the following values:
5946 only evaluate expression once during the filter initialization, or
5947 when the @samp{volume} command is sent
5950 evaluate expression for each incoming frame
5953 Default value is @samp{once}.
5956 The volume expression can contain the following parameters.
5960 frame number (starting at zero)
5963 @item nb_consumed_samples
5964 number of samples consumed by the filter
5966 number of samples in the current frame
5968 original frame position in the file
5974 PTS at start of stream
5976 time at start of stream
5982 last set volume value
5985 Note that when @option{eval} is set to @samp{once} only the
5986 @var{sample_rate} and @var{tb} variables are available, all other
5987 variables will evaluate to NAN.
5989 @subsection Commands
5991 This filter supports the following commands:
5994 Modify the volume expression.
5995 The command accepts the same syntax of the corresponding option.
5997 If the specified expression is not valid, it is kept at its current
6001 @subsection Examples
6005 Halve the input audio volume:
6009 volume=volume=-6.0206dB
6012 In all the above example the named key for @option{volume} can be
6013 omitted, for example like in:
6019 Increase input audio power by 6 decibels using fixed-point precision:
6021 volume=volume=6dB:precision=fixed
6025 Fade volume after time 10 with an annihilation period of 5 seconds:
6027 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6031 @section volumedetect
6033 Detect the volume of the input video.
6035 The filter has no parameters. The input is not modified. Statistics about
6036 the volume will be printed in the log when the input stream end is reached.
6038 In particular it will show the mean volume (root mean square), maximum
6039 volume (on a per-sample basis), and the beginning of a histogram of the
6040 registered volume values (from the maximum value to a cumulated 1/1000 of
6043 All volumes are in decibels relative to the maximum PCM value.
6045 @subsection Examples
6047 Here is an excerpt of the output:
6049 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6050 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6051 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6052 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6053 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6054 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6055 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6056 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6057 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6063 The mean square energy is approximately -27 dB, or 10^-2.7.
6065 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6067 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6070 In other words, raising the volume by +4 dB does not cause any clipping,
6071 raising it by +5 dB causes clipping for 6 samples, etc.
6073 @c man end AUDIO FILTERS
6075 @chapter Audio Sources
6076 @c man begin AUDIO SOURCES
6078 Below is a description of the currently available audio sources.
6082 Buffer audio frames, and make them available to the filter chain.
6084 This source is mainly intended for a programmatic use, in particular
6085 through the interface defined in @file{libavfilter/buffersrc.h}.
6087 It accepts the following parameters:
6091 The timebase which will be used for timestamps of submitted frames. It must be
6092 either a floating-point number or in @var{numerator}/@var{denominator} form.
6095 The sample rate of the incoming audio buffers.
6098 The sample format of the incoming audio buffers.
6099 Either a sample format name or its corresponding integer representation from
6100 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6102 @item channel_layout
6103 The channel layout of the incoming audio buffers.
6104 Either a channel layout name from channel_layout_map in
6105 @file{libavutil/channel_layout.c} or its corresponding integer representation
6106 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6109 The number of channels of the incoming audio buffers.
6110 If both @var{channels} and @var{channel_layout} are specified, then they
6115 @subsection Examples
6118 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6121 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6122 Since the sample format with name "s16p" corresponds to the number
6123 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6126 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6131 Generate an audio signal specified by an expression.
6133 This source accepts in input one or more expressions (one for each
6134 channel), which are evaluated and used to generate a corresponding
6137 This source accepts the following options:
6141 Set the '|'-separated expressions list for each separate channel. In case the
6142 @option{channel_layout} option is not specified, the selected channel layout
6143 depends on the number of provided expressions. Otherwise the last
6144 specified expression is applied to the remaining output channels.
6146 @item channel_layout, c
6147 Set the channel layout. The number of channels in the specified layout
6148 must be equal to the number of specified expressions.
6151 Set the minimum duration of the sourced audio. See
6152 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6153 for the accepted syntax.
6154 Note that the resulting duration may be greater than the specified
6155 duration, as the generated audio is always cut at the end of a
6158 If not specified, or the expressed duration is negative, the audio is
6159 supposed to be generated forever.
6162 Set the number of samples per channel per each output frame,
6165 @item sample_rate, s
6166 Specify the sample rate, default to 44100.
6169 Each expression in @var{exprs} can contain the following constants:
6173 number of the evaluated sample, starting from 0
6176 time of the evaluated sample expressed in seconds, starting from 0
6183 @subsection Examples
6193 Generate a sin signal with frequency of 440 Hz, set sample rate to
6196 aevalsrc="sin(440*2*PI*t):s=8000"
6200 Generate a two channels signal, specify the channel layout (Front
6201 Center + Back Center) explicitly:
6203 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6207 Generate white noise:
6209 aevalsrc="-2+random(0)"
6213 Generate an amplitude modulated signal:
6215 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6219 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6221 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6228 Generate a FIR coefficients using frequency sampling method.
6230 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6232 The filter accepts the following options:
6236 Set number of filter coefficents in output audio stream.
6237 Default value is 1025.
6240 Set frequency points from where magnitude and phase are set.
6241 This must be in non decreasing order, and first element must be 0, while last element
6242 must be 1. Elements are separated by white spaces.
6245 Set magnitude value for every frequency point set by @option{frequency}.
6246 Number of values must be same as number of frequency points.
6247 Values are separated by white spaces.
6250 Set phase value for every frequency point set by @option{frequency}.
6251 Number of values must be same as number of frequency points.
6252 Values are separated by white spaces.
6254 @item sample_rate, r
6255 Set sample rate, default is 44100.
6258 Set number of samples per each frame. Default is 1024.
6261 Set window function. Default is blackman.
6266 The null audio source, return unprocessed audio frames. It is mainly useful
6267 as a template and to be employed in analysis / debugging tools, or as
6268 the source for filters which ignore the input data (for example the sox
6271 This source accepts the following options:
6275 @item channel_layout, cl
6277 Specifies the channel layout, and can be either an integer or a string
6278 representing a channel layout. The default value of @var{channel_layout}
6281 Check the channel_layout_map definition in
6282 @file{libavutil/channel_layout.c} for the mapping between strings and
6283 channel layout values.
6285 @item sample_rate, r
6286 Specifies the sample rate, and defaults to 44100.
6289 Set the number of samples per requested frames.
6292 Set the duration of the sourced audio. See
6293 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6294 for the accepted syntax.
6296 If not specified, or the expressed duration is negative, the audio is
6297 supposed to be generated forever.
6300 @subsection Examples
6304 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6306 anullsrc=r=48000:cl=4
6310 Do the same operation with a more obvious syntax:
6312 anullsrc=r=48000:cl=mono
6316 All the parameters need to be explicitly defined.
6320 Synthesize a voice utterance using the libflite library.
6322 To enable compilation of this filter you need to configure FFmpeg with
6323 @code{--enable-libflite}.
6325 Note that versions of the flite library prior to 2.0 are not thread-safe.
6327 The filter accepts the following options:
6332 If set to 1, list the names of the available voices and exit
6333 immediately. Default value is 0.
6336 Set the maximum number of samples per frame. Default value is 512.
6339 Set the filename containing the text to speak.
6342 Set the text to speak.
6345 Set the voice to use for the speech synthesis. Default value is
6346 @code{kal}. See also the @var{list_voices} option.
6349 @subsection Examples
6353 Read from file @file{speech.txt}, and synthesize the text using the
6354 standard flite voice:
6356 flite=textfile=speech.txt
6360 Read the specified text selecting the @code{slt} voice:
6362 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6366 Input text to ffmpeg:
6368 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6372 Make @file{ffplay} speak the specified text, using @code{flite} and
6373 the @code{lavfi} device:
6375 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6379 For more information about libflite, check:
6380 @url{http://www.festvox.org/flite/}
6384 Generate a noise audio signal.
6386 The filter accepts the following options:
6389 @item sample_rate, r
6390 Specify the sample rate. Default value is 48000 Hz.
6393 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6397 Specify the duration of the generated audio stream. Not specifying this option
6398 results in noise with an infinite length.
6400 @item color, colour, c
6401 Specify the color of noise. Available noise colors are white, pink, brown,
6402 blue, violet and velvet. Default color is white.
6405 Specify a value used to seed the PRNG.
6408 Set the number of samples per each output frame, default is 1024.
6411 @subsection Examples
6416 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6418 anoisesrc=d=60:c=pink:r=44100:a=0.5
6424 Generate odd-tap Hilbert transform FIR coefficients.
6426 The resulting stream can be used with @ref{afir} filter for phase-shifting
6427 the signal by 90 degrees.
6429 This is used in many matrix coding schemes and for analytic signal generation.
6430 The process is often written as a multiplication by i (or j), the imaginary unit.
6432 The filter accepts the following options:
6436 @item sample_rate, s
6437 Set sample rate, default is 44100.
6440 Set length of FIR filter, default is 22051.
6443 Set number of samples per each frame.
6446 Set window function to be used when generating FIR coefficients.
6451 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6453 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6455 The filter accepts the following options:
6458 @item sample_rate, r
6459 Set sample rate, default is 44100.
6462 Set number of samples per each frame. Default is 1024.
6465 Set high-pass frequency. Default is 0.
6468 Set low-pass frequency. Default is 0.
6469 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6470 is higher than 0 then filter will create band-pass filter coefficients,
6471 otherwise band-reject filter coefficients.
6474 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6477 Set Kaiser window beta.
6480 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6483 Enable rounding, by default is disabled.
6486 Set number of taps for high-pass filter.
6489 Set number of taps for low-pass filter.
6494 Generate an audio signal made of a sine wave with amplitude 1/8.
6496 The audio signal is bit-exact.
6498 The filter accepts the following options:
6503 Set the carrier frequency. Default is 440 Hz.
6505 @item beep_factor, b
6506 Enable a periodic beep every second with frequency @var{beep_factor} times
6507 the carrier frequency. Default is 0, meaning the beep is disabled.
6509 @item sample_rate, r
6510 Specify the sample rate, default is 44100.
6513 Specify the duration of the generated audio stream.
6515 @item samples_per_frame
6516 Set the number of samples per output frame.
6518 The expression can contain the following constants:
6522 The (sequential) number of the output audio frame, starting from 0.
6525 The PTS (Presentation TimeStamp) of the output audio frame,
6526 expressed in @var{TB} units.
6529 The PTS of the output audio frame, expressed in seconds.
6532 The timebase of the output audio frames.
6535 Default is @code{1024}.
6538 @subsection Examples
6543 Generate a simple 440 Hz sine wave:
6549 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6553 sine=frequency=220:beep_factor=4:duration=5
6557 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6560 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6564 @c man end AUDIO SOURCES
6566 @chapter Audio Sinks
6567 @c man begin AUDIO SINKS
6569 Below is a description of the currently available audio sinks.
6571 @section abuffersink
6573 Buffer audio frames, and make them available to the end of filter chain.
6575 This sink is mainly intended for programmatic use, in particular
6576 through the interface defined in @file{libavfilter/buffersink.h}
6577 or the options system.
6579 It accepts a pointer to an AVABufferSinkContext structure, which
6580 defines the incoming buffers' formats, to be passed as the opaque
6581 parameter to @code{avfilter_init_filter} for initialization.
6584 Null audio sink; do absolutely nothing with the input audio. It is
6585 mainly useful as a template and for use in analysis / debugging
6588 @c man end AUDIO SINKS
6590 @chapter Video Filters
6591 @c man begin VIDEO FILTERS
6593 When you configure your FFmpeg build, you can disable any of the
6594 existing filters using @code{--disable-filters}.
6595 The configure output will show the video filters included in your
6598 Below is a description of the currently available video filters.
6602 Mark a region of interest in a video frame.
6604 The frame data is passed through unchanged, but metadata is attached
6605 to the frame indicating regions of interest which can affect the
6606 behaviour of later encoding. Multiple regions can be marked by
6607 applying the filter multiple times.
6611 Region distance in pixels from the left edge of the frame.
6613 Region distance in pixels from the top edge of the frame.
6615 Region width in pixels.
6617 Region height in pixels.
6619 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6620 and may contain the following variables:
6623 Width of the input frame.
6625 Height of the input frame.
6629 Quantisation offset to apply within the region.
6631 This must be a real value in the range -1 to +1. A value of zero
6632 indicates no quality change. A negative value asks for better quality
6633 (less quantisation), while a positive value asks for worse quality
6634 (greater quantisation).
6636 The range is calibrated so that the extreme values indicate the
6637 largest possible offset - if the rest of the frame is encoded with the
6638 worst possible quality, an offset of -1 indicates that this region
6639 should be encoded with the best possible quality anyway. Intermediate
6640 values are then interpolated in some codec-dependent way.
6642 For example, in 10-bit H.264 the quantisation parameter varies between
6643 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6644 this region should be encoded with a QP around one-tenth of the full
6645 range better than the rest of the frame. So, if most of the frame
6646 were to be encoded with a QP of around 30, this region would get a QP
6647 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6648 An extreme value of -1 would indicate that this region should be
6649 encoded with the best possible quality regardless of the treatment of
6650 the rest of the frame - that is, should be encoded at a QP of -12.
6652 If set to true, remove any existing regions of interest marked on the
6653 frame before adding the new one.
6656 @subsection Examples
6660 Mark the centre quarter of the frame as interesting.
6662 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6665 Mark the 100-pixel-wide region on the left edge of the frame as very
6666 uninteresting (to be encoded at much lower quality than the rest of
6669 addroi=0:0:100:ih:+1/5
6673 @section alphaextract
6675 Extract the alpha component from the input as a grayscale video. This
6676 is especially useful with the @var{alphamerge} filter.
6680 Add or replace the alpha component of the primary input with the
6681 grayscale value of a second input. This is intended for use with
6682 @var{alphaextract} to allow the transmission or storage of frame
6683 sequences that have alpha in a format that doesn't support an alpha
6686 For example, to reconstruct full frames from a normal YUV-encoded video
6687 and a separate video created with @var{alphaextract}, you might use:
6689 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6694 Amplify differences between current pixel and pixels of adjacent frames in
6695 same pixel location.
6697 This filter accepts the following options:
6701 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6702 For example radius of 3 will instruct filter to calculate average of 7 frames.
6705 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6708 Set threshold for difference amplification. Any difference greater or equal to
6709 this value will not alter source pixel. Default is 10.
6710 Allowed range is from 0 to 65535.
6713 Set tolerance for difference amplification. Any difference lower to
6714 this value will not alter source pixel. Default is 0.
6715 Allowed range is from 0 to 65535.
6718 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6719 This option controls maximum possible value that will decrease source pixel value.
6722 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6723 This option controls maximum possible value that will increase source pixel value.
6726 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6729 @subsection Commands
6731 This filter supports the following @ref{commands} that corresponds to option of same name:
6743 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6744 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6745 Substation Alpha) subtitles files.
6747 This filter accepts the following option in addition to the common options from
6748 the @ref{subtitles} filter:
6752 Set the shaping engine
6754 Available values are:
6757 The default libass shaping engine, which is the best available.
6759 Fast, font-agnostic shaper that can do only substitutions
6761 Slower shaper using OpenType for substitutions and positioning
6764 The default is @code{auto}.
6768 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6770 The filter accepts the following options:
6774 Set threshold A for 1st plane. Default is 0.02.
6775 Valid range is 0 to 0.3.
6778 Set threshold B for 1st plane. Default is 0.04.
6779 Valid range is 0 to 5.
6782 Set threshold A for 2nd plane. Default is 0.02.
6783 Valid range is 0 to 0.3.
6786 Set threshold B for 2nd plane. Default is 0.04.
6787 Valid range is 0 to 5.
6790 Set threshold A for 3rd plane. Default is 0.02.
6791 Valid range is 0 to 0.3.
6794 Set threshold B for 3rd plane. Default is 0.04.
6795 Valid range is 0 to 5.
6797 Threshold A is designed to react on abrupt changes in the input signal and
6798 threshold B is designed to react on continuous changes in the input signal.
6801 Set number of frames filter will use for averaging. Default is 9. Must be odd
6802 number in range [5, 129].
6805 Set what planes of frame filter will use for averaging. Default is all.
6808 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
6809 Alternatively can be set to @code{s} serial.
6811 Parallel can be faster then serial, while other way around is never true.
6812 Parallel will abort early on first change being greater then thresholds, while serial
6813 will continue processing other side of frames if they are equal or below thresholds.
6816 @subsection Commands
6817 This filter supports same @ref{commands} as options except option @code{s}.
6818 The command accepts the same syntax of the corresponding option.
6822 Apply average blur filter.
6824 The filter accepts the following options:
6828 Set horizontal radius size.
6831 Set which planes to filter. By default all planes are filtered.
6834 Set vertical radius size, if zero it will be same as @code{sizeX}.
6835 Default is @code{0}.
6838 @subsection Commands
6839 This filter supports same commands as options.
6840 The command accepts the same syntax of the corresponding option.
6842 If the specified expression is not valid, it is kept at its current
6847 Compute the bounding box for the non-black pixels in the input frame
6850 This filter computes the bounding box containing all the pixels with a
6851 luminance value greater than the minimum allowed value.
6852 The parameters describing the bounding box are printed on the filter
6855 The filter accepts the following option:
6859 Set the minimal luminance value. Default is @code{16}.
6863 Apply bilateral filter, spatial smoothing while preserving edges.
6865 The filter accepts the following options:
6868 Set sigma of gaussian function to calculate spatial weight.
6869 Allowed range is 0 to 512. Default is 0.1.
6872 Set sigma of gaussian function to calculate range weight.
6873 Allowed range is 0 to 1. Default is 0.1.
6876 Set planes to filter. Default is first only.
6879 @section bitplanenoise
6881 Show and measure bit plane noise.
6883 The filter accepts the following options:
6887 Set which plane to analyze. Default is @code{1}.
6890 Filter out noisy pixels from @code{bitplane} set above.
6891 Default is disabled.
6894 @section blackdetect
6896 Detect video intervals that are (almost) completely black. Can be
6897 useful to detect chapter transitions, commercials, or invalid
6900 The filter outputs its detection analysis to both the log as well as
6901 frame metadata. If a black segment of at least the specified minimum
6902 duration is found, a line with the start and end timestamps as well
6903 as duration is printed to the log with level @code{info}. In addition,
6904 a log line with level @code{debug} is printed per frame showing the
6905 black amount detected for that frame.
6907 The filter also attaches metadata to the first frame of a black
6908 segment with key @code{lavfi.black_start} and to the first frame
6909 after the black segment ends with key @code{lavfi.black_end}. The
6910 value is the frame's timestamp. This metadata is added regardless
6911 of the minimum duration specified.
6913 The filter accepts the following options:
6916 @item black_min_duration, d
6917 Set the minimum detected black duration expressed in seconds. It must
6918 be a non-negative floating point number.
6920 Default value is 2.0.
6922 @item picture_black_ratio_th, pic_th
6923 Set the threshold for considering a picture "black".
6924 Express the minimum value for the ratio:
6926 @var{nb_black_pixels} / @var{nb_pixels}
6929 for which a picture is considered black.
6930 Default value is 0.98.
6932 @item pixel_black_th, pix_th
6933 Set the threshold for considering a pixel "black".
6935 The threshold expresses the maximum pixel luminance value for which a
6936 pixel is considered "black". The provided value is scaled according to
6937 the following equation:
6939 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6942 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6943 the input video format, the range is [0-255] for YUV full-range
6944 formats and [16-235] for YUV non full-range formats.
6946 Default value is 0.10.
6949 The following example sets the maximum pixel threshold to the minimum
6950 value, and detects only black intervals of 2 or more seconds:
6952 blackdetect=d=2:pix_th=0.00
6957 Detect frames that are (almost) completely black. Can be useful to
6958 detect chapter transitions or commercials. Output lines consist of
6959 the frame number of the detected frame, the percentage of blackness,
6960 the position in the file if known or -1 and the timestamp in seconds.
6962 In order to display the output lines, you need to set the loglevel at
6963 least to the AV_LOG_INFO value.
6965 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6966 The value represents the percentage of pixels in the picture that
6967 are below the threshold value.
6969 It accepts the following parameters:
6974 The percentage of the pixels that have to be below the threshold; it defaults to
6977 @item threshold, thresh
6978 The threshold below which a pixel value is considered black; it defaults to
6986 Blend two video frames into each other.
6988 The @code{blend} filter takes two input streams and outputs one
6989 stream, the first input is the "top" layer and second input is
6990 "bottom" layer. By default, the output terminates when the longest input terminates.
6992 The @code{tblend} (time blend) filter takes two consecutive frames
6993 from one single stream, and outputs the result obtained by blending
6994 the new frame on top of the old frame.
6996 A description of the accepted options follows.
7004 Set blend mode for specific pixel component or all pixel components in case
7005 of @var{all_mode}. Default value is @code{normal}.
7007 Available values for component modes are:
7049 Set blend opacity for specific pixel component or all pixel components in case
7050 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7057 Set blend expression for specific pixel component or all pixel components in case
7058 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7060 The expressions can use the following variables:
7064 The sequential number of the filtered frame, starting from @code{0}.
7068 the coordinates of the current sample
7072 the width and height of currently filtered plane
7076 Width and height scale for the plane being filtered. It is the
7077 ratio between the dimensions of the current plane to the luma plane,
7078 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7079 the luma plane and @code{0.5,0.5} for the chroma planes.
7082 Time of the current frame, expressed in seconds.
7085 Value of pixel component at current location for first video frame (top layer).
7088 Value of pixel component at current location for second video frame (bottom layer).
7092 The @code{blend} filter also supports the @ref{framesync} options.
7094 @subsection Examples
7098 Apply transition from bottom layer to top layer in first 10 seconds:
7100 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7104 Apply linear horizontal transition from top layer to bottom layer:
7106 blend=all_expr='A*(X/W)+B*(1-X/W)'
7110 Apply 1x1 checkerboard effect:
7112 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7116 Apply uncover left effect:
7118 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7122 Apply uncover down effect:
7124 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7128 Apply uncover up-left effect:
7130 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7134 Split diagonally video and shows top and bottom layer on each side:
7136 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7140 Display differences between the current and the previous frame:
7142 tblend=all_mode=grainextract
7148 Denoise frames using Block-Matching 3D algorithm.
7150 The filter accepts the following options.
7154 Set denoising strength. Default value is 1.
7155 Allowed range is from 0 to 999.9.
7156 The denoising algorithm is very sensitive to sigma, so adjust it
7157 according to the source.
7160 Set local patch size. This sets dimensions in 2D.
7163 Set sliding step for processing blocks. Default value is 4.
7164 Allowed range is from 1 to 64.
7165 Smaller values allows processing more reference blocks and is slower.
7168 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7169 When set to 1, no block matching is done. Larger values allows more blocks
7171 Allowed range is from 1 to 256.
7174 Set radius for search block matching. Default is 9.
7175 Allowed range is from 1 to INT32_MAX.
7178 Set step between two search locations for block matching. Default is 1.
7179 Allowed range is from 1 to 64. Smaller is slower.
7182 Set threshold of mean square error for block matching. Valid range is 0 to
7186 Set thresholding parameter for hard thresholding in 3D transformed domain.
7187 Larger values results in stronger hard-thresholding filtering in frequency
7191 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7192 Default is @code{basic}.
7195 If enabled, filter will use 2nd stream for block matching.
7196 Default is disabled for @code{basic} value of @var{estim} option,
7197 and always enabled if value of @var{estim} is @code{final}.
7200 Set planes to filter. Default is all available except alpha.
7203 @subsection Examples
7207 Basic filtering with bm3d:
7209 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7213 Same as above, but filtering only luma:
7215 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7219 Same as above, but with both estimation modes:
7221 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
7225 Same as above, but prefilter with @ref{nlmeans} filter instead:
7227 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
7233 Apply a boxblur algorithm to the input video.
7235 It accepts the following parameters:
7239 @item luma_radius, lr
7240 @item luma_power, lp
7241 @item chroma_radius, cr
7242 @item chroma_power, cp
7243 @item alpha_radius, ar
7244 @item alpha_power, ap
7248 A description of the accepted options follows.
7251 @item luma_radius, lr
7252 @item chroma_radius, cr
7253 @item alpha_radius, ar
7254 Set an expression for the box radius in pixels used for blurring the
7255 corresponding input plane.
7257 The radius value must be a non-negative number, and must not be
7258 greater than the value of the expression @code{min(w,h)/2} for the
7259 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7262 Default value for @option{luma_radius} is "2". If not specified,
7263 @option{chroma_radius} and @option{alpha_radius} default to the
7264 corresponding value set for @option{luma_radius}.
7266 The expressions can contain the following constants:
7270 The input width and height in pixels.
7274 The input chroma image width and height in pixels.
7278 The horizontal and vertical chroma subsample values. For example, for the
7279 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7282 @item luma_power, lp
7283 @item chroma_power, cp
7284 @item alpha_power, ap
7285 Specify how many times the boxblur filter is applied to the
7286 corresponding plane.
7288 Default value for @option{luma_power} is 2. If not specified,
7289 @option{chroma_power} and @option{alpha_power} default to the
7290 corresponding value set for @option{luma_power}.
7292 A value of 0 will disable the effect.
7295 @subsection Examples
7299 Apply a boxblur filter with the luma, chroma, and alpha radii
7302 boxblur=luma_radius=2:luma_power=1
7307 Set the luma radius to 2, and alpha and chroma radius to 0:
7309 boxblur=2:1:cr=0:ar=0
7313 Set the luma and chroma radii to a fraction of the video dimension:
7315 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7321 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7322 Deinterlacing Filter").
7324 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7325 interpolation algorithms.
7326 It accepts the following parameters:
7330 The interlacing mode to adopt. It accepts one of the following values:
7334 Output one frame for each frame.
7336 Output one frame for each field.
7339 The default value is @code{send_field}.
7342 The picture field parity assumed for the input interlaced video. It accepts one
7343 of the following values:
7347 Assume the top field is first.
7349 Assume the bottom field is first.
7351 Enable automatic detection of field parity.
7354 The default value is @code{auto}.
7355 If the interlacing is unknown or the decoder does not export this information,
7356 top field first will be assumed.
7359 Specify which frames to deinterlace. Accepts one of the following
7364 Deinterlace all frames.
7366 Only deinterlace frames marked as interlaced.
7369 The default value is @code{all}.
7374 Apply Contrast Adaptive Sharpen filter to video stream.
7376 The filter accepts the following options:
7380 Set the sharpening strength. Default value is 0.
7383 Set planes to filter. Default value is to filter all
7384 planes except alpha plane.
7388 Remove all color information for all colors except for certain one.
7390 The filter accepts the following options:
7394 The color which will not be replaced with neutral chroma.
7397 Similarity percentage with the above color.
7398 0.01 matches only the exact key color, while 1.0 matches everything.
7402 0.0 makes pixels either fully gray, or not gray at all.
7403 Higher values result in more preserved color.
7406 Signals that the color passed is already in YUV instead of RGB.
7408 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7409 This can be used to pass exact YUV values as hexadecimal numbers.
7412 @subsection Commands
7413 This filter supports same @ref{commands} as options.
7414 The command accepts the same syntax of the corresponding option.
7416 If the specified expression is not valid, it is kept at its current
7420 YUV colorspace color/chroma keying.
7422 The filter accepts the following options:
7426 The color which will be replaced with transparency.
7429 Similarity percentage with the key color.
7431 0.01 matches only the exact key color, while 1.0 matches everything.
7436 0.0 makes pixels either fully transparent, or not transparent at all.
7438 Higher values result in semi-transparent pixels, with a higher transparency
7439 the more similar the pixels color is to the key color.
7442 Signals that the color passed is already in YUV instead of RGB.
7444 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7445 This can be used to pass exact YUV values as hexadecimal numbers.
7448 @subsection Commands
7449 This filter supports same @ref{commands} as options.
7450 The command accepts the same syntax of the corresponding option.
7452 If the specified expression is not valid, it is kept at its current
7455 @subsection Examples
7459 Make every green pixel in the input image transparent:
7461 ffmpeg -i input.png -vf chromakey=green out.png
7465 Overlay a greenscreen-video on top of a static black background.
7467 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
7472 Reduce chrominance noise.
7474 The filter accepts the following options:
7478 Set threshold for averaging chrominance values.
7479 Sum of absolute difference of U and V pixel components or current
7480 pixel and neighbour pixels lower than this threshold will be used in
7481 averaging. Luma component is left unchanged and is copied to output.
7482 Default value is 30. Allowed range is from 1 to 200.
7485 Set horizontal radius of rectangle used for averaging.
7486 Allowed range is from 1 to 100. Default value is 5.
7489 Set vertical radius of rectangle used for averaging.
7490 Allowed range is from 1 to 100. Default value is 5.
7493 Set horizontal step when averaging. Default value is 1.
7494 Allowed range is from 1 to 50.
7495 Mostly useful to speed-up filtering.
7498 Set vertical step when averaging. Default value is 1.
7499 Allowed range is from 1 to 50.
7500 Mostly useful to speed-up filtering.
7503 @subsection Commands
7504 This filter supports same @ref{commands} as options.
7505 The command accepts the same syntax of the corresponding option.
7507 @section chromashift
7508 Shift chroma pixels horizontally and/or vertically.
7510 The filter accepts the following options:
7513 Set amount to shift chroma-blue horizontally.
7515 Set amount to shift chroma-blue vertically.
7517 Set amount to shift chroma-red horizontally.
7519 Set amount to shift chroma-red vertically.
7521 Set edge mode, can be @var{smear}, default, or @var{warp}.
7524 @subsection Commands
7526 This filter supports the all above options as @ref{commands}.
7530 Display CIE color diagram with pixels overlaid onto it.
7532 The filter accepts the following options:
7547 @item uhdtv, rec2020
7561 Set what gamuts to draw.
7563 See @code{system} option for available values.
7566 Set ciescope size, by default set to 512.
7569 Set intensity used to map input pixel values to CIE diagram.
7572 Set contrast used to draw tongue colors that are out of active color system gamut.
7575 Correct gamma displayed on scope, by default enabled.
7578 Show white point on CIE diagram, by default disabled.
7581 Set input gamma. Used only with XYZ input color space.
7586 Visualize information exported by some codecs.
7588 Some codecs can export information through frames using side-data or other
7589 means. For example, some MPEG based codecs export motion vectors through the
7590 @var{export_mvs} flag in the codec @option{flags2} option.
7592 The filter accepts the following option:
7596 Set motion vectors to visualize.
7598 Available flags for @var{mv} are:
7602 forward predicted MVs of P-frames
7604 forward predicted MVs of B-frames
7606 backward predicted MVs of B-frames
7610 Display quantization parameters using the chroma planes.
7613 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7615 Available flags for @var{mv_type} are:
7619 forward predicted MVs
7621 backward predicted MVs
7624 @item frame_type, ft
7625 Set frame type to visualize motion vectors of.
7627 Available flags for @var{frame_type} are:
7631 intra-coded frames (I-frames)
7633 predicted frames (P-frames)
7635 bi-directionally predicted frames (B-frames)
7639 @subsection Examples
7643 Visualize forward predicted MVs of all frames using @command{ffplay}:
7645 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7649 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7651 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7655 @section colorbalance
7656 Modify intensity of primary colors (red, green and blue) of input frames.
7658 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7659 regions for the red-cyan, green-magenta or blue-yellow balance.
7661 A positive adjustment value shifts the balance towards the primary color, a negative
7662 value towards the complementary color.
7664 The filter accepts the following options:
7670 Adjust red, green and blue shadows (darkest pixels).
7675 Adjust red, green and blue midtones (medium pixels).
7680 Adjust red, green and blue highlights (brightest pixels).
7682 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7685 Preserve lightness when changing color balance. Default is disabled.
7688 @subsection Examples
7692 Add red color cast to shadows:
7698 @subsection Commands
7700 This filter supports the all above options as @ref{commands}.
7702 @section colorchannelmixer
7704 Adjust video input frames by re-mixing color channels.
7706 This filter modifies a color channel by adding the values associated to
7707 the other channels of the same pixels. For example if the value to
7708 modify is red, the output value will be:
7710 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7713 The filter accepts the following options:
7720 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7721 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7727 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7728 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7734 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7735 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7741 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7742 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7744 Allowed ranges for options are @code{[-2.0, 2.0]}.
7747 @subsection Examples
7751 Convert source to grayscale:
7753 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7756 Simulate sepia tones:
7758 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7762 @subsection Commands
7764 This filter supports the all above options as @ref{commands}.
7767 RGB colorspace color keying.
7769 The filter accepts the following options:
7773 The color which will be replaced with transparency.
7776 Similarity percentage with the key color.
7778 0.01 matches only the exact key color, while 1.0 matches everything.
7783 0.0 makes pixels either fully transparent, or not transparent at all.
7785 Higher values result in semi-transparent pixels, with a higher transparency
7786 the more similar the pixels color is to the key color.
7789 @subsection Examples
7793 Make every green pixel in the input image transparent:
7795 ffmpeg -i input.png -vf colorkey=green out.png
7799 Overlay a greenscreen-video on top of a static background image.
7801 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
7805 @subsection Commands
7806 This filter supports same @ref{commands} as options.
7807 The command accepts the same syntax of the corresponding option.
7809 If the specified expression is not valid, it is kept at its current
7813 Remove all color information for all RGB colors except for certain one.
7815 The filter accepts the following options:
7819 The color which will not be replaced with neutral gray.
7822 Similarity percentage with the above color.
7823 0.01 matches only the exact key color, while 1.0 matches everything.
7826 Blend percentage. 0.0 makes pixels fully gray.
7827 Higher values result in more preserved color.
7830 @subsection Commands
7831 This filter supports same @ref{commands} as options.
7832 The command accepts the same syntax of the corresponding option.
7834 If the specified expression is not valid, it is kept at its current
7837 @section colorlevels
7839 Adjust video input frames using levels.
7841 The filter accepts the following options:
7848 Adjust red, green, blue and alpha input black point.
7849 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7855 Adjust red, green, blue and alpha input white point.
7856 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7858 Input levels are used to lighten highlights (bright tones), darken shadows
7859 (dark tones), change the balance of bright and dark tones.
7865 Adjust red, green, blue and alpha output black point.
7866 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7872 Adjust red, green, blue and alpha output white point.
7873 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7875 Output levels allows manual selection of a constrained output level range.
7878 @subsection Examples
7882 Make video output darker:
7884 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7890 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7894 Make video output lighter:
7896 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7900 Increase brightness:
7902 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7906 @subsection Commands
7908 This filter supports the all above options as @ref{commands}.
7910 @section colormatrix
7912 Convert color matrix.
7914 The filter accepts the following options:
7919 Specify the source and destination color matrix. Both values must be
7922 The accepted values are:
7950 For example to convert from BT.601 to SMPTE-240M, use the command:
7952 colormatrix=bt601:smpte240m
7957 Convert colorspace, transfer characteristics or color primaries.
7958 Input video needs to have an even size.
7960 The filter accepts the following options:
7965 Specify all color properties at once.
7967 The accepted values are:
7997 Specify output colorspace.
7999 The accepted values are:
8008 BT.470BG or BT.601-6 625
8011 SMPTE-170M or BT.601-6 525
8020 BT.2020 with non-constant luminance
8026 Specify output transfer characteristics.
8028 The accepted values are:
8040 Constant gamma of 2.2
8043 Constant gamma of 2.8
8046 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8064 BT.2020 for 10-bits content
8067 BT.2020 for 12-bits content
8073 Specify output color primaries.
8075 The accepted values are:
8084 BT.470BG or BT.601-6 625
8087 SMPTE-170M or BT.601-6 525
8111 Specify output color range.
8113 The accepted values are:
8116 TV (restricted) range
8119 MPEG (restricted) range
8130 Specify output color format.
8132 The accepted values are:
8135 YUV 4:2:0 planar 8-bits
8138 YUV 4:2:0 planar 10-bits
8141 YUV 4:2:0 planar 12-bits
8144 YUV 4:2:2 planar 8-bits
8147 YUV 4:2:2 planar 10-bits
8150 YUV 4:2:2 planar 12-bits
8153 YUV 4:4:4 planar 8-bits
8156 YUV 4:4:4 planar 10-bits
8159 YUV 4:4:4 planar 12-bits
8164 Do a fast conversion, which skips gamma/primary correction. This will take
8165 significantly less CPU, but will be mathematically incorrect. To get output
8166 compatible with that produced by the colormatrix filter, use fast=1.
8169 Specify dithering mode.
8171 The accepted values are:
8177 Floyd-Steinberg dithering
8181 Whitepoint adaptation mode.
8183 The accepted values are:
8186 Bradford whitepoint adaptation
8189 von Kries whitepoint adaptation
8192 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8196 Override all input properties at once. Same accepted values as @ref{all}.
8199 Override input colorspace. Same accepted values as @ref{space}.
8202 Override input color primaries. Same accepted values as @ref{primaries}.
8205 Override input transfer characteristics. Same accepted values as @ref{trc}.
8208 Override input color range. Same accepted values as @ref{range}.
8212 The filter converts the transfer characteristics, color space and color
8213 primaries to the specified user values. The output value, if not specified,
8214 is set to a default value based on the "all" property. If that property is
8215 also not specified, the filter will log an error. The output color range and
8216 format default to the same value as the input color range and format. The
8217 input transfer characteristics, color space, color primaries and color range
8218 should be set on the input data. If any of these are missing, the filter will
8219 log an error and no conversion will take place.
8221 For example to convert the input to SMPTE-240M, use the command:
8223 colorspace=smpte240m
8226 @section convolution
8228 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8230 The filter accepts the following options:
8237 Set matrix for each plane.
8238 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8239 and from 1 to 49 odd number of signed integers in @var{row} mode.
8245 Set multiplier for calculated value for each plane.
8246 If unset or 0, it will be sum of all matrix elements.
8252 Set bias for each plane. This value is added to the result of the multiplication.
8253 Useful for making the overall image brighter or darker. Default is 0.0.
8259 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8260 Default is @var{square}.
8263 @subsection Examples
8269 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"
8275 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"
8281 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"
8287 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"
8291 Apply laplacian edge detector which includes diagonals:
8293 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"
8299 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"
8305 Apply 2D convolution of video stream in frequency domain using second stream
8308 The filter accepts the following options:
8312 Set which planes to process.
8315 Set which impulse video frames will be processed, can be @var{first}
8316 or @var{all}. Default is @var{all}.
8319 The @code{convolve} filter also supports the @ref{framesync} options.
8323 Copy the input video source unchanged to the output. This is mainly useful for
8328 Video filtering on GPU using Apple's CoreImage API on OSX.
8330 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8331 processed by video hardware. However, software-based OpenGL implementations
8332 exist which means there is no guarantee for hardware processing. It depends on
8335 There are many filters and image generators provided by Apple that come with a
8336 large variety of options. The filter has to be referenced by its name along
8339 The coreimage filter accepts the following options:
8342 List all available filters and generators along with all their respective
8343 options as well as possible minimum and maximum values along with the default
8350 Specify all filters by their respective name and options.
8351 Use @var{list_filters} to determine all valid filter names and options.
8352 Numerical options are specified by a float value and are automatically clamped
8353 to their respective value range. Vector and color options have to be specified
8354 by a list of space separated float values. Character escaping has to be done.
8355 A special option name @code{default} is available to use default options for a
8358 It is required to specify either @code{default} or at least one of the filter options.
8359 All omitted options are used with their default values.
8360 The syntax of the filter string is as follows:
8362 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8366 Specify a rectangle where the output of the filter chain is copied into the
8367 input image. It is given by a list of space separated float values:
8369 output_rect=x\ y\ width\ height
8371 If not given, the output rectangle equals the dimensions of the input image.
8372 The output rectangle is automatically cropped at the borders of the input
8373 image. Negative values are valid for each component.
8375 output_rect=25\ 25\ 100\ 100
8379 Several filters can be chained for successive processing without GPU-HOST
8380 transfers allowing for fast processing of complex filter chains.
8381 Currently, only filters with zero (generators) or exactly one (filters) input
8382 image and one output image are supported. Also, transition filters are not yet
8385 Some filters generate output images with additional padding depending on the
8386 respective filter kernel. The padding is automatically removed to ensure the
8387 filter output has the same size as the input image.
8389 For image generators, the size of the output image is determined by the
8390 previous output image of the filter chain or the input image of the whole
8391 filterchain, respectively. The generators do not use the pixel information of
8392 this image to generate their output. However, the generated output is
8393 blended onto this image, resulting in partial or complete coverage of the
8396 The @ref{coreimagesrc} video source can be used for generating input images
8397 which are directly fed into the filter chain. By using it, providing input
8398 images by another video source or an input video is not required.
8400 @subsection Examples
8405 List all filters available:
8407 coreimage=list_filters=true
8411 Use the CIBoxBlur filter with default options to blur an image:
8413 coreimage=filter=CIBoxBlur@@default
8417 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8418 its center at 100x100 and a radius of 50 pixels:
8420 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8424 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8425 given as complete and escaped command-line for Apple's standard bash shell:
8427 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8433 Cover a rectangular object
8435 It accepts the following options:
8439 Filepath of the optional cover image, needs to be in yuv420.
8444 It accepts the following values:
8447 cover it by the supplied image
8449 cover it by interpolating the surrounding pixels
8452 Default value is @var{blur}.
8455 @subsection Examples
8459 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8461 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8467 Crop the input video to given dimensions.
8469 It accepts the following parameters:
8473 The width of the output video. It defaults to @code{iw}.
8474 This expression is evaluated only once during the filter
8475 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8478 The height of the output video. It defaults to @code{ih}.
8479 This expression is evaluated only once during the filter
8480 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8483 The horizontal position, in the input video, of the left edge of the output
8484 video. It defaults to @code{(in_w-out_w)/2}.
8485 This expression is evaluated per-frame.
8488 The vertical position, in the input video, of the top edge of the output video.
8489 It defaults to @code{(in_h-out_h)/2}.
8490 This expression is evaluated per-frame.
8493 If set to 1 will force the output display aspect ratio
8494 to be the same of the input, by changing the output sample aspect
8495 ratio. It defaults to 0.
8498 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8499 width/height/x/y as specified and will not be rounded to nearest smaller value.
8503 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8504 expressions containing the following constants:
8509 The computed values for @var{x} and @var{y}. They are evaluated for
8514 The input width and height.
8518 These are the same as @var{in_w} and @var{in_h}.
8522 The output (cropped) width and height.
8526 These are the same as @var{out_w} and @var{out_h}.
8529 same as @var{iw} / @var{ih}
8532 input sample aspect ratio
8535 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8539 horizontal and vertical chroma subsample values. For example for the
8540 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8543 The number of the input frame, starting from 0.
8546 the position in the file of the input frame, NAN if unknown
8549 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8553 The expression for @var{out_w} may depend on the value of @var{out_h},
8554 and the expression for @var{out_h} may depend on @var{out_w}, but they
8555 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8556 evaluated after @var{out_w} and @var{out_h}.
8558 The @var{x} and @var{y} parameters specify the expressions for the
8559 position of the top-left corner of the output (non-cropped) area. They
8560 are evaluated for each frame. If the evaluated value is not valid, it
8561 is approximated to the nearest valid value.
8563 The expression for @var{x} may depend on @var{y}, and the expression
8564 for @var{y} may depend on @var{x}.
8566 @subsection Examples
8570 Crop area with size 100x100 at position (12,34).
8575 Using named options, the example above becomes:
8577 crop=w=100:h=100:x=12:y=34
8581 Crop the central input area with size 100x100:
8587 Crop the central input area with size 2/3 of the input video:
8589 crop=2/3*in_w:2/3*in_h
8593 Crop the input video central square:
8600 Delimit the rectangle with the top-left corner placed at position
8601 100:100 and the right-bottom corner corresponding to the right-bottom
8602 corner of the input image.
8604 crop=in_w-100:in_h-100:100:100
8608 Crop 10 pixels from the left and right borders, and 20 pixels from
8609 the top and bottom borders
8611 crop=in_w-2*10:in_h-2*20
8615 Keep only the bottom right quarter of the input image:
8617 crop=in_w/2:in_h/2:in_w/2:in_h/2
8621 Crop height for getting Greek harmony:
8623 crop=in_w:1/PHI*in_w
8627 Apply trembling effect:
8629 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)
8633 Apply erratic camera effect depending on timestamp:
8635 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)"
8639 Set x depending on the value of y:
8641 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8645 @subsection Commands
8647 This filter supports the following commands:
8653 Set width/height of the output video and the horizontal/vertical position
8655 The command accepts the same syntax of the corresponding option.
8657 If the specified expression is not valid, it is kept at its current
8663 Auto-detect the crop size.
8665 It calculates the necessary cropping parameters and prints the
8666 recommended parameters via the logging system. The detected dimensions
8667 correspond to the non-black area of the input video.
8669 It accepts the following parameters:
8674 Set higher black value threshold, which can be optionally specified
8675 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8676 value greater to the set value is considered non-black. It defaults to 24.
8677 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8678 on the bitdepth of the pixel format.
8681 The value which the width/height should be divisible by. It defaults to
8682 16. The offset is automatically adjusted to center the video. Use 2 to
8683 get only even dimensions (needed for 4:2:2 video). 16 is best when
8684 encoding to most video codecs.
8686 @item reset_count, reset
8687 Set the counter that determines after how many frames cropdetect will
8688 reset the previously detected largest video area and start over to
8689 detect the current optimal crop area. Default value is 0.
8691 This can be useful when channel logos distort the video area. 0
8692 indicates 'never reset', and returns the largest area encountered during
8699 Delay video filtering until a given wallclock timestamp. The filter first
8700 passes on @option{preroll} amount of frames, then it buffers at most
8701 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8702 it forwards the buffered frames and also any subsequent frames coming in its
8705 The filter can be used synchronize the output of multiple ffmpeg processes for
8706 realtime output devices like decklink. By putting the delay in the filtering
8707 chain and pre-buffering frames the process can pass on data to output almost
8708 immediately after the target wallclock timestamp is reached.
8710 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8716 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8719 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8722 The maximum duration of content to buffer before waiting for the cue expressed
8723 in seconds. Default is 0.
8730 Apply color adjustments using curves.
8732 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8733 component (red, green and blue) has its values defined by @var{N} key points
8734 tied from each other using a smooth curve. The x-axis represents the pixel
8735 values from the input frame, and the y-axis the new pixel values to be set for
8738 By default, a component curve is defined by the two points @var{(0;0)} and
8739 @var{(1;1)}. This creates a straight line where each original pixel value is
8740 "adjusted" to its own value, which means no change to the image.
8742 The filter allows you to redefine these two points and add some more. A new
8743 curve (using a natural cubic spline interpolation) will be define to pass
8744 smoothly through all these new coordinates. The new defined points needs to be
8745 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8746 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8747 the vector spaces, the values will be clipped accordingly.
8749 The filter accepts the following options:
8753 Select one of the available color presets. This option can be used in addition
8754 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8755 options takes priority on the preset values.
8756 Available presets are:
8759 @item color_negative
8762 @item increase_contrast
8764 @item linear_contrast
8765 @item medium_contrast
8767 @item strong_contrast
8770 Default is @code{none}.
8772 Set the master key points. These points will define a second pass mapping. It
8773 is sometimes called a "luminance" or "value" mapping. It can be used with
8774 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8775 post-processing LUT.
8777 Set the key points for the red component.
8779 Set the key points for the green component.
8781 Set the key points for the blue component.
8783 Set the key points for all components (not including master).
8784 Can be used in addition to the other key points component
8785 options. In this case, the unset component(s) will fallback on this
8786 @option{all} setting.
8788 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
8790 Save Gnuplot script of the curves in specified file.
8793 To avoid some filtergraph syntax conflicts, each key points list need to be
8794 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
8796 @subsection Examples
8800 Increase slightly the middle level of blue:
8802 curves=blue='0/0 0.5/0.58 1/1'
8808 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'
8810 Here we obtain the following coordinates for each components:
8813 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8815 @code{(0;0) (0.50;0.48) (1;1)}
8817 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8821 The previous example can also be achieved with the associated built-in preset:
8823 curves=preset=vintage
8833 Use a Photoshop preset and redefine the points of the green component:
8835 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8839 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8840 and @command{gnuplot}:
8842 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8843 gnuplot -p /tmp/curves.plt
8849 Video data analysis filter.
8851 This filter shows hexadecimal pixel values of part of video.
8853 The filter accepts the following options:
8857 Set output video size.
8860 Set x offset from where to pick pixels.
8863 Set y offset from where to pick pixels.
8866 Set scope mode, can be one of the following:
8869 Draw hexadecimal pixel values with white color on black background.
8872 Draw hexadecimal pixel values with input video pixel color on black
8876 Draw hexadecimal pixel values on color background picked from input video,
8877 the text color is picked in such way so its always visible.
8881 Draw rows and columns numbers on left and top of video.
8884 Set background opacity.
8887 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
8891 Apply Directional blur filter.
8893 The filter accepts the following options:
8897 Set angle of directional blur. Default is @code{45}.
8900 Set radius of directional blur. Default is @code{5}.
8903 Set which planes to filter. By default all planes are filtered.
8906 @subsection Commands
8907 This filter supports same @ref{commands} as options.
8908 The command accepts the same syntax of the corresponding option.
8910 If the specified expression is not valid, it is kept at its current
8915 Denoise frames using 2D DCT (frequency domain filtering).
8917 This filter is not designed for real time.
8919 The filter accepts the following options:
8923 Set the noise sigma constant.
8925 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8926 coefficient (absolute value) below this threshold with be dropped.
8928 If you need a more advanced filtering, see @option{expr}.
8930 Default is @code{0}.
8933 Set number overlapping pixels for each block. Since the filter can be slow, you
8934 may want to reduce this value, at the cost of a less effective filter and the
8935 risk of various artefacts.
8937 If the overlapping value doesn't permit processing the whole input width or
8938 height, a warning will be displayed and according borders won't be denoised.
8940 Default value is @var{blocksize}-1, which is the best possible setting.
8943 Set the coefficient factor expression.
8945 For each coefficient of a DCT block, this expression will be evaluated as a
8946 multiplier value for the coefficient.
8948 If this is option is set, the @option{sigma} option will be ignored.
8950 The absolute value of the coefficient can be accessed through the @var{c}
8954 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8955 @var{blocksize}, which is the width and height of the processed blocks.
8957 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8958 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8959 on the speed processing. Also, a larger block size does not necessarily means a
8963 @subsection Examples
8965 Apply a denoise with a @option{sigma} of @code{4.5}:
8970 The same operation can be achieved using the expression system:
8972 dctdnoiz=e='gte(c, 4.5*3)'
8975 Violent denoise using a block size of @code{16x16}:
8982 Remove banding artifacts from input video.
8983 It works by replacing banded pixels with average value of referenced pixels.
8985 The filter accepts the following options:
8992 Set banding detection threshold for each plane. Default is 0.02.
8993 Valid range is 0.00003 to 0.5.
8994 If difference between current pixel and reference pixel is less than threshold,
8995 it will be considered as banded.
8998 Banding detection range in pixels. Default is 16. If positive, random number
8999 in range 0 to set value will be used. If negative, exact absolute value
9001 The range defines square of four pixels around current pixel.
9004 Set direction in radians from which four pixel will be compared. If positive,
9005 random direction from 0 to set direction will be picked. If negative, exact of
9006 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9007 will pick only pixels on same row and -PI/2 will pick only pixels on same
9011 If enabled, current pixel is compared with average value of all four
9012 surrounding pixels. The default is enabled. If disabled current pixel is
9013 compared with all four surrounding pixels. The pixel is considered banded
9014 if only all four differences with surrounding pixels are less than threshold.
9017 If enabled, current pixel is changed if and only if all pixel components are banded,
9018 e.g. banding detection threshold is triggered for all color components.
9019 The default is disabled.
9024 Remove blocking artifacts from input video.
9026 The filter accepts the following options:
9030 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9031 This controls what kind of deblocking is applied.
9034 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9040 Set blocking detection thresholds. Allowed range is 0 to 1.
9041 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9042 Using higher threshold gives more deblocking strength.
9043 Setting @var{alpha} controls threshold detection at exact edge of block.
9044 Remaining options controls threshold detection near the edge. Each one for
9045 below/above or left/right. Setting any of those to @var{0} disables
9049 Set planes to filter. Default is to filter all available planes.
9052 @subsection Examples
9056 Deblock using weak filter and block size of 4 pixels.
9058 deblock=filter=weak:block=4
9062 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9063 deblocking more edges.
9065 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9069 Similar as above, but filter only first plane.
9071 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9075 Similar as above, but filter only second and third plane.
9077 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9084 Drop duplicated frames at regular intervals.
9086 The filter accepts the following options:
9090 Set the number of frames from which one will be dropped. Setting this to
9091 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9092 Default is @code{5}.
9095 Set the threshold for duplicate detection. If the difference metric for a frame
9096 is less than or equal to this value, then it is declared as duplicate. Default
9100 Set scene change threshold. Default is @code{15}.
9104 Set the size of the x and y-axis blocks used during metric calculations.
9105 Larger blocks give better noise suppression, but also give worse detection of
9106 small movements. Must be a power of two. Default is @code{32}.
9109 Mark main input as a pre-processed input and activate clean source input
9110 stream. This allows the input to be pre-processed with various filters to help
9111 the metrics calculation while keeping the frame selection lossless. When set to
9112 @code{1}, the first stream is for the pre-processed input, and the second
9113 stream is the clean source from where the kept frames are chosen. Default is
9117 Set whether or not chroma is considered in the metric calculations. Default is
9123 Apply 2D deconvolution of video stream in frequency domain using second stream
9126 The filter accepts the following options:
9130 Set which planes to process.
9133 Set which impulse video frames will be processed, can be @var{first}
9134 or @var{all}. Default is @var{all}.
9137 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9138 and height are not same and not power of 2 or if stream prior to convolving
9142 The @code{deconvolve} filter also supports the @ref{framesync} options.
9146 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9148 It accepts the following options:
9152 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9153 @var{rainbows} for cross-color reduction.
9156 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9159 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9162 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9165 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9170 Apply deflate effect to the video.
9172 This filter replaces the pixel by the local(3x3) average by taking into account
9173 only values lower than the pixel.
9175 It accepts the following options:
9182 Limit the maximum change for each plane, default is 65535.
9183 If 0, plane will remain unchanged.
9186 @subsection Commands
9188 This filter supports the all above options as @ref{commands}.
9192 Remove temporal frame luminance variations.
9194 It accepts the following options:
9198 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9201 Set averaging mode to smooth temporal luminance variations.
9203 Available values are:
9228 Do not actually modify frame. Useful when one only wants metadata.
9233 Remove judder produced by partially interlaced telecined content.
9235 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9236 source was partially telecined content then the output of @code{pullup,dejudder}
9237 will have a variable frame rate. May change the recorded frame rate of the
9238 container. Aside from that change, this filter will not affect constant frame
9241 The option available in this filter is:
9245 Specify the length of the window over which the judder repeats.
9247 Accepts any integer greater than 1. Useful values are:
9251 If the original was telecined from 24 to 30 fps (Film to NTSC).
9254 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9257 If a mixture of the two.
9260 The default is @samp{4}.
9265 Suppress a TV station logo by a simple interpolation of the surrounding
9266 pixels. Just set a rectangle covering the logo and watch it disappear
9267 (and sometimes something even uglier appear - your mileage may vary).
9269 It accepts the following parameters:
9274 Specify the top left corner coordinates of the logo. They must be
9279 Specify the width and height of the logo to clear. They must be
9283 Specify the thickness of the fuzzy edge of the rectangle (added to
9284 @var{w} and @var{h}). The default value is 1. This option is
9285 deprecated, setting higher values should no longer be necessary and
9289 When set to 1, a green rectangle is drawn on the screen to simplify
9290 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9291 The default value is 0.
9293 The rectangle is drawn on the outermost pixels which will be (partly)
9294 replaced with interpolated values. The values of the next pixels
9295 immediately outside this rectangle in each direction will be used to
9296 compute the interpolated pixel values inside the rectangle.
9300 @subsection Examples
9304 Set a rectangle covering the area with top left corner coordinates 0,0
9305 and size 100x77, and a band of size 10:
9307 delogo=x=0:y=0:w=100:h=77:band=10
9315 Remove the rain in the input image/video by applying the derain methods based on
9316 convolutional neural networks. Supported models:
9320 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9321 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9324 Training as well as model generation scripts are provided in
9325 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9327 Native model files (.model) can be generated from TensorFlow model
9328 files (.pb) by using tools/python/convert.py
9330 The filter accepts the following options:
9334 Specify which filter to use. This option accepts the following values:
9338 Derain filter. To conduct derain filter, you need to use a derain model.
9341 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9343 Default value is @samp{derain}.
9346 Specify which DNN backend to use for model loading and execution. This option accepts
9347 the following values:
9351 Native implementation of DNN loading and execution.
9354 TensorFlow backend. To enable this backend you
9355 need to install the TensorFlow for C library (see
9356 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9357 @code{--enable-libtensorflow}
9359 Default value is @samp{native}.
9362 Set path to model file specifying network architecture and its parameters.
9363 Note that different backends use different file formats. TensorFlow and native
9364 backend can load files for only its format.
9367 It can also be finished with @ref{dnn_processing} filter.
9371 Attempt to fix small changes in horizontal and/or vertical shift. This
9372 filter helps remove camera shake from hand-holding a camera, bumping a
9373 tripod, moving on a vehicle, etc.
9375 The filter accepts the following options:
9383 Specify a rectangular area where to limit the search for motion
9385 If desired the search for motion vectors can be limited to a
9386 rectangular area of the frame defined by its top left corner, width
9387 and height. These parameters have the same meaning as the drawbox
9388 filter which can be used to visualise the position of the bounding
9391 This is useful when simultaneous movement of subjects within the frame
9392 might be confused for camera motion by the motion vector search.
9394 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9395 then the full frame is used. This allows later options to be set
9396 without specifying the bounding box for the motion vector search.
9398 Default - search the whole frame.
9402 Specify the maximum extent of movement in x and y directions in the
9403 range 0-64 pixels. Default 16.
9406 Specify how to generate pixels to fill blanks at the edge of the
9407 frame. Available values are:
9410 Fill zeroes at blank locations
9412 Original image at blank locations
9414 Extruded edge value at blank locations
9416 Mirrored edge at blank locations
9418 Default value is @samp{mirror}.
9421 Specify the blocksize to use for motion search. Range 4-128 pixels,
9425 Specify the contrast threshold for blocks. Only blocks with more than
9426 the specified contrast (difference between darkest and lightest
9427 pixels) will be considered. Range 1-255, default 125.
9430 Specify the search strategy. Available values are:
9433 Set exhaustive search
9435 Set less exhaustive search.
9437 Default value is @samp{exhaustive}.
9440 If set then a detailed log of the motion search is written to the
9447 Remove unwanted contamination of foreground colors, caused by reflected color of
9448 greenscreen or bluescreen.
9450 This filter accepts the following options:
9454 Set what type of despill to use.
9457 Set how spillmap will be generated.
9460 Set how much to get rid of still remaining spill.
9463 Controls amount of red in spill area.
9466 Controls amount of green in spill area.
9467 Should be -1 for greenscreen.
9470 Controls amount of blue in spill area.
9471 Should be -1 for bluescreen.
9474 Controls brightness of spill area, preserving colors.
9477 Modify alpha from generated spillmap.
9480 @subsection Commands
9482 This filter supports the all above options as @ref{commands}.
9486 Apply an exact inverse of the telecine operation. It requires a predefined
9487 pattern specified using the pattern option which must be the same as that passed
9488 to the telecine filter.
9490 This filter accepts the following options:
9499 The default value is @code{top}.
9503 A string of numbers representing the pulldown pattern you wish to apply.
9504 The default value is @code{23}.
9507 A number representing position of the first frame with respect to the telecine
9508 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9513 Apply dilation effect to the video.
9515 This filter replaces the pixel by the local(3x3) maximum.
9517 It accepts the following options:
9524 Limit the maximum change for each plane, default is 65535.
9525 If 0, plane will remain unchanged.
9528 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9531 Flags to local 3x3 coordinates maps like this:
9538 @subsection Commands
9540 This filter supports the all above options as @ref{commands}.
9544 Displace pixels as indicated by second and third input stream.
9546 It takes three input streams and outputs one stream, the first input is the
9547 source, and second and third input are displacement maps.
9549 The second input specifies how much to displace pixels along the
9550 x-axis, while the third input specifies how much to displace pixels
9552 If one of displacement map streams terminates, last frame from that
9553 displacement map will be used.
9555 Note that once generated, displacements maps can be reused over and over again.
9557 A description of the accepted options follows.
9561 Set displace behavior for pixels that are out of range.
9563 Available values are:
9566 Missing pixels are replaced by black pixels.
9569 Adjacent pixels will spread out to replace missing pixels.
9572 Out of range pixels are wrapped so they point to pixels of other side.
9575 Out of range pixels will be replaced with mirrored pixels.
9577 Default is @samp{smear}.
9581 @subsection Examples
9585 Add ripple effect to rgb input of video size hd720:
9587 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
9591 Add wave effect to rgb input of video size hd720:
9593 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
9597 @anchor{dnn_processing}
9598 @section dnn_processing
9600 Do image processing with deep neural networks. It works together with another filter
9601 which converts the pixel format of the Frame to what the dnn network requires.
9603 The filter accepts the following options:
9607 Specify which DNN backend to use for model loading and execution. This option accepts
9608 the following values:
9612 Native implementation of DNN loading and execution.
9615 TensorFlow backend. To enable this backend you
9616 need to install the TensorFlow for C library (see
9617 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9618 @code{--enable-libtensorflow}
9621 OpenVINO backend. To enable this backend you
9622 need to build and install the OpenVINO for C library (see
9623 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9624 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9625 be needed if the header files and libraries are not installed into system path)
9629 Default value is @samp{native}.
9632 Set path to model file specifying network architecture and its parameters.
9633 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9634 backend can load files for only its format.
9636 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9639 Set the input name of the dnn network.
9642 Set the output name of the dnn network.
9646 @subsection Examples
9650 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9652 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9656 Halve the pixel value of the frame with format gray32f:
9658 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
9662 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9664 ./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
9668 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9670 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9677 Draw a colored box on the input image.
9679 It accepts the following parameters:
9684 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9688 The expressions which specify the width and height of the box; if 0 they are interpreted as
9689 the input width and height. It defaults to 0.
9692 Specify the color of the box to write. For the general syntax of this option,
9693 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9694 value @code{invert} is used, the box edge color is the same as the
9695 video with inverted luma.
9698 The expression which sets the thickness of the box edge.
9699 A value of @code{fill} will create a filled box. Default value is @code{3}.
9701 See below for the list of accepted constants.
9704 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
9705 will overwrite the video's color and alpha pixels.
9706 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
9709 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9710 following constants:
9714 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9718 horizontal and vertical chroma subsample values. For example for the
9719 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9723 The input width and height.
9726 The input sample aspect ratio.
9730 The x and y offset coordinates where the box is drawn.
9734 The width and height of the drawn box.
9737 The thickness of the drawn box.
9739 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9740 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9744 @subsection Examples
9748 Draw a black box around the edge of the input image:
9754 Draw a box with color red and an opacity of 50%:
9756 drawbox=10:20:200:60:red@@0.5
9759 The previous example can be specified as:
9761 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
9765 Fill the box with pink color:
9767 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
9771 Draw a 2-pixel red 2.40:1 mask:
9773 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
9777 @subsection Commands
9778 This filter supports same commands as options.
9779 The command accepts the same syntax of the corresponding option.
9781 If the specified expression is not valid, it is kept at its current
9786 Draw a graph using input video metadata.
9788 It accepts the following parameters:
9792 Set 1st frame metadata key from which metadata values will be used to draw a graph.
9795 Set 1st foreground color expression.
9798 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
9801 Set 2nd foreground color expression.
9804 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
9807 Set 3rd foreground color expression.
9810 Set 4th frame metadata key from which metadata values will be used to draw a graph.
9813 Set 4th foreground color expression.
9816 Set minimal value of metadata value.
9819 Set maximal value of metadata value.
9822 Set graph background color. Default is white.
9827 Available values for mode is:
9834 Default is @code{line}.
9839 Available values for slide is:
9842 Draw new frame when right border is reached.
9845 Replace old columns with new ones.
9848 Scroll from right to left.
9851 Scroll from left to right.
9854 Draw single picture.
9857 Default is @code{frame}.
9860 Set size of graph video. For the syntax of this option, check the
9861 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
9862 The default value is @code{900x256}.
9865 Set the output frame rate. Default value is @code{25}.
9867 The foreground color expressions can use the following variables:
9870 Minimal value of metadata value.
9873 Maximal value of metadata value.
9876 Current metadata key value.
9879 The color is defined as 0xAABBGGRR.
9882 Example using metadata from @ref{signalstats} filter:
9884 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
9887 Example using metadata from @ref{ebur128} filter:
9889 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
9894 Draw a grid on the input image.
9896 It accepts the following parameters:
9901 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
9905 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
9906 input width and height, respectively, minus @code{thickness}, so image gets
9907 framed. Default to 0.
9910 Specify the color of the grid. For the general syntax of this option,
9911 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9912 value @code{invert} is used, the grid color is the same as the
9913 video with inverted luma.
9916 The expression which sets the thickness of the grid line. Default value is @code{1}.
9918 See below for the list of accepted constants.
9921 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
9922 will overwrite the video's color and alpha pixels.
9923 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
9926 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9927 following constants:
9931 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9935 horizontal and vertical chroma subsample values. For example for the
9936 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9940 The input grid cell width and height.
9943 The input sample aspect ratio.
9947 The x and y coordinates of some point of grid intersection (meant to configure offset).
9951 The width and height of the drawn cell.
9954 The thickness of the drawn cell.
9956 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9957 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9961 @subsection Examples
9965 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
9967 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
9971 Draw a white 3x3 grid with an opacity of 50%:
9973 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
9977 @subsection Commands
9978 This filter supports same commands as options.
9979 The command accepts the same syntax of the corresponding option.
9981 If the specified expression is not valid, it is kept at its current
9987 Draw a text string or text from a specified file on top of a video, using the
9988 libfreetype library.
9990 To enable compilation of this filter, you need to configure FFmpeg with
9991 @code{--enable-libfreetype}.
9992 To enable default font fallback and the @var{font} option you need to
9993 configure FFmpeg with @code{--enable-libfontconfig}.
9994 To enable the @var{text_shaping} option, you need to configure FFmpeg with
9995 @code{--enable-libfribidi}.
9999 It accepts the following parameters:
10004 Used to draw a box around text using the background color.
10005 The value must be either 1 (enable) or 0 (disable).
10006 The default value of @var{box} is 0.
10009 Set the width of the border to be drawn around the box using @var{boxcolor}.
10010 The default value of @var{boxborderw} is 0.
10013 The color to be used for drawing box around text. For the syntax of this
10014 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10016 The default value of @var{boxcolor} is "white".
10019 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10020 The default value of @var{line_spacing} is 0.
10023 Set the width of the border to be drawn around the text using @var{bordercolor}.
10024 The default value of @var{borderw} is 0.
10027 Set the color to be used for drawing border around text. For the syntax of this
10028 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10030 The default value of @var{bordercolor} is "black".
10033 Select how the @var{text} is expanded. Can be either @code{none},
10034 @code{strftime} (deprecated) or
10035 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10039 Set a start time for the count. Value is in microseconds. Only applied
10040 in the deprecated strftime expansion mode. To emulate in normal expansion
10041 mode use the @code{pts} function, supplying the start time (in seconds)
10042 as the second argument.
10045 If true, check and fix text coords to avoid clipping.
10048 The color to be used for drawing fonts. For the syntax of this option, check
10049 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10051 The default value of @var{fontcolor} is "black".
10053 @item fontcolor_expr
10054 String which is expanded the same way as @var{text} to obtain dynamic
10055 @var{fontcolor} value. By default this option has empty value and is not
10056 processed. When this option is set, it overrides @var{fontcolor} option.
10059 The font family to be used for drawing text. By default Sans.
10062 The font file to be used for drawing text. The path must be included.
10063 This parameter is mandatory if the fontconfig support is disabled.
10066 Draw the text applying alpha blending. The value can
10067 be a number between 0.0 and 1.0.
10068 The expression accepts the same variables @var{x, y} as well.
10069 The default value is 1.
10070 Please see @var{fontcolor_expr}.
10073 The font size to be used for drawing text.
10074 The default value of @var{fontsize} is 16.
10077 If set to 1, attempt to shape the text (for example, reverse the order of
10078 right-to-left text and join Arabic characters) before drawing it.
10079 Otherwise, just draw the text exactly as given.
10080 By default 1 (if supported).
10082 @item ft_load_flags
10083 The flags to be used for loading the fonts.
10085 The flags map the corresponding flags supported by libfreetype, and are
10086 a combination of the following values:
10093 @item vertical_layout
10094 @item force_autohint
10097 @item ignore_global_advance_width
10099 @item ignore_transform
10101 @item linear_design
10105 Default value is "default".
10107 For more information consult the documentation for the FT_LOAD_*
10111 The color to be used for drawing a shadow behind the drawn text. For the
10112 syntax of this option, check the @ref{color syntax,,"Color" section in the
10113 ffmpeg-utils manual,ffmpeg-utils}.
10115 The default value of @var{shadowcolor} is "black".
10119 The x and y offsets for the text shadow position with respect to the
10120 position of the text. They can be either positive or negative
10121 values. The default value for both is "0".
10124 The starting frame number for the n/frame_num variable. The default value
10128 The size in number of spaces to use for rendering the tab.
10129 Default value is 4.
10132 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10133 format. It can be used with or without text parameter. @var{timecode_rate}
10134 option must be specified.
10136 @item timecode_rate, rate, r
10137 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10138 integer. Minimum value is "1".
10139 Drop-frame timecode is supported for frame rates 30 & 60.
10142 If set to 1, the output of the timecode option will wrap around at 24 hours.
10143 Default is 0 (disabled).
10146 The text string to be drawn. The text must be a sequence of UTF-8
10147 encoded characters.
10148 This parameter is mandatory if no file is specified with the parameter
10152 A text file containing text to be drawn. The text must be a sequence
10153 of UTF-8 encoded characters.
10155 This parameter is mandatory if no text string is specified with the
10156 parameter @var{text}.
10158 If both @var{text} and @var{textfile} are specified, an error is thrown.
10161 If set to 1, the @var{textfile} will be reloaded before each frame.
10162 Be sure to update it atomically, or it may be read partially, or even fail.
10166 The expressions which specify the offsets where text will be drawn
10167 within the video frame. They are relative to the top/left border of the
10170 The default value of @var{x} and @var{y} is "0".
10172 See below for the list of accepted constants and functions.
10175 The parameters for @var{x} and @var{y} are expressions containing the
10176 following constants and functions:
10180 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10184 horizontal and vertical chroma subsample values. For example for the
10185 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10188 the height of each text line
10196 @item max_glyph_a, ascent
10197 the maximum distance from the baseline to the highest/upper grid
10198 coordinate used to place a glyph outline point, for all the rendered
10200 It is a positive value, due to the grid's orientation with the Y axis
10203 @item max_glyph_d, descent
10204 the maximum distance from the baseline to the lowest grid coordinate
10205 used to place a glyph outline point, for all the rendered glyphs.
10206 This is a negative value, due to the grid's orientation, with the Y axis
10210 maximum glyph height, that is the maximum height for all the glyphs
10211 contained in the rendered text, it is equivalent to @var{ascent} -
10215 maximum glyph width, that is the maximum width for all the glyphs
10216 contained in the rendered text
10219 the number of input frame, starting from 0
10221 @item rand(min, max)
10222 return a random number included between @var{min} and @var{max}
10225 The input sample aspect ratio.
10228 timestamp expressed in seconds, NAN if the input timestamp is unknown
10231 the height of the rendered text
10234 the width of the rendered text
10238 the x and y offset coordinates where the text is drawn.
10240 These parameters allow the @var{x} and @var{y} expressions to refer
10241 to each other, so you can for example specify @code{y=x/dar}.
10244 A one character description of the current frame's picture type.
10247 The current packet's position in the input file or stream
10248 (in bytes, from the start of the input). A value of -1 indicates
10249 this info is not available.
10252 The current packet's duration, in seconds.
10255 The current packet's size (in bytes).
10258 @anchor{drawtext_expansion}
10259 @subsection Text expansion
10261 If @option{expansion} is set to @code{strftime},
10262 the filter recognizes strftime() sequences in the provided text and
10263 expands them accordingly. Check the documentation of strftime(). This
10264 feature is deprecated.
10266 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10268 If @option{expansion} is set to @code{normal} (which is the default),
10269 the following expansion mechanism is used.
10271 The backslash character @samp{\}, followed by any character, always expands to
10272 the second character.
10274 Sequences of the form @code{%@{...@}} are expanded. The text between the
10275 braces is a function name, possibly followed by arguments separated by ':'.
10276 If the arguments contain special characters or delimiters (':' or '@}'),
10277 they should be escaped.
10279 Note that they probably must also be escaped as the value for the
10280 @option{text} option in the filter argument string and as the filter
10281 argument in the filtergraph description, and possibly also for the shell,
10282 that makes up to four levels of escaping; using a text file avoids these
10285 The following functions are available:
10290 The expression evaluation result.
10292 It must take one argument specifying the expression to be evaluated,
10293 which accepts the same constants and functions as the @var{x} and
10294 @var{y} values. Note that not all constants should be used, for
10295 example the text size is not known when evaluating the expression, so
10296 the constants @var{text_w} and @var{text_h} will have an undefined
10299 @item expr_int_format, eif
10300 Evaluate the expression's value and output as formatted integer.
10302 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10303 The second argument specifies the output format. Allowed values are @samp{x},
10304 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10305 @code{printf} function.
10306 The third parameter is optional and sets the number of positions taken by the output.
10307 It can be used to add padding with zeros from the left.
10310 The time at which the filter is running, expressed in UTC.
10311 It can accept an argument: a strftime() format string.
10314 The time at which the filter is running, expressed in the local time zone.
10315 It can accept an argument: a strftime() format string.
10318 Frame metadata. Takes one or two arguments.
10320 The first argument is mandatory and specifies the metadata key.
10322 The second argument is optional and specifies a default value, used when the
10323 metadata key is not found or empty.
10325 Available metadata can be identified by inspecting entries
10326 starting with TAG included within each frame section
10327 printed by running @code{ffprobe -show_frames}.
10329 String metadata generated in filters leading to
10330 the drawtext filter are also available.
10333 The frame number, starting from 0.
10336 A one character description of the current picture type.
10339 The timestamp of the current frame.
10340 It can take up to three arguments.
10342 The first argument is the format of the timestamp; it defaults to @code{flt}
10343 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10344 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10345 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10346 @code{localtime} stands for the timestamp of the frame formatted as
10347 local time zone time.
10349 The second argument is an offset added to the timestamp.
10351 If the format is set to @code{hms}, a third argument @code{24HH} may be
10352 supplied to present the hour part of the formatted timestamp in 24h format
10355 If the format is set to @code{localtime} or @code{gmtime},
10356 a third argument may be supplied: a strftime() format string.
10357 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10360 @subsection Commands
10362 This filter supports altering parameters via commands:
10365 Alter existing filter parameters.
10367 Syntax for the argument is the same as for filter invocation, e.g.
10370 fontsize=56:fontcolor=green:text='Hello World'
10373 Full filter invocation with sendcmd would look like this:
10376 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10380 If the entire argument can't be parsed or applied as valid values then the filter will
10381 continue with its existing parameters.
10383 @subsection Examples
10387 Draw "Test Text" with font FreeSerif, using the default values for the
10388 optional parameters.
10391 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10395 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10396 and y=50 (counting from the top-left corner of the screen), text is
10397 yellow with a red box around it. Both the text and the box have an
10401 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10402 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10405 Note that the double quotes are not necessary if spaces are not used
10406 within the parameter list.
10409 Show the text at the center of the video frame:
10411 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10415 Show the text at a random position, switching to a new position every 30 seconds:
10417 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)"
10421 Show a text line sliding from right to left in the last row of the video
10422 frame. The file @file{LONG_LINE} is assumed to contain a single line
10425 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10429 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10431 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10435 Draw a single green letter "g", at the center of the input video.
10436 The glyph baseline is placed at half screen height.
10438 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10442 Show text for 1 second every 3 seconds:
10444 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10448 Use fontconfig to set the font. Note that the colons need to be escaped.
10450 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10454 Draw "Test Text" with font size dependent on height of the video.
10456 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10460 Print the date of a real-time encoding (see strftime(3)):
10462 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10466 Show text fading in and out (appearing/disappearing):
10469 DS=1.0 # display start
10470 DE=10.0 # display end
10471 FID=1.5 # fade in duration
10472 FOD=5 # fade out duration
10473 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 @}"
10477 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10478 and the @option{fontsize} value are included in the @option{y} offset.
10480 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10481 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10485 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10486 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10487 must have option @option{-export_path_metadata 1} for the special metadata fields
10488 to be available for filters.
10490 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10495 For more information about libfreetype, check:
10496 @url{http://www.freetype.org/}.
10498 For more information about fontconfig, check:
10499 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10501 For more information about libfribidi, check:
10502 @url{http://fribidi.org/}.
10504 @section edgedetect
10506 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10508 The filter accepts the following options:
10513 Set low and high threshold values used by the Canny thresholding
10516 The high threshold selects the "strong" edge pixels, which are then
10517 connected through 8-connectivity with the "weak" edge pixels selected
10518 by the low threshold.
10520 @var{low} and @var{high} threshold values must be chosen in the range
10521 [0,1], and @var{low} should be lesser or equal to @var{high}.
10523 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10527 Define the drawing mode.
10531 Draw white/gray wires on black background.
10534 Mix the colors to create a paint/cartoon effect.
10537 Apply Canny edge detector on all selected planes.
10539 Default value is @var{wires}.
10542 Select planes for filtering. By default all available planes are filtered.
10545 @subsection Examples
10549 Standard edge detection with custom values for the hysteresis thresholding:
10551 edgedetect=low=0.1:high=0.4
10555 Painting effect without thresholding:
10557 edgedetect=mode=colormix:high=0
10563 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10565 For each input image, the filter will compute the optimal mapping from
10566 the input to the output given the codebook length, that is the number
10567 of distinct output colors.
10569 This filter accepts the following options.
10572 @item codebook_length, l
10573 Set codebook length. The value must be a positive integer, and
10574 represents the number of distinct output colors. Default value is 256.
10577 Set the maximum number of iterations to apply for computing the optimal
10578 mapping. The higher the value the better the result and the higher the
10579 computation time. Default value is 1.
10582 Set a random seed, must be an integer included between 0 and
10583 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10584 will try to use a good random seed on a best effort basis.
10587 Set pal8 output pixel format. This option does not work with codebook
10588 length greater than 256.
10593 Measure graylevel entropy in histogram of color channels of video frames.
10595 It accepts the following parameters:
10599 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10601 @var{diff} mode measures entropy of histogram delta values, absolute differences
10602 between neighbour histogram values.
10606 Set brightness, contrast, saturation and approximate gamma adjustment.
10608 The filter accepts the following options:
10612 Set the contrast expression. The value must be a float value in range
10613 @code{-1000.0} to @code{1000.0}. The default value is "1".
10616 Set the brightness expression. The value must be a float value in
10617 range @code{-1.0} to @code{1.0}. The default value is "0".
10620 Set the saturation expression. The value must be a float in
10621 range @code{0.0} to @code{3.0}. The default value is "1".
10624 Set the gamma expression. The value must be a float in range
10625 @code{0.1} to @code{10.0}. The default value is "1".
10628 Set the gamma expression for red. The value must be a float in
10629 range @code{0.1} to @code{10.0}. The default value is "1".
10632 Set the gamma expression for green. The value must be a float in range
10633 @code{0.1} to @code{10.0}. The default value is "1".
10636 Set the gamma expression for blue. 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 weight expression. It can be used to reduce the effect
10641 of a high gamma value on bright image areas, e.g. keep them from
10642 getting overamplified and just plain white. The value must be a float
10643 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10644 gamma correction all the way down while @code{1.0} leaves it at its
10645 full strength. Default is "1".
10648 Set when the expressions for brightness, contrast, saturation and
10649 gamma expressions are evaluated.
10651 It accepts the following values:
10654 only evaluate expressions once during the filter initialization or
10655 when a command is processed
10658 evaluate expressions for each incoming frame
10661 Default value is @samp{init}.
10664 The expressions accept the following parameters:
10667 frame count of the input frame starting from 0
10670 byte position of the corresponding packet in the input file, NAN if
10674 frame rate of the input video, NAN if the input frame rate is unknown
10677 timestamp expressed in seconds, NAN if the input timestamp is unknown
10680 @subsection Commands
10681 The filter supports the following commands:
10685 Set the contrast expression.
10688 Set the brightness expression.
10691 Set the saturation expression.
10694 Set the gamma expression.
10697 Set the gamma_r expression.
10700 Set gamma_g expression.
10703 Set gamma_b expression.
10706 Set gamma_weight expression.
10708 The command accepts the same syntax of the corresponding option.
10710 If the specified expression is not valid, it is kept at its current
10717 Apply erosion effect to the video.
10719 This filter replaces the pixel by the local(3x3) minimum.
10721 It accepts the following options:
10728 Limit the maximum change for each plane, default is 65535.
10729 If 0, plane will remain unchanged.
10732 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
10735 Flags to local 3x3 coordinates maps like this:
10742 @subsection Commands
10744 This filter supports the all above options as @ref{commands}.
10746 @section extractplanes
10748 Extract color channel components from input video stream into
10749 separate grayscale video streams.
10751 The filter accepts the following option:
10755 Set plane(s) to extract.
10757 Available values for planes are:
10768 Choosing planes not available in the input will result in an error.
10769 That means you cannot select @code{r}, @code{g}, @code{b} planes
10770 with @code{y}, @code{u}, @code{v} planes at same time.
10773 @subsection Examples
10777 Extract luma, u and v color channel component from input video frame
10778 into 3 grayscale outputs:
10780 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
10786 Apply a fade-in/out effect to the input video.
10788 It accepts the following parameters:
10792 The effect type can be either "in" for a fade-in, or "out" for a fade-out
10794 Default is @code{in}.
10796 @item start_frame, s
10797 Specify the number of the frame to start applying the fade
10798 effect at. Default is 0.
10801 The number of frames that the fade effect lasts. At the end of the
10802 fade-in effect, the output video will have the same intensity as the input video.
10803 At the end of the fade-out transition, the output video will be filled with the
10804 selected @option{color}.
10808 If set to 1, fade only alpha channel, if one exists on the input.
10809 Default value is 0.
10811 @item start_time, st
10812 Specify the timestamp (in seconds) of the frame to start to apply the fade
10813 effect. If both start_frame and start_time are specified, the fade will start at
10814 whichever comes last. Default is 0.
10817 The number of seconds for which the fade effect has to last. At the end of the
10818 fade-in effect the output video will have the same intensity as the input video,
10819 at the end of the fade-out transition the output video will be filled with the
10820 selected @option{color}.
10821 If both duration and nb_frames are specified, duration is used. Default is 0
10822 (nb_frames is used by default).
10825 Specify the color of the fade. Default is "black".
10828 @subsection Examples
10832 Fade in the first 30 frames of video:
10837 The command above is equivalent to:
10843 Fade out the last 45 frames of a 200-frame video:
10846 fade=type=out:start_frame=155:nb_frames=45
10850 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
10852 fade=in:0:25, fade=out:975:25
10856 Make the first 5 frames yellow, then fade in from frame 5-24:
10858 fade=in:5:20:color=yellow
10862 Fade in alpha over first 25 frames of video:
10864 fade=in:0:25:alpha=1
10868 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
10870 fade=t=in:st=5.5:d=0.5
10876 Denoise frames using 3D FFT (frequency domain filtering).
10878 The filter accepts the following options:
10882 Set the noise sigma constant. This sets denoising strength.
10883 Default value is 1. Allowed range is from 0 to 30.
10884 Using very high sigma with low overlap may give blocking artifacts.
10887 Set amount of denoising. By default all detected noise is reduced.
10888 Default value is 1. Allowed range is from 0 to 1.
10891 Set size of block, Default is 4, can be 3, 4, 5 or 6.
10892 Actual size of block in pixels is 2 to power of @var{block}, so by default
10893 block size in pixels is 2^4 which is 16.
10896 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
10899 Set number of previous frames to use for denoising. By default is set to 0.
10902 Set number of next frames to to use for denoising. By default is set to 0.
10905 Set planes which will be filtered, by default are all available filtered
10910 Apply arbitrary expressions to samples in frequency domain
10914 Adjust the dc value (gain) of the luma plane of the image. The filter
10915 accepts an integer value in range @code{0} to @code{1000}. The default
10916 value is set to @code{0}.
10919 Adjust the dc value (gain) of the 1st chroma plane of the image. The
10920 filter accepts an integer value in range @code{0} to @code{1000}. The
10921 default value is set to @code{0}.
10924 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
10925 filter accepts an integer value in range @code{0} to @code{1000}. The
10926 default value is set to @code{0}.
10929 Set the frequency domain weight expression for the luma plane.
10932 Set the frequency domain weight expression for the 1st chroma plane.
10935 Set the frequency domain weight expression for the 2nd chroma plane.
10938 Set when the expressions are evaluated.
10940 It accepts the following values:
10943 Only evaluate expressions once during the filter initialization.
10946 Evaluate expressions for each incoming frame.
10949 Default value is @samp{init}.
10951 The filter accepts the following variables:
10954 The coordinates of the current sample.
10958 The width and height of the image.
10961 The number of input frame, starting from 0.
10964 @subsection Examples
10970 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
10976 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
10982 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
10988 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
10995 Extract a single field from an interlaced image using stride
10996 arithmetic to avoid wasting CPU time. The output frames are marked as
10999 The filter accepts the following options:
11003 Specify whether to extract the top (if the value is @code{0} or
11004 @code{top}) or the bottom field (if the value is @code{1} or
11010 Create new frames by copying the top and bottom fields from surrounding frames
11011 supplied as numbers by the hint file.
11015 Set file containing hints: absolute/relative frame numbers.
11017 There must be one line for each frame in a clip. Each line must contain two
11018 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11019 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11020 is current frame number for @code{absolute} mode or out of [-1, 1] range
11021 for @code{relative} mode. First number tells from which frame to pick up top
11022 field and second number tells from which frame to pick up bottom field.
11024 If optionally followed by @code{+} output frame will be marked as interlaced,
11025 else if followed by @code{-} output frame will be marked as progressive, else
11026 it will be marked same as input frame.
11027 If optionally followed by @code{t} output frame will use only top field, or in
11028 case of @code{b} it will use only bottom field.
11029 If line starts with @code{#} or @code{;} that line is skipped.
11032 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11035 Example of first several lines of @code{hint} file for @code{relative} mode:
11037 0,0 - # first frame
11038 1,0 - # second frame, use third's frame top field and second's frame bottom field
11039 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11054 @section fieldmatch
11056 Field matching filter for inverse telecine. It is meant to reconstruct the
11057 progressive frames from a telecined stream. The filter does not drop duplicated
11058 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11059 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11061 The separation of the field matching and the decimation is notably motivated by
11062 the possibility of inserting a de-interlacing filter fallback between the two.
11063 If the source has mixed telecined and real interlaced content,
11064 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11065 But these remaining combed frames will be marked as interlaced, and thus can be
11066 de-interlaced by a later filter such as @ref{yadif} before decimation.
11068 In addition to the various configuration options, @code{fieldmatch} can take an
11069 optional second stream, activated through the @option{ppsrc} option. If
11070 enabled, the frames reconstruction will be based on the fields and frames from
11071 this second stream. This allows the first input to be pre-processed in order to
11072 help the various algorithms of the filter, while keeping the output lossless
11073 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11074 or brightness/contrast adjustments can help.
11076 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11077 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11078 which @code{fieldmatch} is based on. While the semantic and usage are very
11079 close, some behaviour and options names can differ.
11081 The @ref{decimate} filter currently only works for constant frame rate input.
11082 If your input has mixed telecined (30fps) and progressive content with a lower
11083 framerate like 24fps use the following filterchain to produce the necessary cfr
11084 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11086 The filter accepts the following options:
11090 Specify the assumed field order of the input stream. Available values are:
11094 Auto detect parity (use FFmpeg's internal parity value).
11096 Assume bottom field first.
11098 Assume top field first.
11101 Note that it is sometimes recommended not to trust the parity announced by the
11104 Default value is @var{auto}.
11107 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11108 sense that it won't risk creating jerkiness due to duplicate frames when
11109 possible, but if there are bad edits or blended fields it will end up
11110 outputting combed frames when a good match might actually exist. On the other
11111 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11112 but will almost always find a good frame if there is one. The other values are
11113 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11114 jerkiness and creating duplicate frames versus finding good matches in sections
11115 with bad edits, orphaned fields, blended fields, etc.
11117 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11119 Available values are:
11123 2-way matching (p/c)
11125 2-way matching, and trying 3rd match if still combed (p/c + n)
11127 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11129 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11130 still combed (p/c + n + u/b)
11132 3-way matching (p/c/n)
11134 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11135 detected as combed (p/c/n + u/b)
11138 The parenthesis at the end indicate the matches that would be used for that
11139 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11142 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11145 Default value is @var{pc_n}.
11148 Mark the main input stream as a pre-processed input, and enable the secondary
11149 input stream as the clean source to pick the fields from. See the filter
11150 introduction for more details. It is similar to the @option{clip2} feature from
11153 Default value is @code{0} (disabled).
11156 Set the field to match from. It is recommended to set this to the same value as
11157 @option{order} unless you experience matching failures with that setting. In
11158 certain circumstances changing the field that is used to match from can have a
11159 large impact on matching performance. Available values are:
11163 Automatic (same value as @option{order}).
11165 Match from the bottom field.
11167 Match from the top field.
11170 Default value is @var{auto}.
11173 Set whether or not chroma is included during the match comparisons. In most
11174 cases it is recommended to leave this enabled. You should set this to @code{0}
11175 only if your clip has bad chroma problems such as heavy rainbowing or other
11176 artifacts. Setting this to @code{0} could also be used to speed things up at
11177 the cost of some accuracy.
11179 Default value is @code{1}.
11183 These define an exclusion band which excludes the lines between @option{y0} and
11184 @option{y1} from being included in the field matching decision. An exclusion
11185 band can be used to ignore subtitles, a logo, or other things that may
11186 interfere with the matching. @option{y0} sets the starting scan line and
11187 @option{y1} sets the ending line; all lines in between @option{y0} and
11188 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11189 @option{y0} and @option{y1} to the same value will disable the feature.
11190 @option{y0} and @option{y1} defaults to @code{0}.
11193 Set the scene change detection threshold as a percentage of maximum change on
11194 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11195 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11196 @option{scthresh} is @code{[0.0, 100.0]}.
11198 Default value is @code{12.0}.
11201 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11202 account the combed scores of matches when deciding what match to use as the
11203 final match. Available values are:
11207 No final matching based on combed scores.
11209 Combed scores are only used when a scene change is detected.
11211 Use combed scores all the time.
11214 Default is @var{sc}.
11217 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11218 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11219 Available values are:
11223 No forced calculation.
11225 Force p/c/n calculations.
11227 Force p/c/n/u/b calculations.
11230 Default value is @var{none}.
11233 This is the area combing threshold used for combed frame detection. This
11234 essentially controls how "strong" or "visible" combing must be to be detected.
11235 Larger values mean combing must be more visible and smaller values mean combing
11236 can be less visible or strong and still be detected. Valid settings are from
11237 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11238 be detected as combed). This is basically a pixel difference value. A good
11239 range is @code{[8, 12]}.
11241 Default value is @code{9}.
11244 Sets whether or not chroma is considered in the combed frame decision. Only
11245 disable this if your source has chroma problems (rainbowing, etc.) that are
11246 causing problems for the combed frame detection with chroma enabled. Actually,
11247 using @option{chroma}=@var{0} is usually more reliable, except for the case
11248 where there is chroma only combing in the source.
11250 Default value is @code{0}.
11254 Respectively set the x-axis and y-axis size of the window used during combed
11255 frame detection. This has to do with the size of the area in which
11256 @option{combpel} pixels are required to be detected as combed for a frame to be
11257 declared combed. See the @option{combpel} parameter description for more info.
11258 Possible values are any number that is a power of 2 starting at 4 and going up
11261 Default value is @code{16}.
11264 The number of combed pixels inside any of the @option{blocky} by
11265 @option{blockx} size blocks on the frame for the frame to be detected as
11266 combed. While @option{cthresh} controls how "visible" the combing must be, this
11267 setting controls "how much" combing there must be in any localized area (a
11268 window defined by the @option{blockx} and @option{blocky} settings) on the
11269 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11270 which point no frames will ever be detected as combed). This setting is known
11271 as @option{MI} in TFM/VFM vocabulary.
11273 Default value is @code{80}.
11276 @anchor{p/c/n/u/b meaning}
11277 @subsection p/c/n/u/b meaning
11279 @subsubsection p/c/n
11281 We assume the following telecined stream:
11284 Top fields: 1 2 2 3 4
11285 Bottom fields: 1 2 3 4 4
11288 The numbers correspond to the progressive frame the fields relate to. Here, the
11289 first two frames are progressive, the 3rd and 4th are combed, and so on.
11291 When @code{fieldmatch} is configured to run a matching from bottom
11292 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11297 B 1 2 3 4 4 <-- matching reference
11306 As a result of the field matching, we can see that some frames get duplicated.
11307 To perform a complete inverse telecine, you need to rely on a decimation filter
11308 after this operation. See for instance the @ref{decimate} filter.
11310 The same operation now matching from top fields (@option{field}=@var{top})
11315 T 1 2 2 3 4 <-- matching reference
11325 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11326 basically, they refer to the frame and field of the opposite parity:
11329 @item @var{p} matches the field of the opposite parity in the previous frame
11330 @item @var{c} matches the field of the opposite parity in the current frame
11331 @item @var{n} matches the field of the opposite parity in the next frame
11336 The @var{u} and @var{b} matching are a bit special in the sense that they match
11337 from the opposite parity flag. In the following examples, we assume that we are
11338 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11339 'x' is placed above and below each matched fields.
11341 With bottom matching (@option{field}=@var{bottom}):
11346 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11347 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11355 With top matching (@option{field}=@var{top}):
11360 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11361 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11369 @subsection Examples
11371 Simple IVTC of a top field first telecined stream:
11373 fieldmatch=order=tff:combmatch=none, decimate
11376 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11378 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11381 @section fieldorder
11383 Transform the field order of the input video.
11385 It accepts the following parameters:
11390 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11391 for bottom field first.
11394 The default value is @samp{tff}.
11396 The transformation is done by shifting the picture content up or down
11397 by one line, and filling the remaining line with appropriate picture content.
11398 This method is consistent with most broadcast field order converters.
11400 If the input video is not flagged as being interlaced, or it is already
11401 flagged as being of the required output field order, then this filter does
11402 not alter the incoming video.
11404 It is very useful when converting to or from PAL DV material,
11405 which is bottom field first.
11409 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11412 @section fifo, afifo
11414 Buffer input images and send them when they are requested.
11416 It is mainly useful when auto-inserted by the libavfilter
11419 It does not take parameters.
11421 @section fillborders
11423 Fill borders of the input video, without changing video stream dimensions.
11424 Sometimes video can have garbage at the four edges and you may not want to
11425 crop video input to keep size multiple of some number.
11427 This filter accepts the following options:
11431 Number of pixels to fill from left border.
11434 Number of pixels to fill from right border.
11437 Number of pixels to fill from top border.
11440 Number of pixels to fill from bottom border.
11445 It accepts the following values:
11448 fill pixels using outermost pixels
11451 fill pixels using mirroring
11454 fill pixels with constant value
11457 Default is @var{smear}.
11460 Set color for pixels in fixed mode. Default is @var{black}.
11463 @subsection Commands
11464 This filter supports same @ref{commands} as options.
11465 The command accepts the same syntax of the corresponding option.
11467 If the specified expression is not valid, it is kept at its current
11472 Find a rectangular object
11474 It accepts the following options:
11478 Filepath of the object image, needs to be in gray8.
11481 Detection threshold, default is 0.5.
11484 Number of mipmaps, default is 3.
11486 @item xmin, ymin, xmax, ymax
11487 Specifies the rectangle in which to search.
11490 @subsection Examples
11494 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11496 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11502 Flood area with values of same pixel components with another values.
11504 It accepts the following options:
11507 Set pixel x coordinate.
11510 Set pixel y coordinate.
11513 Set source #0 component value.
11516 Set source #1 component value.
11519 Set source #2 component value.
11522 Set source #3 component value.
11525 Set destination #0 component value.
11528 Set destination #1 component value.
11531 Set destination #2 component value.
11534 Set destination #3 component value.
11540 Convert the input video to one of the specified pixel formats.
11541 Libavfilter will try to pick one that is suitable as input to
11544 It accepts the following parameters:
11548 A '|'-separated list of pixel format names, such as
11549 "pix_fmts=yuv420p|monow|rgb24".
11553 @subsection Examples
11557 Convert the input video to the @var{yuv420p} format
11559 format=pix_fmts=yuv420p
11562 Convert the input video to any of the formats in the list
11564 format=pix_fmts=yuv420p|yuv444p|yuv410p
11571 Convert the video to specified constant frame rate by duplicating or dropping
11572 frames as necessary.
11574 It accepts the following parameters:
11578 The desired output frame rate. The default is @code{25}.
11581 Assume the first PTS should be the given value, in seconds. This allows for
11582 padding/trimming at the start of stream. By default, no assumption is made
11583 about the first frame's expected PTS, so no padding or trimming is done.
11584 For example, this could be set to 0 to pad the beginning with duplicates of
11585 the first frame if a video stream starts after the audio stream or to trim any
11586 frames with a negative PTS.
11589 Timestamp (PTS) rounding method.
11591 Possible values are:
11598 round towards -infinity
11600 round towards +infinity
11604 The default is @code{near}.
11607 Action performed when reading the last frame.
11609 Possible values are:
11612 Use same timestamp rounding method as used for other frames.
11614 Pass through last frame if input duration has not been reached yet.
11616 The default is @code{round}.
11620 Alternatively, the options can be specified as a flat string:
11621 @var{fps}[:@var{start_time}[:@var{round}]].
11623 See also the @ref{setpts} filter.
11625 @subsection Examples
11629 A typical usage in order to set the fps to 25:
11635 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11637 fps=fps=film:round=near
11643 Pack two different video streams into a stereoscopic video, setting proper
11644 metadata on supported codecs. The two views should have the same size and
11645 framerate and processing will stop when the shorter video ends. Please note
11646 that you may conveniently adjust view properties with the @ref{scale} and
11649 It accepts the following parameters:
11653 The desired packing format. Supported values are:
11658 The views are next to each other (default).
11661 The views are on top of each other.
11664 The views are packed by line.
11667 The views are packed by column.
11670 The views are temporally interleaved.
11679 # Convert left and right views into a frame-sequential video
11680 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11682 # Convert views into a side-by-side video with the same output resolution as the input
11683 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
11688 Change the frame rate by interpolating new video output frames from the source
11691 This filter is not designed to function correctly with interlaced media. If
11692 you wish to change the frame rate of interlaced media then you are required
11693 to deinterlace before this filter and re-interlace after this filter.
11695 A description of the accepted options follows.
11699 Specify the output frames per second. This option can also be specified
11700 as a value alone. The default is @code{50}.
11703 Specify the start of a range where the output frame will be created as a
11704 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11705 the default is @code{15}.
11708 Specify the end of a range where the output frame will be created as a
11709 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11710 the default is @code{240}.
11713 Specify the level at which a scene change is detected as a value between
11714 0 and 100 to indicate a new scene; a low value reflects a low
11715 probability for the current frame to introduce a new scene, while a higher
11716 value means the current frame is more likely to be one.
11717 The default is @code{8.2}.
11720 Specify flags influencing the filter process.
11722 Available value for @var{flags} is:
11725 @item scene_change_detect, scd
11726 Enable scene change detection using the value of the option @var{scene}.
11727 This flag is enabled by default.
11733 Select one frame every N-th frame.
11735 This filter accepts the following option:
11738 Select frame after every @code{step} frames.
11739 Allowed values are positive integers higher than 0. Default value is @code{1}.
11742 @section freezedetect
11744 Detect frozen video.
11746 This filter logs a message and sets frame metadata when it detects that the
11747 input video has no significant change in content during a specified duration.
11748 Video freeze detection calculates the mean average absolute difference of all
11749 the components of video frames and compares it to a noise floor.
11751 The printed times and duration are expressed in seconds. The
11752 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
11753 whose timestamp equals or exceeds the detection duration and it contains the
11754 timestamp of the first frame of the freeze. The
11755 @code{lavfi.freezedetect.freeze_duration} and
11756 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
11759 The filter accepts the following options:
11763 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
11764 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
11768 Set freeze duration until notification (default is 2 seconds).
11771 @section freezeframes
11773 Freeze video frames.
11775 This filter freezes video frames using frame from 2nd input.
11777 The filter accepts the following options:
11781 Set number of first frame from which to start freeze.
11784 Set number of last frame from which to end freeze.
11787 Set number of frame from 2nd input which will be used instead of replaced frames.
11793 Apply a frei0r effect to the input video.
11795 To enable the compilation of this filter, you need to install the frei0r
11796 header and configure FFmpeg with @code{--enable-frei0r}.
11798 It accepts the following parameters:
11803 The name of the frei0r effect to load. If the environment variable
11804 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
11805 directories specified by the colon-separated list in @env{FREI0R_PATH}.
11806 Otherwise, the standard frei0r paths are searched, in this order:
11807 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
11808 @file{/usr/lib/frei0r-1/}.
11810 @item filter_params
11811 A '|'-separated list of parameters to pass to the frei0r effect.
11815 A frei0r effect parameter can be a boolean (its value is either
11816 "y" or "n"), a double, a color (specified as
11817 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
11818 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
11819 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
11820 a position (specified as @var{X}/@var{Y}, where
11821 @var{X} and @var{Y} are floating point numbers) and/or a string.
11823 The number and types of parameters depend on the loaded effect. If an
11824 effect parameter is not specified, the default value is set.
11826 @subsection Examples
11830 Apply the distort0r effect, setting the first two double parameters:
11832 frei0r=filter_name=distort0r:filter_params=0.5|0.01
11836 Apply the colordistance effect, taking a color as the first parameter:
11838 frei0r=colordistance:0.2/0.3/0.4
11839 frei0r=colordistance:violet
11840 frei0r=colordistance:0x112233
11844 Apply the perspective effect, specifying the top left and top right image
11847 frei0r=perspective:0.2/0.2|0.8/0.2
11851 For more information, see
11852 @url{http://frei0r.dyne.org}
11854 @subsection Commands
11856 This filter supports the @option{filter_params} option as @ref{commands}.
11860 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
11862 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
11863 processing filter, one of them is performed once per block, not per pixel.
11864 This allows for much higher speed.
11866 The filter accepts the following options:
11870 Set quality. This option defines the number of levels for averaging. It accepts
11871 an integer in the range 4-5. Default value is @code{4}.
11874 Force a constant quantization parameter. It accepts an integer in range 0-63.
11875 If not set, the filter will use the QP from the video stream (if available).
11878 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
11879 more details but also more artifacts, while higher values make the image smoother
11880 but also blurrier. Default value is @code{0} − PSNR optimal.
11882 @item use_bframe_qp
11883 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
11884 option may cause flicker since the B-Frames have often larger QP. Default is
11885 @code{0} (not enabled).
11891 Apply Gaussian blur filter.
11893 The filter accepts the following options:
11897 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
11900 Set number of steps for Gaussian approximation. Default is @code{1}.
11903 Set which planes to filter. By default all planes are filtered.
11906 Set vertical sigma, if negative it will be same as @code{sigma}.
11907 Default is @code{-1}.
11910 @subsection Commands
11911 This filter supports same commands as options.
11912 The command accepts the same syntax of the corresponding option.
11914 If the specified expression is not valid, it is kept at its current
11919 Apply generic equation to each pixel.
11921 The filter accepts the following options:
11924 @item lum_expr, lum
11925 Set the luminance expression.
11927 Set the chrominance blue expression.
11929 Set the chrominance red expression.
11930 @item alpha_expr, a
11931 Set the alpha expression.
11933 Set the red expression.
11934 @item green_expr, g
11935 Set the green expression.
11937 Set the blue expression.
11940 The colorspace is selected according to the specified options. If one
11941 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
11942 options is specified, the filter will automatically select a YCbCr
11943 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
11944 @option{blue_expr} options is specified, it will select an RGB
11947 If one of the chrominance expression is not defined, it falls back on the other
11948 one. If no alpha expression is specified it will evaluate to opaque value.
11949 If none of chrominance expressions are specified, they will evaluate
11950 to the luminance expression.
11952 The expressions can use the following variables and functions:
11956 The sequential number of the filtered frame, starting from @code{0}.
11960 The coordinates of the current sample.
11964 The width and height of the image.
11968 Width and height scale depending on the currently filtered plane. It is the
11969 ratio between the corresponding luma plane number of pixels and the current
11970 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
11971 @code{0.5,0.5} for chroma planes.
11974 Time of the current frame, expressed in seconds.
11977 Return the value of the pixel at location (@var{x},@var{y}) of the current
11981 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
11985 Return the value of the pixel at location (@var{x},@var{y}) of the
11986 blue-difference chroma plane. Return 0 if there is no such plane.
11989 Return the value of the pixel at location (@var{x},@var{y}) of the
11990 red-difference chroma plane. Return 0 if there is no such plane.
11995 Return the value of the pixel at location (@var{x},@var{y}) of the
11996 red/green/blue component. Return 0 if there is no such component.
11999 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12000 plane. Return 0 if there is no such plane.
12002 @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)
12003 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12004 sums of samples within a rectangle. See the functions without the sum postfix.
12006 @item interpolation
12007 Set one of interpolation methods:
12012 Default is bilinear.
12015 For functions, if @var{x} and @var{y} are outside the area, the value will be
12016 automatically clipped to the closer edge.
12018 Please note that this filter can use multiple threads in which case each slice
12019 will have its own expression state. If you want to use only a single expression
12020 state because your expressions depend on previous state then you should limit
12021 the number of filter threads to 1.
12023 @subsection Examples
12027 Flip the image horizontally:
12033 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12034 wavelength of 100 pixels:
12036 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12040 Generate a fancy enigmatic moving light:
12042 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
12046 Generate a quick emboss effect:
12048 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12052 Modify RGB components depending on pixel position:
12054 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12058 Create a radial gradient that is the same size as the input (also see
12059 the @ref{vignette} filter):
12061 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12067 Fix the banding artifacts that are sometimes introduced into nearly flat
12068 regions by truncation to 8-bit color depth.
12069 Interpolate the gradients that should go where the bands are, and
12072 It is designed for playback only. Do not use it prior to
12073 lossy compression, because compression tends to lose the dither and
12074 bring back the bands.
12076 It accepts the following parameters:
12081 The maximum amount by which the filter will change any one pixel. This is also
12082 the threshold for detecting nearly flat regions. Acceptable values range from
12083 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12087 The neighborhood to fit the gradient to. A larger radius makes for smoother
12088 gradients, but also prevents the filter from modifying the pixels near detailed
12089 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12090 values will be clipped to the valid range.
12094 Alternatively, the options can be specified as a flat string:
12095 @var{strength}[:@var{radius}]
12097 @subsection Examples
12101 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12107 Specify radius, omitting the strength (which will fall-back to the default
12115 @anchor{graphmonitor}
12116 @section graphmonitor
12117 Show various filtergraph stats.
12119 With this filter one can debug complete filtergraph.
12120 Especially issues with links filling with queued frames.
12122 The filter accepts the following options:
12126 Set video output size. Default is @var{hd720}.
12129 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12132 Set output mode, can be @var{fulll} or @var{compact}.
12133 In @var{compact} mode only filters with some queued frames have displayed stats.
12136 Set flags which enable which stats are shown in video.
12138 Available values for flags are:
12141 Display number of queued frames in each link.
12143 @item frame_count_in
12144 Display number of frames taken from filter.
12146 @item frame_count_out
12147 Display number of frames given out from filter.
12150 Display current filtered frame pts.
12153 Display current filtered frame time.
12156 Display time base for filter link.
12159 Display used format for filter link.
12162 Display video size or number of audio channels in case of audio used by filter link.
12165 Display video frame rate or sample rate in case of audio used by filter link.
12168 Display link output status.
12172 Set upper limit for video rate of output stream, Default value is @var{25}.
12173 This guarantee that output video frame rate will not be higher than this value.
12177 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12178 and corrects the scene colors accordingly.
12180 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12182 The filter accepts the following options:
12186 The order of differentiation to be applied on the scene. Must be chosen in the range
12187 [0,2] and default value is 1.
12190 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12191 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12192 max value instead of calculating Minkowski distance.
12195 The standard deviation of Gaussian blur to be applied on the scene. Must be
12196 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12197 can't be equal to 0 if @var{difford} is greater than 0.
12200 @subsection Examples
12206 greyedge=difford=1:minknorm=5:sigma=2
12212 greyedge=difford=1:minknorm=0:sigma=2
12220 Apply a Hald CLUT to a video stream.
12222 First input is the video stream to process, and second one is the Hald CLUT.
12223 The Hald CLUT input can be a simple picture or a complete video stream.
12225 The filter accepts the following options:
12229 Force termination when the shortest input terminates. Default is @code{0}.
12231 Continue applying the last CLUT after the end of the stream. A value of
12232 @code{0} disable the filter after the last frame of the CLUT is reached.
12233 Default is @code{1}.
12236 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12237 filters share the same internals).
12239 This filter also supports the @ref{framesync} options.
12241 More information about the Hald CLUT can be found on Eskil Steenberg's website
12242 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12244 @subsection Workflow examples
12246 @subsubsection Hald CLUT video stream
12248 Generate an identity Hald CLUT stream altered with various effects:
12250 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
12253 Note: make sure you use a lossless codec.
12255 Then use it with @code{haldclut} to apply it on some random stream:
12257 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12260 The Hald CLUT will be applied to the 10 first seconds (duration of
12261 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12262 to the remaining frames of the @code{mandelbrot} stream.
12264 @subsubsection Hald CLUT with preview
12266 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12267 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12268 biggest possible square starting at the top left of the picture. The remaining
12269 padding pixels (bottom or right) will be ignored. This area can be used to add
12270 a preview of the Hald CLUT.
12272 Typically, the following generated Hald CLUT will be supported by the
12273 @code{haldclut} filter:
12276 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12277 pad=iw+320 [padded_clut];
12278 smptebars=s=320x256, split [a][b];
12279 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12280 [main][b] overlay=W-320" -frames:v 1 clut.png
12283 It contains the original and a preview of the effect of the CLUT: SMPTE color
12284 bars are displayed on the right-top, and below the same color bars processed by
12287 Then, the effect of this Hald CLUT can be visualized with:
12289 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12294 Flip the input video horizontally.
12296 For example, to horizontally flip the input video with @command{ffmpeg}:
12298 ffmpeg -i in.avi -vf "hflip" out.avi
12302 This filter applies a global color histogram equalization on a
12305 It can be used to correct video that has a compressed range of pixel
12306 intensities. The filter redistributes the pixel intensities to
12307 equalize their distribution across the intensity range. It may be
12308 viewed as an "automatically adjusting contrast filter". This filter is
12309 useful only for correcting degraded or poorly captured source
12312 The filter accepts the following options:
12316 Determine the amount of equalization to be applied. As the strength
12317 is reduced, the distribution of pixel intensities more-and-more
12318 approaches that of the input frame. The value must be a float number
12319 in the range [0,1] and defaults to 0.200.
12322 Set the maximum intensity that can generated and scale the output
12323 values appropriately. The strength should be set as desired and then
12324 the intensity can be limited if needed to avoid washing-out. The value
12325 must be a float number in the range [0,1] and defaults to 0.210.
12328 Set the antibanding level. If enabled the filter will randomly vary
12329 the luminance of output pixels by a small amount to avoid banding of
12330 the histogram. Possible values are @code{none}, @code{weak} or
12331 @code{strong}. It defaults to @code{none}.
12337 Compute and draw a color distribution histogram for the input video.
12339 The computed histogram is a representation of the color component
12340 distribution in an image.
12342 Standard histogram displays the color components distribution in an image.
12343 Displays color graph for each color component. Shows distribution of
12344 the Y, U, V, A or R, G, B components, depending on input format, in the
12345 current frame. Below each graph a color component scale meter is shown.
12347 The filter accepts the following options:
12351 Set height of level. Default value is @code{200}.
12352 Allowed range is [50, 2048].
12355 Set height of color scale. Default value is @code{12}.
12356 Allowed range is [0, 40].
12360 It accepts the following values:
12363 Per color component graphs are placed below each other.
12366 Per color component graphs are placed side by side.
12369 Presents information identical to that in the @code{parade}, except
12370 that the graphs representing color components are superimposed directly
12373 Default is @code{stack}.
12376 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12377 Default is @code{linear}.
12380 Set what color components to display.
12381 Default is @code{7}.
12384 Set foreground opacity. Default is @code{0.7}.
12387 Set background opacity. Default is @code{0.5}.
12390 @subsection Examples
12395 Calculate and draw histogram:
12397 ffplay -i input -vf histogram
12405 This is a high precision/quality 3d denoise filter. It aims to reduce
12406 image noise, producing smooth images and making still images really
12407 still. It should enhance compressibility.
12409 It accepts the following optional parameters:
12413 A non-negative floating point number which specifies spatial luma strength.
12414 It defaults to 4.0.
12416 @item chroma_spatial
12417 A non-negative floating point number which specifies spatial chroma strength.
12418 It defaults to 3.0*@var{luma_spatial}/4.0.
12421 A floating point number which specifies luma temporal strength. It defaults to
12422 6.0*@var{luma_spatial}/4.0.
12425 A floating point number which specifies chroma temporal strength. It defaults to
12426 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12429 @subsection Commands
12430 This filter supports same @ref{commands} as options.
12431 The command accepts the same syntax of the corresponding option.
12433 If the specified expression is not valid, it is kept at its current
12436 @anchor{hwdownload}
12437 @section hwdownload
12439 Download hardware frames to system memory.
12441 The input must be in hardware frames, and the output a non-hardware format.
12442 Not all formats will be supported on the output - it may be necessary to insert
12443 an additional @option{format} filter immediately following in the graph to get
12444 the output in a supported format.
12448 Map hardware frames to system memory or to another device.
12450 This filter has several different modes of operation; which one is used depends
12451 on the input and output formats:
12454 Hardware frame input, normal frame output
12456 Map the input frames to system memory and pass them to the output. If the
12457 original hardware frame is later required (for example, after overlaying
12458 something else on part of it), the @option{hwmap} filter can be used again
12459 in the next mode to retrieve it.
12461 Normal frame input, hardware frame output
12463 If the input is actually a software-mapped hardware frame, then unmap it -
12464 that is, return the original hardware frame.
12466 Otherwise, a device must be provided. Create new hardware surfaces on that
12467 device for the output, then map them back to the software format at the input
12468 and give those frames to the preceding filter. This will then act like the
12469 @option{hwupload} filter, but may be able to avoid an additional copy when
12470 the input is already in a compatible format.
12472 Hardware frame input and output
12474 A device must be supplied for the output, either directly or with the
12475 @option{derive_device} option. The input and output devices must be of
12476 different types and compatible - the exact meaning of this is
12477 system-dependent, but typically it means that they must refer to the same
12478 underlying hardware context (for example, refer to the same graphics card).
12480 If the input frames were originally created on the output device, then unmap
12481 to retrieve the original frames.
12483 Otherwise, map the frames to the output device - create new hardware frames
12484 on the output corresponding to the frames on the input.
12487 The following additional parameters are accepted:
12491 Set the frame mapping mode. Some combination of:
12494 The mapped frame should be readable.
12496 The mapped frame should be writeable.
12498 The mapping will always overwrite the entire frame.
12500 This may improve performance in some cases, as the original contents of the
12501 frame need not be loaded.
12503 The mapping must not involve any copying.
12505 Indirect mappings to copies of frames are created in some cases where either
12506 direct mapping is not possible or it would have unexpected properties.
12507 Setting this flag ensures that the mapping is direct and will fail if that is
12510 Defaults to @var{read+write} if not specified.
12512 @item derive_device @var{type}
12513 Rather than using the device supplied at initialisation, instead derive a new
12514 device of type @var{type} from the device the input frames exist on.
12517 In a hardware to hardware mapping, map in reverse - create frames in the sink
12518 and map them back to the source. This may be necessary in some cases where
12519 a mapping in one direction is required but only the opposite direction is
12520 supported by the devices being used.
12522 This option is dangerous - it may break the preceding filter in undefined
12523 ways if there are any additional constraints on that filter's output.
12524 Do not use it without fully understanding the implications of its use.
12530 Upload system memory frames to hardware surfaces.
12532 The device to upload to must be supplied when the filter is initialised. If
12533 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12534 option or with the @option{derive_device} option. The input and output devices
12535 must be of different types and compatible - the exact meaning of this is
12536 system-dependent, but typically it means that they must refer to the same
12537 underlying hardware context (for example, refer to the same graphics card).
12539 The following additional parameters are accepted:
12542 @item derive_device @var{type}
12543 Rather than using the device supplied at initialisation, instead derive a new
12544 device of type @var{type} from the device the input frames exist on.
12547 @anchor{hwupload_cuda}
12548 @section hwupload_cuda
12550 Upload system memory frames to a CUDA device.
12552 It accepts the following optional parameters:
12556 The number of the CUDA device to use
12561 Apply a high-quality magnification filter designed for pixel art. This filter
12562 was originally created by Maxim Stepin.
12564 It accepts the following option:
12568 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12569 @code{hq3x} and @code{4} for @code{hq4x}.
12570 Default is @code{3}.
12574 Stack input videos horizontally.
12576 All streams must be of same pixel format and of same height.
12578 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12579 to create same output.
12581 The filter accepts the following option:
12585 Set number of input streams. Default is 2.
12588 If set to 1, force the output to terminate when the shortest input
12589 terminates. Default value is 0.
12594 Modify the hue and/or the saturation of the input.
12596 It accepts the following parameters:
12600 Specify the hue angle as a number of degrees. It accepts an expression,
12601 and defaults to "0".
12604 Specify the saturation in the [-10,10] range. It accepts an expression and
12608 Specify the hue angle as a number of radians. It accepts an
12609 expression, and defaults to "0".
12612 Specify the brightness in the [-10,10] range. It accepts an expression and
12616 @option{h} and @option{H} are mutually exclusive, and can't be
12617 specified at the same time.
12619 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12620 expressions containing the following constants:
12624 frame count of the input frame starting from 0
12627 presentation timestamp of the input frame expressed in time base units
12630 frame rate of the input video, NAN if the input frame rate is unknown
12633 timestamp expressed in seconds, NAN if the input timestamp is unknown
12636 time base of the input video
12639 @subsection Examples
12643 Set the hue to 90 degrees and the saturation to 1.0:
12649 Same command but expressing the hue in radians:
12655 Rotate hue and make the saturation swing between 0
12656 and 2 over a period of 1 second:
12658 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12662 Apply a 3 seconds saturation fade-in effect starting at 0:
12664 hue="s=min(t/3\,1)"
12667 The general fade-in expression can be written as:
12669 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12673 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12675 hue="s=max(0\, min(1\, (8-t)/3))"
12678 The general fade-out expression can be written as:
12680 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12685 @subsection Commands
12687 This filter supports the following commands:
12693 Modify the hue and/or the saturation and/or brightness of the input video.
12694 The command accepts the same syntax of the corresponding option.
12696 If the specified expression is not valid, it is kept at its current
12700 @section hysteresis
12702 Grow first stream into second stream by connecting components.
12703 This makes it possible to build more robust edge masks.
12705 This filter accepts the following options:
12709 Set which planes will be processed as bitmap, unprocessed planes will be
12710 copied from first stream.
12711 By default value 0xf, all planes will be processed.
12714 Set threshold which is used in filtering. If pixel component value is higher than
12715 this value filter algorithm for connecting components is activated.
12716 By default value is 0.
12719 The @code{hysteresis} filter also supports the @ref{framesync} options.
12723 Detect video interlacing type.
12725 This filter tries to detect if the input frames are interlaced, progressive,
12726 top or bottom field first. It will also try to detect fields that are
12727 repeated between adjacent frames (a sign of telecine).
12729 Single frame detection considers only immediately adjacent frames when classifying each frame.
12730 Multiple frame detection incorporates the classification history of previous frames.
12732 The filter will log these metadata values:
12735 @item single.current_frame
12736 Detected type of current frame using single-frame detection. One of:
12737 ``tff'' (top field first), ``bff'' (bottom field first),
12738 ``progressive'', or ``undetermined''
12741 Cumulative number of frames detected as top field first using single-frame detection.
12744 Cumulative number of frames detected as top field first using multiple-frame detection.
12747 Cumulative number of frames detected as bottom field first using single-frame detection.
12749 @item multiple.current_frame
12750 Detected type of current frame using multiple-frame detection. One of:
12751 ``tff'' (top field first), ``bff'' (bottom field first),
12752 ``progressive'', or ``undetermined''
12755 Cumulative number of frames detected as bottom field first using multiple-frame detection.
12757 @item single.progressive
12758 Cumulative number of frames detected as progressive using single-frame detection.
12760 @item multiple.progressive
12761 Cumulative number of frames detected as progressive using multiple-frame detection.
12763 @item single.undetermined
12764 Cumulative number of frames that could not be classified using single-frame detection.
12766 @item multiple.undetermined
12767 Cumulative number of frames that could not be classified using multiple-frame detection.
12769 @item repeated.current_frame
12770 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
12772 @item repeated.neither
12773 Cumulative number of frames with no repeated field.
12776 Cumulative number of frames with the top field repeated from the previous frame's top field.
12778 @item repeated.bottom
12779 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
12782 The filter accepts the following options:
12786 Set interlacing threshold.
12788 Set progressive threshold.
12790 Threshold for repeated field detection.
12792 Number of frames after which a given frame's contribution to the
12793 statistics is halved (i.e., it contributes only 0.5 to its
12794 classification). The default of 0 means that all frames seen are given
12795 full weight of 1.0 forever.
12796 @item analyze_interlaced_flag
12797 When this is not 0 then idet will use the specified number of frames to determine
12798 if the interlaced flag is accurate, it will not count undetermined frames.
12799 If the flag is found to be accurate it will be used without any further
12800 computations, if it is found to be inaccurate it will be cleared without any
12801 further computations. This allows inserting the idet filter as a low computational
12802 method to clean up the interlaced flag
12807 Deinterleave or interleave fields.
12809 This filter allows one to process interlaced images fields without
12810 deinterlacing them. Deinterleaving splits the input frame into 2
12811 fields (so called half pictures). Odd lines are moved to the top
12812 half of the output image, even lines to the bottom half.
12813 You can process (filter) them independently and then re-interleave them.
12815 The filter accepts the following options:
12819 @item chroma_mode, c
12820 @item alpha_mode, a
12821 Available values for @var{luma_mode}, @var{chroma_mode} and
12822 @var{alpha_mode} are:
12828 @item deinterleave, d
12829 Deinterleave fields, placing one above the other.
12831 @item interleave, i
12832 Interleave fields. Reverse the effect of deinterleaving.
12834 Default value is @code{none}.
12836 @item luma_swap, ls
12837 @item chroma_swap, cs
12838 @item alpha_swap, as
12839 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
12842 @subsection Commands
12844 This filter supports the all above options as @ref{commands}.
12848 Apply inflate effect to the video.
12850 This filter replaces the pixel by the local(3x3) average by taking into account
12851 only values higher than the pixel.
12853 It accepts the following options:
12860 Limit the maximum change for each plane, default is 65535.
12861 If 0, plane will remain unchanged.
12864 @subsection Commands
12866 This filter supports the all above options as @ref{commands}.
12870 Simple interlacing filter from progressive contents. This interleaves upper (or
12871 lower) lines from odd frames with lower (or upper) lines from even frames,
12872 halving the frame rate and preserving image height.
12875 Original Original New Frame
12876 Frame 'j' Frame 'j+1' (tff)
12877 ========== =========== ==================
12878 Line 0 --------------------> Frame 'j' Line 0
12879 Line 1 Line 1 ----> Frame 'j+1' Line 1
12880 Line 2 ---------------------> Frame 'j' Line 2
12881 Line 3 Line 3 ----> Frame 'j+1' Line 3
12883 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
12886 It accepts the following optional parameters:
12890 This determines whether the interlaced frame is taken from the even
12891 (tff - default) or odd (bff) lines of the progressive frame.
12894 Vertical lowpass filter to avoid twitter interlacing and
12895 reduce moire patterns.
12899 Disable vertical lowpass filter
12902 Enable linear filter (default)
12905 Enable complex filter. This will slightly less reduce twitter and moire
12906 but better retain detail and subjective sharpness impression.
12913 Deinterlace input video by applying Donald Graft's adaptive kernel
12914 deinterling. Work on interlaced parts of a video to produce
12915 progressive frames.
12917 The description of the accepted parameters follows.
12921 Set the threshold which affects the filter's tolerance when
12922 determining if a pixel line must be processed. It must be an integer
12923 in the range [0,255] and defaults to 10. A value of 0 will result in
12924 applying the process on every pixels.
12927 Paint pixels exceeding the threshold value to white if set to 1.
12931 Set the fields order. Swap fields if set to 1, leave fields alone if
12935 Enable additional sharpening if set to 1. Default is 0.
12938 Enable twoway sharpening if set to 1. Default is 0.
12941 @subsection Examples
12945 Apply default values:
12947 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
12951 Enable additional sharpening:
12957 Paint processed pixels in white:
12965 Slowly update darker pixels.
12967 This filter makes short flashes of light appear longer.
12968 This filter accepts the following options:
12972 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
12975 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
12978 @section lenscorrection
12980 Correct radial lens distortion
12982 This filter can be used to correct for radial distortion as can result from the use
12983 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
12984 one can use tools available for example as part of opencv or simply trial-and-error.
12985 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
12986 and extract the k1 and k2 coefficients from the resulting matrix.
12988 Note that effectively the same filter is available in the open-source tools Krita and
12989 Digikam from the KDE project.
12991 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
12992 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
12993 brightness distribution, so you may want to use both filters together in certain
12994 cases, though you will have to take care of ordering, i.e. whether vignetting should
12995 be applied before or after lens correction.
12997 @subsection Options
12999 The filter accepts the following options:
13003 Relative x-coordinate of the focal point of the image, and thereby the center of the
13004 distortion. This value has a range [0,1] and is expressed as fractions of the image
13005 width. Default is 0.5.
13007 Relative y-coordinate of the focal point of the image, and thereby the center of the
13008 distortion. This value has a range [0,1] and is expressed as fractions of the image
13009 height. Default is 0.5.
13011 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13012 no correction. Default is 0.
13014 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13015 0 means no correction. Default is 0.
13018 The formula that generates the correction is:
13020 @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)
13022 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13023 distances from the focal point in the source and target images, respectively.
13027 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13029 The @code{lensfun} filter requires the camera make, camera model, and lens model
13030 to apply the lens correction. The filter will load the lensfun database and
13031 query it to find the corresponding camera and lens entries in the database. As
13032 long as these entries can be found with the given options, the filter can
13033 perform corrections on frames. Note that incomplete strings will result in the
13034 filter choosing the best match with the given options, and the filter will
13035 output the chosen camera and lens models (logged with level "info"). You must
13036 provide the make, camera model, and lens model as they are required.
13038 The filter accepts the following options:
13042 The make of the camera (for example, "Canon"). This option is required.
13045 The model of the camera (for example, "Canon EOS 100D"). This option is
13049 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13050 option is required.
13053 The type of correction to apply. The following values are valid options:
13057 Enables fixing lens vignetting.
13060 Enables fixing lens geometry. This is the default.
13063 Enables fixing chromatic aberrations.
13066 Enables fixing lens vignetting and lens geometry.
13069 Enables fixing lens vignetting and chromatic aberrations.
13072 Enables fixing both lens geometry and chromatic aberrations.
13075 Enables all possible corrections.
13079 The focal length of the image/video (zoom; expected constant for video). For
13080 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13081 range should be chosen when using that lens. Default 18.
13084 The aperture of the image/video (expected constant for video). Note that
13085 aperture is only used for vignetting correction. Default 3.5.
13087 @item focus_distance
13088 The focus distance of the image/video (expected constant for video). Note that
13089 focus distance is only used for vignetting and only slightly affects the
13090 vignetting correction process. If unknown, leave it at the default value (which
13094 The scale factor which is applied after transformation. After correction the
13095 video is no longer necessarily rectangular. This parameter controls how much of
13096 the resulting image is visible. The value 0 means that a value will be chosen
13097 automatically such that there is little or no unmapped area in the output
13098 image. 1.0 means that no additional scaling is done. Lower values may result
13099 in more of the corrected image being visible, while higher values may avoid
13100 unmapped areas in the output.
13102 @item target_geometry
13103 The target geometry of the output image/video. The following values are valid
13107 @item rectilinear (default)
13110 @item equirectangular
13111 @item fisheye_orthographic
13112 @item fisheye_stereographic
13113 @item fisheye_equisolid
13114 @item fisheye_thoby
13117 Apply the reverse of image correction (instead of correcting distortion, apply
13120 @item interpolation
13121 The type of interpolation used when correcting distortion. The following values
13126 @item linear (default)
13131 @subsection Examples
13135 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13136 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13140 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
13144 Apply the same as before, but only for the first 5 seconds of video.
13147 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
13154 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13155 score between two input videos.
13157 The obtained VMAF score is printed through the logging system.
13159 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13160 After installing the library it can be enabled using:
13161 @code{./configure --enable-libvmaf}.
13162 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13164 The filter has following options:
13168 Set the model path which is to be used for SVM.
13169 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13172 Set the file path to be used to store logs.
13175 Set the format of the log file (csv, json or xml).
13177 @item enable_transform
13178 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13179 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13180 Default value: @code{false}
13183 Invokes the phone model which will generate VMAF scores higher than in the
13184 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13185 Default value: @code{false}
13188 Enables computing psnr along with vmaf.
13189 Default value: @code{false}
13192 Enables computing ssim along with vmaf.
13193 Default value: @code{false}
13196 Enables computing ms_ssim along with vmaf.
13197 Default value: @code{false}
13200 Set the pool method to be used for computing vmaf.
13201 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13204 Set number of threads to be used when computing vmaf.
13205 Default value: @code{0}, which makes use of all available logical processors.
13208 Set interval for frame subsampling used when computing vmaf.
13209 Default value: @code{1}
13211 @item enable_conf_interval
13212 Enables confidence interval.
13213 Default value: @code{false}
13216 This filter also supports the @ref{framesync} options.
13218 @subsection Examples
13221 On the below examples the input file @file{main.mpg} being processed is
13222 compared with the reference file @file{ref.mpg}.
13225 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13229 Example with options:
13231 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13235 Example with options and different containers:
13237 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 -
13243 Limits the pixel components values to the specified range [min, max].
13245 The filter accepts the following options:
13249 Lower bound. Defaults to the lowest allowed value for the input.
13252 Upper bound. Defaults to the highest allowed value for the input.
13255 Specify which planes will be processed. Defaults to all available.
13262 The filter accepts the following options:
13266 Set the number of loops. Setting this value to -1 will result in infinite loops.
13270 Set maximal size in number of frames. Default is 0.
13273 Set first frame of loop. Default is 0.
13276 @subsection Examples
13280 Loop single first frame infinitely:
13282 loop=loop=-1:size=1:start=0
13286 Loop single first frame 10 times:
13288 loop=loop=10:size=1:start=0
13292 Loop 10 first frames 5 times:
13294 loop=loop=5:size=10:start=0
13300 Apply a 1D LUT to an input video.
13302 The filter accepts the following options:
13306 Set the 1D LUT file name.
13308 Currently supported formats:
13317 Select interpolation mode.
13319 Available values are:
13323 Use values from the nearest defined point.
13325 Interpolate values using the linear interpolation.
13327 Interpolate values using the cosine interpolation.
13329 Interpolate values using the cubic interpolation.
13331 Interpolate values using the spline interpolation.
13338 Apply a 3D LUT to an input video.
13340 The filter accepts the following options:
13344 Set the 3D LUT file name.
13346 Currently supported formats:
13360 Select interpolation mode.
13362 Available values are:
13366 Use values from the nearest defined point.
13368 Interpolate values using the 8 points defining a cube.
13370 Interpolate values using a tetrahedron.
13376 Turn certain luma values into transparency.
13378 The filter accepts the following options:
13382 Set the luma which will be used as base for transparency.
13383 Default value is @code{0}.
13386 Set the range of luma values to be keyed out.
13387 Default value is @code{0.01}.
13390 Set the range of softness. Default value is @code{0}.
13391 Use this to control gradual transition from zero to full transparency.
13394 @subsection Commands
13395 This filter supports same @ref{commands} as options.
13396 The command accepts the same syntax of the corresponding option.
13398 If the specified expression is not valid, it is kept at its current
13401 @section lut, lutrgb, lutyuv
13403 Compute a look-up table for binding each pixel component input value
13404 to an output value, and apply it to the input video.
13406 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13407 to an RGB input video.
13409 These filters accept the following parameters:
13412 set first pixel component expression
13414 set second pixel component expression
13416 set third pixel component expression
13418 set fourth pixel component expression, corresponds to the alpha component
13421 set red component expression
13423 set green component expression
13425 set blue component expression
13427 alpha component expression
13430 set Y/luminance component expression
13432 set U/Cb component expression
13434 set V/Cr component expression
13437 Each of them specifies the expression to use for computing the lookup table for
13438 the corresponding pixel component values.
13440 The exact component associated to each of the @var{c*} options depends on the
13443 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13444 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13446 The expressions can contain the following constants and functions:
13451 The input width and height.
13454 The input value for the pixel component.
13457 The input value, clipped to the @var{minval}-@var{maxval} range.
13460 The maximum value for the pixel component.
13463 The minimum value for the pixel component.
13466 The negated value for the pixel component value, clipped to the
13467 @var{minval}-@var{maxval} range; it corresponds to the expression
13468 "maxval-clipval+minval".
13471 The computed value in @var{val}, clipped to the
13472 @var{minval}-@var{maxval} range.
13474 @item gammaval(gamma)
13475 The computed gamma correction value of the pixel component value,
13476 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13478 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13482 All expressions default to "val".
13484 @subsection Examples
13488 Negate input video:
13490 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13491 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13494 The above is the same as:
13496 lutrgb="r=negval:g=negval:b=negval"
13497 lutyuv="y=negval:u=negval:v=negval"
13507 Remove chroma components, turning the video into a graytone image:
13509 lutyuv="u=128:v=128"
13513 Apply a luma burning effect:
13519 Remove green and blue components:
13525 Set a constant alpha channel value on input:
13527 format=rgba,lutrgb=a="maxval-minval/2"
13531 Correct luminance gamma by a factor of 0.5:
13533 lutyuv=y=gammaval(0.5)
13537 Discard least significant bits of luma:
13539 lutyuv=y='bitand(val, 128+64+32)'
13543 Technicolor like effect:
13545 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13549 @section lut2, tlut2
13551 The @code{lut2} filter takes two input streams and outputs one
13554 The @code{tlut2} (time lut2) filter takes two consecutive frames
13555 from one single stream.
13557 This filter accepts the following parameters:
13560 set first pixel component expression
13562 set second pixel component expression
13564 set third pixel component expression
13566 set fourth pixel component expression, corresponds to the alpha component
13569 set output bit depth, only available for @code{lut2} filter. By default is 0,
13570 which means bit depth is automatically picked from first input format.
13573 The @code{lut2} filter also supports the @ref{framesync} options.
13575 Each of them specifies the expression to use for computing the lookup table for
13576 the corresponding pixel component values.
13578 The exact component associated to each of the @var{c*} options depends on the
13581 The expressions can contain the following constants:
13586 The input width and height.
13589 The first input value for the pixel component.
13592 The second input value for the pixel component.
13595 The first input video bit depth.
13598 The second input video bit depth.
13601 All expressions default to "x".
13603 @subsection Examples
13607 Highlight differences between two RGB video streams:
13609 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)'
13613 Highlight differences between two YUV video streams:
13615 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)'
13619 Show max difference between two video streams:
13621 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)))'
13625 @section maskedclamp
13627 Clamp the first input stream with the second input and third input stream.
13629 Returns the value of first stream to be between second input
13630 stream - @code{undershoot} and third input stream + @code{overshoot}.
13632 This filter accepts the following options:
13635 Default value is @code{0}.
13638 Default value is @code{0}.
13641 Set which planes will be processed as bitmap, unprocessed planes will be
13642 copied from first stream.
13643 By default value 0xf, all planes will be processed.
13648 Merge the second and third input stream into output stream using absolute differences
13649 between second input stream and first input stream and absolute difference between
13650 third input stream and first input stream. The picked value will be from second input
13651 stream if second absolute difference is greater than first one or from third input stream
13654 This filter accepts the following options:
13657 Set which planes will be processed as bitmap, unprocessed planes will be
13658 copied from first stream.
13659 By default value 0xf, all planes will be processed.
13662 @section maskedmerge
13664 Merge the first input stream with the second input stream using per pixel
13665 weights in the third input stream.
13667 A value of 0 in the third stream pixel component means that pixel component
13668 from first stream is returned unchanged, while maximum value (eg. 255 for
13669 8-bit videos) means that pixel component from second stream is returned
13670 unchanged. Intermediate values define the amount of merging between both
13671 input stream's pixel components.
13673 This filter accepts the following options:
13676 Set which planes will be processed as bitmap, unprocessed planes will be
13677 copied from first stream.
13678 By default value 0xf, all planes will be processed.
13683 Merge the second and third input stream into output stream using absolute differences
13684 between second input stream and first input stream and absolute difference between
13685 third input stream and first input stream. The picked value will be from second input
13686 stream if second absolute difference is less than first one or from third input stream
13689 This filter accepts the following options:
13692 Set which planes will be processed as bitmap, unprocessed planes will be
13693 copied from first stream.
13694 By default value 0xf, all planes will be processed.
13697 @section maskedthreshold
13698 Pick pixels comparing absolute difference of two video streams with fixed
13701 If absolute difference between pixel component of first and second video
13702 stream is equal or lower than user supplied threshold than pixel component
13703 from first video stream is picked, otherwise pixel component from second
13704 video stream is picked.
13706 This filter accepts the following options:
13709 Set threshold used when picking pixels from absolute difference from two input
13713 Set which planes will be processed as bitmap, unprocessed planes will be
13714 copied from second stream.
13715 By default value 0xf, all planes will be processed.
13719 Create mask from input video.
13721 For example it is useful to create motion masks after @code{tblend} filter.
13723 This filter accepts the following options:
13727 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
13730 Set high threshold. Any pixel component higher than this value will be set to max value
13731 allowed for current pixel format.
13734 Set planes to filter, by default all available planes are filtered.
13737 Fill all frame pixels with this value.
13740 Set max average pixel value for frame. If sum of all pixel components is higher that this
13741 average, output frame will be completely filled with value set by @var{fill} option.
13742 Typically useful for scene changes when used in combination with @code{tblend} filter.
13747 Apply motion-compensation deinterlacing.
13749 It needs one field per frame as input and must thus be used together
13750 with yadif=1/3 or equivalent.
13752 This filter accepts the following options:
13755 Set the deinterlacing mode.
13757 It accepts one of the following values:
13762 use iterative motion estimation
13764 like @samp{slow}, but use multiple reference frames.
13766 Default value is @samp{fast}.
13769 Set the picture field parity assumed for the input video. It must be
13770 one of the following values:
13774 assume top field first
13776 assume bottom field first
13779 Default value is @samp{bff}.
13782 Set per-block quantization parameter (QP) used by the internal
13785 Higher values should result in a smoother motion vector field but less
13786 optimal individual vectors. Default value is 1.
13791 Pick median pixel from certain rectangle defined by radius.
13793 This filter accepts the following options:
13797 Set horizontal radius size. Default value is @code{1}.
13798 Allowed range is integer from 1 to 127.
13801 Set which planes to process. Default is @code{15}, which is all available planes.
13804 Set vertical radius size. Default value is @code{0}.
13805 Allowed range is integer from 0 to 127.
13806 If it is 0, value will be picked from horizontal @code{radius} option.
13809 Set median percentile. Default value is @code{0.5}.
13810 Default value of @code{0.5} will pick always median values, while @code{0} will pick
13811 minimum values, and @code{1} maximum values.
13814 @subsection Commands
13815 This filter supports same @ref{commands} as options.
13816 The command accepts the same syntax of the corresponding option.
13818 If the specified expression is not valid, it is kept at its current
13821 @section mergeplanes
13823 Merge color channel components from several video streams.
13825 The filter accepts up to 4 input streams, and merge selected input
13826 planes to the output video.
13828 This filter accepts the following options:
13831 Set input to output plane mapping. Default is @code{0}.
13833 The mappings is specified as a bitmap. It should be specified as a
13834 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
13835 mapping for the first plane of the output stream. 'A' sets the number of
13836 the input stream to use (from 0 to 3), and 'a' the plane number of the
13837 corresponding input to use (from 0 to 3). The rest of the mappings is
13838 similar, 'Bb' describes the mapping for the output stream second
13839 plane, 'Cc' describes the mapping for the output stream third plane and
13840 'Dd' describes the mapping for the output stream fourth plane.
13843 Set output pixel format. Default is @code{yuva444p}.
13846 @subsection Examples
13850 Merge three gray video streams of same width and height into single video stream:
13852 [a0][a1][a2]mergeplanes=0x001020:yuv444p
13856 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
13858 [a0][a1]mergeplanes=0x00010210:yuva444p
13862 Swap Y and A plane in yuva444p stream:
13864 format=yuva444p,mergeplanes=0x03010200:yuva444p
13868 Swap U and V plane in yuv420p stream:
13870 format=yuv420p,mergeplanes=0x000201:yuv420p
13874 Cast a rgb24 clip to yuv444p:
13876 format=rgb24,mergeplanes=0x000102:yuv444p
13882 Estimate and export motion vectors using block matching algorithms.
13883 Motion vectors are stored in frame side data to be used by other filters.
13885 This filter accepts the following options:
13888 Specify the motion estimation method. Accepts one of the following values:
13892 Exhaustive search algorithm.
13894 Three step search algorithm.
13896 Two dimensional logarithmic search algorithm.
13898 New three step search algorithm.
13900 Four step search algorithm.
13902 Diamond search algorithm.
13904 Hexagon-based search algorithm.
13906 Enhanced predictive zonal search algorithm.
13908 Uneven multi-hexagon search algorithm.
13910 Default value is @samp{esa}.
13913 Macroblock size. Default @code{16}.
13916 Search parameter. Default @code{7}.
13919 @section midequalizer
13921 Apply Midway Image Equalization effect using two video streams.
13923 Midway Image Equalization adjusts a pair of images to have the same
13924 histogram, while maintaining their dynamics as much as possible. It's
13925 useful for e.g. matching exposures from a pair of stereo cameras.
13927 This filter has two inputs and one output, which must be of same pixel format, but
13928 may be of different sizes. The output of filter is first input adjusted with
13929 midway histogram of both inputs.
13931 This filter accepts the following option:
13935 Set which planes to process. Default is @code{15}, which is all available planes.
13938 @section minterpolate
13940 Convert the video to specified frame rate using motion interpolation.
13942 This filter accepts the following options:
13945 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}.
13948 Motion interpolation mode. Following values are accepted:
13951 Duplicate previous or next frame for interpolating new ones.
13953 Blend source frames. Interpolated frame is mean of previous and next frames.
13955 Motion compensated interpolation. Following options are effective when this mode is selected:
13959 Motion compensation mode. Following values are accepted:
13962 Overlapped block motion compensation.
13964 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
13966 Default mode is @samp{obmc}.
13969 Motion estimation mode. Following values are accepted:
13972 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
13974 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
13976 Default mode is @samp{bilat}.
13979 The algorithm to be used for motion estimation. Following values are accepted:
13982 Exhaustive search algorithm.
13984 Three step search algorithm.
13986 Two dimensional logarithmic search algorithm.
13988 New three step search algorithm.
13990 Four step search algorithm.
13992 Diamond search algorithm.
13994 Hexagon-based search algorithm.
13996 Enhanced predictive zonal search algorithm.
13998 Uneven multi-hexagon search algorithm.
14000 Default algorithm is @samp{epzs}.
14003 Macroblock size. Default @code{16}.
14006 Motion estimation search parameter. Default @code{32}.
14009 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).
14014 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:
14017 Disable scene change detection.
14019 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14021 Default method is @samp{fdiff}.
14023 @item scd_threshold
14024 Scene change detection threshold. Default is @code{10.}.
14029 Mix several video input streams into one video stream.
14031 A description of the accepted options follows.
14035 The number of inputs. If unspecified, it defaults to 2.
14038 Specify weight of each input video stream as sequence.
14039 Each weight is separated by space. If number of weights
14040 is smaller than number of @var{frames} last specified
14041 weight will be used for all remaining unset weights.
14044 Specify scale, if it is set it will be multiplied with sum
14045 of each weight multiplied with pixel values to give final destination
14046 pixel value. By default @var{scale} is auto scaled to sum of weights.
14049 Specify how end of stream is determined.
14052 The duration of the longest input. (default)
14055 The duration of the shortest input.
14058 The duration of the first input.
14062 @section mpdecimate
14064 Drop frames that do not differ greatly from the previous frame in
14065 order to reduce frame rate.
14067 The main use of this filter is for very-low-bitrate encoding
14068 (e.g. streaming over dialup modem), but it could in theory be used for
14069 fixing movies that were inverse-telecined incorrectly.
14071 A description of the accepted options follows.
14075 Set the maximum number of consecutive frames which can be dropped (if
14076 positive), or the minimum interval between dropped frames (if
14077 negative). If the value is 0, the frame is dropped disregarding the
14078 number of previous sequentially dropped frames.
14080 Default value is 0.
14085 Set the dropping threshold values.
14087 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14088 represent actual pixel value differences, so a threshold of 64
14089 corresponds to 1 unit of difference for each pixel, or the same spread
14090 out differently over the block.
14092 A frame is a candidate for dropping if no 8x8 blocks differ by more
14093 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14094 meaning the whole image) differ by more than a threshold of @option{lo}.
14096 Default value for @option{hi} is 64*12, default value for @option{lo} is
14097 64*5, and default value for @option{frac} is 0.33.
14103 Negate (invert) the input video.
14105 It accepts the following option:
14110 With value 1, it negates the alpha component, if present. Default value is 0.
14116 Denoise frames using Non-Local Means algorithm.
14118 Each pixel is adjusted by looking for other pixels with similar contexts. This
14119 context similarity is defined by comparing their surrounding patches of size
14120 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14123 Note that the research area defines centers for patches, which means some
14124 patches will be made of pixels outside that research area.
14126 The filter accepts the following options.
14130 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14133 Set patch size. Default is 7. Must be odd number in range [0, 99].
14136 Same as @option{p} but for chroma planes.
14138 The default value is @var{0} and means automatic.
14141 Set research size. Default is 15. Must be odd number in range [0, 99].
14144 Same as @option{r} but for chroma planes.
14146 The default value is @var{0} and means automatic.
14151 Deinterlace video using neural network edge directed interpolation.
14153 This filter accepts the following options:
14157 Mandatory option, without binary file filter can not work.
14158 Currently file can be found here:
14159 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14162 Set which frames to deinterlace, by default it is @code{all}.
14163 Can be @code{all} or @code{interlaced}.
14166 Set mode of operation.
14168 Can be one of the following:
14172 Use frame flags, both fields.
14174 Use frame flags, single field.
14176 Use top field only.
14178 Use bottom field only.
14180 Use both fields, top first.
14182 Use both fields, bottom first.
14186 Set which planes to process, by default filter process all frames.
14189 Set size of local neighborhood around each pixel, used by the predictor neural
14192 Can be one of the following:
14205 Set the number of neurons in predictor neural network.
14206 Can be one of the following:
14217 Controls the number of different neural network predictions that are blended
14218 together to compute the final output value. Can be @code{fast}, default or
14222 Set which set of weights to use in the predictor.
14223 Can be one of the following:
14227 weights trained to minimize absolute error
14229 weights trained to minimize squared error
14233 Controls whether or not the prescreener neural network is used to decide
14234 which pixels should be processed by the predictor neural network and which
14235 can be handled by simple cubic interpolation.
14236 The prescreener is trained to know whether cubic interpolation will be
14237 sufficient for a pixel or whether it should be predicted by the predictor nn.
14238 The computational complexity of the prescreener nn is much less than that of
14239 the predictor nn. Since most pixels can be handled by cubic interpolation,
14240 using the prescreener generally results in much faster processing.
14241 The prescreener is pretty accurate, so the difference between using it and not
14242 using it is almost always unnoticeable.
14244 Can be one of the following:
14252 Default is @code{new}.
14255 Set various debugging flags.
14260 Force libavfilter not to use any of the specified pixel formats for the
14261 input to the next filter.
14263 It accepts the following parameters:
14267 A '|'-separated list of pixel format names, such as
14268 pix_fmts=yuv420p|monow|rgb24".
14272 @subsection Examples
14276 Force libavfilter to use a format different from @var{yuv420p} for the
14277 input to the vflip filter:
14279 noformat=pix_fmts=yuv420p,vflip
14283 Convert the input video to any of the formats not contained in the list:
14285 noformat=yuv420p|yuv444p|yuv410p
14291 Add noise on video input frame.
14293 The filter accepts the following options:
14301 Set noise seed for specific pixel component or all pixel components in case
14302 of @var{all_seed}. Default value is @code{123457}.
14304 @item all_strength, alls
14305 @item c0_strength, c0s
14306 @item c1_strength, c1s
14307 @item c2_strength, c2s
14308 @item c3_strength, c3s
14309 Set noise strength for specific pixel component or all pixel components in case
14310 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14312 @item all_flags, allf
14313 @item c0_flags, c0f
14314 @item c1_flags, c1f
14315 @item c2_flags, c2f
14316 @item c3_flags, c3f
14317 Set pixel component flags or set flags for all components if @var{all_flags}.
14318 Available values for component flags are:
14321 averaged temporal noise (smoother)
14323 mix random noise with a (semi)regular pattern
14325 temporal noise (noise pattern changes between frames)
14327 uniform noise (gaussian otherwise)
14331 @subsection Examples
14333 Add temporal and uniform noise to input video:
14335 noise=alls=20:allf=t+u
14340 Normalize RGB video (aka histogram stretching, contrast stretching).
14341 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14343 For each channel of each frame, the filter computes the input range and maps
14344 it linearly to the user-specified output range. The output range defaults
14345 to the full dynamic range from pure black to pure white.
14347 Temporal smoothing can be used on the input range to reduce flickering (rapid
14348 changes in brightness) caused when small dark or bright objects enter or leave
14349 the scene. This is similar to the auto-exposure (automatic gain control) on a
14350 video camera, and, like a video camera, it may cause a period of over- or
14351 under-exposure of the video.
14353 The R,G,B channels can be normalized independently, which may cause some
14354 color shifting, or linked together as a single channel, which prevents
14355 color shifting. Linked normalization preserves hue. Independent normalization
14356 does not, so it can be used to remove some color casts. Independent and linked
14357 normalization can be combined in any ratio.
14359 The normalize filter accepts the following options:
14364 Colors which define the output range. The minimum input value is mapped to
14365 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14366 The defaults are black and white respectively. Specifying white for
14367 @var{blackpt} and black for @var{whitept} will give color-inverted,
14368 normalized video. Shades of grey can be used to reduce the dynamic range
14369 (contrast). Specifying saturated colors here can create some interesting
14373 The number of previous frames to use for temporal smoothing. The input range
14374 of each channel is smoothed using a rolling average over the current frame
14375 and the @var{smoothing} previous frames. The default is 0 (no temporal
14379 Controls the ratio of independent (color shifting) channel normalization to
14380 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14381 independent. Defaults to 1.0 (fully independent).
14384 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14385 expensive no-op. Defaults to 1.0 (full strength).
14389 @subsection Commands
14390 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14391 The command accepts the same syntax of the corresponding option.
14393 If the specified expression is not valid, it is kept at its current
14396 @subsection Examples
14398 Stretch video contrast to use the full dynamic range, with no temporal
14399 smoothing; may flicker depending on the source content:
14401 normalize=blackpt=black:whitept=white:smoothing=0
14404 As above, but with 50 frames of temporal smoothing; flicker should be
14405 reduced, depending on the source content:
14407 normalize=blackpt=black:whitept=white:smoothing=50
14410 As above, but with hue-preserving linked channel normalization:
14412 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14415 As above, but with half strength:
14417 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14420 Map the darkest input color to red, the brightest input color to cyan:
14422 normalize=blackpt=red:whitept=cyan
14427 Pass the video source unchanged to the output.
14430 Optical Character Recognition
14432 This filter uses Tesseract for optical character recognition. To enable
14433 compilation of this filter, you need to configure FFmpeg with
14434 @code{--enable-libtesseract}.
14436 It accepts the following options:
14440 Set datapath to tesseract data. Default is to use whatever was
14441 set at installation.
14444 Set language, default is "eng".
14447 Set character whitelist.
14450 Set character blacklist.
14453 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14454 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14458 Apply a video transform using libopencv.
14460 To enable this filter, install the libopencv library and headers and
14461 configure FFmpeg with @code{--enable-libopencv}.
14463 It accepts the following parameters:
14468 The name of the libopencv filter to apply.
14470 @item filter_params
14471 The parameters to pass to the libopencv filter. If not specified, the default
14472 values are assumed.
14476 Refer to the official libopencv documentation for more precise
14478 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14480 Several libopencv filters are supported; see the following subsections.
14485 Dilate an image by using a specific structuring element.
14486 It corresponds to the libopencv function @code{cvDilate}.
14488 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14490 @var{struct_el} represents a structuring element, and has the syntax:
14491 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14493 @var{cols} and @var{rows} represent the number of columns and rows of
14494 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14495 point, and @var{shape} the shape for the structuring element. @var{shape}
14496 must be "rect", "cross", "ellipse", or "custom".
14498 If the value for @var{shape} is "custom", it must be followed by a
14499 string of the form "=@var{filename}". The file with name
14500 @var{filename} is assumed to represent a binary image, with each
14501 printable character corresponding to a bright pixel. When a custom
14502 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14503 or columns and rows of the read file are assumed instead.
14505 The default value for @var{struct_el} is "3x3+0x0/rect".
14507 @var{nb_iterations} specifies the number of times the transform is
14508 applied to the image, and defaults to 1.
14512 # Use the default values
14515 # Dilate using a structuring element with a 5x5 cross, iterating two times
14516 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14518 # Read the shape from the file diamond.shape, iterating two times.
14519 # The file diamond.shape may contain a pattern of characters like this
14525 # The specified columns and rows are ignored
14526 # but the anchor point coordinates are not
14527 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14532 Erode an image by using a specific structuring element.
14533 It corresponds to the libopencv function @code{cvErode}.
14535 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14536 with the same syntax and semantics as the @ref{dilate} filter.
14540 Smooth the input video.
14542 The filter takes the following parameters:
14543 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14545 @var{type} is the type of smooth filter to apply, and must be one of
14546 the following values: "blur", "blur_no_scale", "median", "gaussian",
14547 or "bilateral". The default value is "gaussian".
14549 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14550 depends on the smooth type. @var{param1} and
14551 @var{param2} accept integer positive values or 0. @var{param3} and
14552 @var{param4} accept floating point values.
14554 The default value for @var{param1} is 3. The default value for the
14555 other parameters is 0.
14557 These parameters correspond to the parameters assigned to the
14558 libopencv function @code{cvSmooth}.
14560 @section oscilloscope
14562 2D Video Oscilloscope.
14564 Useful to measure spatial impulse, step responses, chroma delays, etc.
14566 It accepts the following parameters:
14570 Set scope center x position.
14573 Set scope center y position.
14576 Set scope size, relative to frame diagonal.
14579 Set scope tilt/rotation.
14585 Set trace center x position.
14588 Set trace center y position.
14591 Set trace width, relative to width of frame.
14594 Set trace height, relative to height of frame.
14597 Set which components to trace. By default it traces first three components.
14600 Draw trace grid. By default is enabled.
14603 Draw some statistics. By default is enabled.
14606 Draw scope. By default is enabled.
14609 @subsection Commands
14610 This filter supports same @ref{commands} as options.
14611 The command accepts the same syntax of the corresponding option.
14613 If the specified expression is not valid, it is kept at its current
14616 @subsection Examples
14620 Inspect full first row of video frame.
14622 oscilloscope=x=0.5:y=0:s=1
14626 Inspect full last row of video frame.
14628 oscilloscope=x=0.5:y=1:s=1
14632 Inspect full 5th line of video frame of height 1080.
14634 oscilloscope=x=0.5:y=5/1080:s=1
14638 Inspect full last column of video frame.
14640 oscilloscope=x=1:y=0.5:s=1:t=1
14648 Overlay one video on top of another.
14650 It takes two inputs and has one output. The first input is the "main"
14651 video on which the second input is overlaid.
14653 It accepts the following parameters:
14655 A description of the accepted options follows.
14660 Set the expression for the x and y coordinates of the overlaid video
14661 on the main video. Default value is "0" for both expressions. In case
14662 the expression is invalid, it is set to a huge value (meaning that the
14663 overlay will not be displayed within the output visible area).
14666 See @ref{framesync}.
14669 Set when the expressions for @option{x}, and @option{y} are evaluated.
14671 It accepts the following values:
14674 only evaluate expressions once during the filter initialization or
14675 when a command is processed
14678 evaluate expressions for each incoming frame
14681 Default value is @samp{frame}.
14684 See @ref{framesync}.
14687 Set the format for the output video.
14689 It accepts the following values:
14692 force YUV420 output
14695 force YUV420p10 output
14698 force YUV422 output
14701 force YUV422p10 output
14704 force YUV444 output
14707 force packed RGB output
14710 force planar RGB output
14713 automatically pick format
14716 Default value is @samp{yuv420}.
14719 See @ref{framesync}.
14722 Set format of alpha of the overlaid video, it can be @var{straight} or
14723 @var{premultiplied}. Default is @var{straight}.
14726 The @option{x}, and @option{y} expressions can contain the following
14732 The main input width and height.
14736 The overlay input width and height.
14740 The computed values for @var{x} and @var{y}. They are evaluated for
14745 horizontal and vertical chroma subsample values of the output
14746 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
14750 the number of input frame, starting from 0
14753 the position in the file of the input frame, NAN if unknown
14756 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
14760 This filter also supports the @ref{framesync} options.
14762 Note that the @var{n}, @var{pos}, @var{t} variables are available only
14763 when evaluation is done @emph{per frame}, and will evaluate to NAN
14764 when @option{eval} is set to @samp{init}.
14766 Be aware that frames are taken from each input video in timestamp
14767 order, hence, if their initial timestamps differ, it is a good idea
14768 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
14769 have them begin in the same zero timestamp, as the example for
14770 the @var{movie} filter does.
14772 You can chain together more overlays but you should test the
14773 efficiency of such approach.
14775 @subsection Commands
14777 This filter supports the following commands:
14781 Modify the x and y of the overlay input.
14782 The command accepts the same syntax of the corresponding option.
14784 If the specified expression is not valid, it is kept at its current
14788 @subsection Examples
14792 Draw the overlay at 10 pixels from the bottom right corner of the main
14795 overlay=main_w-overlay_w-10:main_h-overlay_h-10
14798 Using named options the example above becomes:
14800 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
14804 Insert a transparent PNG logo in the bottom left corner of the input,
14805 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
14807 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
14811 Insert 2 different transparent PNG logos (second logo on bottom
14812 right corner) using the @command{ffmpeg} tool:
14814 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
14818 Add a transparent color layer on top of the main video; @code{WxH}
14819 must specify the size of the main input to the overlay filter:
14821 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
14825 Play an original video and a filtered version (here with the deshake
14826 filter) side by side using the @command{ffplay} tool:
14828 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
14831 The above command is the same as:
14833 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
14837 Make a sliding overlay appearing from the left to the right top part of the
14838 screen starting since time 2:
14840 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
14844 Compose output by putting two input videos side to side:
14846 ffmpeg -i left.avi -i right.avi -filter_complex "
14847 nullsrc=size=200x100 [background];
14848 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
14849 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
14850 [background][left] overlay=shortest=1 [background+left];
14851 [background+left][right] overlay=shortest=1:x=100 [left+right]
14856 Mask 10-20 seconds of a video by applying the delogo filter to a section
14858 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
14859 -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]'
14864 Chain several overlays in cascade:
14866 nullsrc=s=200x200 [bg];
14867 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
14868 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
14869 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
14870 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
14871 [in3] null, [mid2] overlay=100:100 [out0]
14876 @anchor{overlay_cuda}
14877 @section overlay_cuda
14879 Overlay one video on top of another.
14881 This is the CUDA variant of the @ref{overlay} filter.
14882 It only accepts CUDA frames. The underlying input pixel formats have to match.
14884 It takes two inputs and has one output. The first input is the "main"
14885 video on which the second input is overlaid.
14887 It accepts the following parameters:
14892 Set the x and y coordinates of the overlaid video on the main video.
14893 Default value is "0" for both expressions.
14896 See @ref{framesync}.
14899 See @ref{framesync}.
14902 See @ref{framesync}.
14906 This filter also supports the @ref{framesync} options.
14910 Apply Overcomplete Wavelet denoiser.
14912 The filter accepts the following options:
14918 Larger depth values will denoise lower frequency components more, but
14919 slow down filtering.
14921 Must be an int in the range 8-16, default is @code{8}.
14923 @item luma_strength, ls
14926 Must be a double value in the range 0-1000, default is @code{1.0}.
14928 @item chroma_strength, cs
14929 Set chroma strength.
14931 Must be a double value in the range 0-1000, default is @code{1.0}.
14937 Add paddings to the input image, and place the original input at the
14938 provided @var{x}, @var{y} coordinates.
14940 It accepts the following parameters:
14945 Specify an expression for the size of the output image with the
14946 paddings added. If the value for @var{width} or @var{height} is 0, the
14947 corresponding input size is used for the output.
14949 The @var{width} expression can reference the value set by the
14950 @var{height} expression, and vice versa.
14952 The default value of @var{width} and @var{height} is 0.
14956 Specify the offsets to place the input image at within the padded area,
14957 with respect to the top/left border of the output image.
14959 The @var{x} expression can reference the value set by the @var{y}
14960 expression, and vice versa.
14962 The default value of @var{x} and @var{y} is 0.
14964 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
14965 so the input image is centered on the padded area.
14968 Specify the color of the padded area. For the syntax of this option,
14969 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
14970 manual,ffmpeg-utils}.
14972 The default value of @var{color} is "black".
14975 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
14977 It accepts the following values:
14981 Only evaluate expressions once during the filter initialization or when
14982 a command is processed.
14985 Evaluate expressions for each incoming frame.
14989 Default value is @samp{init}.
14992 Pad to aspect instead to a resolution.
14996 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
14997 options are expressions containing the following constants:
15002 The input video width and height.
15006 These are the same as @var{in_w} and @var{in_h}.
15010 The output width and height (the size of the padded area), as
15011 specified by the @var{width} and @var{height} expressions.
15015 These are the same as @var{out_w} and @var{out_h}.
15019 The x and y offsets as specified by the @var{x} and @var{y}
15020 expressions, or NAN if not yet specified.
15023 same as @var{iw} / @var{ih}
15026 input sample aspect ratio
15029 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15033 The horizontal and vertical chroma subsample values. For example for the
15034 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15037 @subsection Examples
15041 Add paddings with the color "violet" to the input video. The output video
15042 size is 640x480, and the top-left corner of the input video is placed at
15045 pad=640:480:0:40:violet
15048 The example above is equivalent to the following command:
15050 pad=width=640:height=480:x=0:y=40:color=violet
15054 Pad the input to get an output with dimensions increased by 3/2,
15055 and put the input video at the center of the padded area:
15057 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15061 Pad the input to get a squared output with size equal to the maximum
15062 value between the input width and height, and put the input video at
15063 the center of the padded area:
15065 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15069 Pad the input to get a final w/h ratio of 16:9:
15071 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15075 In case of anamorphic video, in order to set the output display aspect
15076 correctly, it is necessary to use @var{sar} in the expression,
15077 according to the relation:
15079 (ih * X / ih) * sar = output_dar
15080 X = output_dar / sar
15083 Thus the previous example needs to be modified to:
15085 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15089 Double the output size and put the input video in the bottom-right
15090 corner of the output padded area:
15092 pad="2*iw:2*ih:ow-iw:oh-ih"
15096 @anchor{palettegen}
15097 @section palettegen
15099 Generate one palette for a whole video stream.
15101 It accepts the following options:
15105 Set the maximum number of colors to quantize in the palette.
15106 Note: the palette will still contain 256 colors; the unused palette entries
15109 @item reserve_transparent
15110 Create a palette of 255 colors maximum and reserve the last one for
15111 transparency. Reserving the transparency color is useful for GIF optimization.
15112 If not set, the maximum of colors in the palette will be 256. You probably want
15113 to disable this option for a standalone image.
15116 @item transparency_color
15117 Set the color that will be used as background for transparency.
15120 Set statistics mode.
15122 It accepts the following values:
15125 Compute full frame histograms.
15127 Compute histograms only for the part that differs from previous frame. This
15128 might be relevant to give more importance to the moving part of your input if
15129 the background is static.
15131 Compute new histogram for each frame.
15134 Default value is @var{full}.
15137 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15138 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15139 color quantization of the palette. This information is also visible at
15140 @var{info} logging level.
15142 @subsection Examples
15146 Generate a representative palette of a given video using @command{ffmpeg}:
15148 ffmpeg -i input.mkv -vf palettegen palette.png
15152 @section paletteuse
15154 Use a palette to downsample an input video stream.
15156 The filter takes two inputs: one video stream and a palette. The palette must
15157 be a 256 pixels image.
15159 It accepts the following options:
15163 Select dithering mode. Available algorithms are:
15166 Ordered 8x8 bayer dithering (deterministic)
15168 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15169 Note: this dithering is sometimes considered "wrong" and is included as a
15171 @item floyd_steinberg
15172 Floyd and Steingberg dithering (error diffusion)
15174 Frankie Sierra dithering v2 (error diffusion)
15176 Frankie Sierra dithering v2 "Lite" (error diffusion)
15179 Default is @var{sierra2_4a}.
15182 When @var{bayer} dithering is selected, this option defines the scale of the
15183 pattern (how much the crosshatch pattern is visible). A low value means more
15184 visible pattern for less banding, and higher value means less visible pattern
15185 at the cost of more banding.
15187 The option must be an integer value in the range [0,5]. Default is @var{2}.
15190 If set, define the zone to process
15194 Only the changing rectangle will be reprocessed. This is similar to GIF
15195 cropping/offsetting compression mechanism. This option can be useful for speed
15196 if only a part of the image is changing, and has use cases such as limiting the
15197 scope of the error diffusal @option{dither} to the rectangle that bounds the
15198 moving scene (it leads to more deterministic output if the scene doesn't change
15199 much, and as a result less moving noise and better GIF compression).
15202 Default is @var{none}.
15205 Take new palette for each output frame.
15207 @item alpha_threshold
15208 Sets the alpha threshold for transparency. Alpha values above this threshold
15209 will be treated as completely opaque, and values below this threshold will be
15210 treated as completely transparent.
15212 The option must be an integer value in the range [0,255]. Default is @var{128}.
15215 @subsection Examples
15219 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15220 using @command{ffmpeg}:
15222 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15226 @section perspective
15228 Correct perspective of video not recorded perpendicular to the screen.
15230 A description of the accepted parameters follows.
15241 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15242 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15243 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15244 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15245 then the corners of the source will be sent to the specified coordinates.
15247 The expressions can use the following variables:
15252 the width and height of video frame.
15256 Output frame count.
15259 @item interpolation
15260 Set interpolation for perspective correction.
15262 It accepts the following values:
15268 Default value is @samp{linear}.
15271 Set interpretation of coordinate options.
15273 It accepts the following values:
15277 Send point in the source specified by the given coordinates to
15278 the corners of the destination.
15280 @item 1, destination
15282 Send the corners of the source to the point in the destination specified
15283 by the given coordinates.
15285 Default value is @samp{source}.
15289 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15291 It accepts the following values:
15294 only evaluate expressions once during the filter initialization or
15295 when a command is processed
15298 evaluate expressions for each incoming frame
15301 Default value is @samp{init}.
15306 Delay interlaced video by one field time so that the field order changes.
15308 The intended use is to fix PAL movies that have been captured with the
15309 opposite field order to the film-to-video transfer.
15311 A description of the accepted parameters follows.
15317 It accepts the following values:
15320 Capture field order top-first, transfer bottom-first.
15321 Filter will delay the bottom field.
15324 Capture field order bottom-first, transfer top-first.
15325 Filter will delay the top field.
15328 Capture and transfer with the same field order. This mode only exists
15329 for the documentation of the other options to refer to, but if you
15330 actually select it, the filter will faithfully do nothing.
15333 Capture field order determined automatically by field flags, transfer
15335 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15336 basis using field flags. If no field information is available,
15337 then this works just like @samp{u}.
15340 Capture unknown or varying, transfer opposite.
15341 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15342 analyzing the images and selecting the alternative that produces best
15343 match between the fields.
15346 Capture top-first, transfer unknown or varying.
15347 Filter selects among @samp{t} and @samp{p} using image analysis.
15350 Capture bottom-first, transfer unknown or varying.
15351 Filter selects among @samp{b} and @samp{p} using image analysis.
15354 Capture determined by field flags, transfer unknown or varying.
15355 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15356 image analysis. If no field information is available, then this works just
15357 like @samp{U}. This is the default mode.
15360 Both capture and transfer unknown or varying.
15361 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15365 @section photosensitivity
15366 Reduce various flashes in video, so to help users with epilepsy.
15368 It accepts the following options:
15371 Set how many frames to use when filtering. Default is 30.
15374 Set detection threshold factor. Default is 1.
15378 Set how many pixels to skip when sampling frames. Default is 1.
15379 Allowed range is from 1 to 1024.
15382 Leave frames unchanged. Default is disabled.
15385 @section pixdesctest
15387 Pixel format descriptor test filter, mainly useful for internal
15388 testing. The output video should be equal to the input video.
15392 format=monow, pixdesctest
15395 can be used to test the monowhite pixel format descriptor definition.
15399 Display sample values of color channels. Mainly useful for checking color
15400 and levels. Minimum supported resolution is 640x480.
15402 The filters accept the following options:
15406 Set scope X position, relative offset on X axis.
15409 Set scope Y position, relative offset on Y axis.
15418 Set window opacity. This window also holds statistics about pixel area.
15421 Set window X position, relative offset on X axis.
15424 Set window Y position, relative offset on Y axis.
15429 Enable the specified chain of postprocessing subfilters using libpostproc. This
15430 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15431 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15432 Each subfilter and some options have a short and a long name that can be used
15433 interchangeably, i.e. dr/dering are the same.
15435 The filters accept the following options:
15439 Set postprocessing subfilters string.
15442 All subfilters share common options to determine their scope:
15446 Honor the quality commands for this subfilter.
15449 Do chrominance filtering, too (default).
15452 Do luminance filtering only (no chrominance).
15455 Do chrominance filtering only (no luminance).
15458 These options can be appended after the subfilter name, separated by a '|'.
15460 Available subfilters are:
15463 @item hb/hdeblock[|difference[|flatness]]
15464 Horizontal deblocking filter
15467 Difference factor where higher values mean more deblocking (default: @code{32}).
15469 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15472 @item vb/vdeblock[|difference[|flatness]]
15473 Vertical deblocking filter
15476 Difference factor where higher values mean more deblocking (default: @code{32}).
15478 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15481 @item ha/hadeblock[|difference[|flatness]]
15482 Accurate horizontal deblocking filter
15485 Difference factor where higher values mean more deblocking (default: @code{32}).
15487 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15490 @item va/vadeblock[|difference[|flatness]]
15491 Accurate vertical deblocking filter
15494 Difference factor where higher values mean more deblocking (default: @code{32}).
15496 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15500 The horizontal and vertical deblocking filters share the difference and
15501 flatness values so you cannot set different horizontal and vertical
15505 @item h1/x1hdeblock
15506 Experimental horizontal deblocking filter
15508 @item v1/x1vdeblock
15509 Experimental vertical deblocking filter
15514 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15517 larger -> stronger filtering
15519 larger -> stronger filtering
15521 larger -> stronger filtering
15524 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15527 Stretch luminance to @code{0-255}.
15530 @item lb/linblenddeint
15531 Linear blend deinterlacing filter that deinterlaces the given block by
15532 filtering all lines with a @code{(1 2 1)} filter.
15534 @item li/linipoldeint
15535 Linear interpolating deinterlacing filter that deinterlaces the given block by
15536 linearly interpolating every second line.
15538 @item ci/cubicipoldeint
15539 Cubic interpolating deinterlacing filter deinterlaces the given block by
15540 cubically interpolating every second line.
15542 @item md/mediandeint
15543 Median deinterlacing filter that deinterlaces the given block by applying a
15544 median filter to every second line.
15546 @item fd/ffmpegdeint
15547 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15548 second line with a @code{(-1 4 2 4 -1)} filter.
15551 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15552 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15554 @item fq/forceQuant[|quantizer]
15555 Overrides the quantizer table from the input with the constant quantizer you
15563 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15566 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15569 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15572 @subsection Examples
15576 Apply horizontal and vertical deblocking, deringing and automatic
15577 brightness/contrast:
15583 Apply default filters without brightness/contrast correction:
15589 Apply default filters and temporal denoiser:
15591 pp=default/tmpnoise|1|2|3
15595 Apply deblocking on luminance only, and switch vertical deblocking on or off
15596 automatically depending on available CPU time:
15603 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15604 similar to spp = 6 with 7 point DCT, where only the center sample is
15607 The filter accepts the following options:
15611 Force a constant quantization parameter. It accepts an integer in range
15612 0 to 63. If not set, the filter will use the QP from the video stream
15616 Set thresholding mode. Available modes are:
15620 Set hard thresholding.
15622 Set soft thresholding (better de-ringing effect, but likely blurrier).
15624 Set medium thresholding (good results, default).
15628 @section premultiply
15629 Apply alpha premultiply effect to input video stream using first plane
15630 of second stream as alpha.
15632 Both streams must have same dimensions and same pixel format.
15634 The filter accepts the following option:
15638 Set which planes will be processed, unprocessed planes will be copied.
15639 By default value 0xf, all planes will be processed.
15642 Do not require 2nd input for processing, instead use alpha plane from input stream.
15646 Apply prewitt operator to input video stream.
15648 The filter accepts the following option:
15652 Set which planes will be processed, unprocessed planes will be copied.
15653 By default value 0xf, all planes will be processed.
15656 Set value which will be multiplied with filtered result.
15659 Set value which will be added to filtered result.
15662 @section pseudocolor
15664 Alter frame colors in video with pseudocolors.
15666 This filter accepts the following options:
15670 set pixel first component expression
15673 set pixel second component expression
15676 set pixel third component expression
15679 set pixel fourth component expression, corresponds to the alpha component
15682 set component to use as base for altering colors
15685 Each of them specifies the expression to use for computing the lookup table for
15686 the corresponding pixel component values.
15688 The expressions can contain the following constants and functions:
15693 The input width and height.
15696 The input value for the pixel component.
15698 @item ymin, umin, vmin, amin
15699 The minimum allowed component value.
15701 @item ymax, umax, vmax, amax
15702 The maximum allowed component value.
15705 All expressions default to "val".
15707 @subsection Examples
15711 Change too high luma values to gradient:
15713 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'"
15719 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
15720 Ratio) between two input videos.
15722 This filter takes in input two input videos, the first input is
15723 considered the "main" source and is passed unchanged to the
15724 output. The second input is used as a "reference" video for computing
15727 Both video inputs must have the same resolution and pixel format for
15728 this filter to work correctly. Also it assumes that both inputs
15729 have the same number of frames, which are compared one by one.
15731 The obtained average PSNR is printed through the logging system.
15733 The filter stores the accumulated MSE (mean squared error) of each
15734 frame, and at the end of the processing it is averaged across all frames
15735 equally, and the following formula is applied to obtain the PSNR:
15738 PSNR = 10*log10(MAX^2/MSE)
15741 Where MAX is the average of the maximum values of each component of the
15744 The description of the accepted parameters follows.
15747 @item stats_file, f
15748 If specified the filter will use the named file to save the PSNR of
15749 each individual frame. When filename equals "-" the data is sent to
15752 @item stats_version
15753 Specifies which version of the stats file format to use. Details of
15754 each format are written below.
15755 Default value is 1.
15757 @item stats_add_max
15758 Determines whether the max value is output to the stats log.
15759 Default value is 0.
15760 Requires stats_version >= 2. If this is set and stats_version < 2,
15761 the filter will return an error.
15764 This filter also supports the @ref{framesync} options.
15766 The file printed if @var{stats_file} is selected, contains a sequence of
15767 key/value pairs of the form @var{key}:@var{value} for each compared
15770 If a @var{stats_version} greater than 1 is specified, a header line precedes
15771 the list of per-frame-pair stats, with key value pairs following the frame
15772 format with the following parameters:
15775 @item psnr_log_version
15776 The version of the log file format. Will match @var{stats_version}.
15779 A comma separated list of the per-frame-pair parameters included in
15783 A description of each shown per-frame-pair parameter follows:
15787 sequential number of the input frame, starting from 1
15790 Mean Square Error pixel-by-pixel average difference of the compared
15791 frames, averaged over all the image components.
15793 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
15794 Mean Square Error pixel-by-pixel average difference of the compared
15795 frames for the component specified by the suffix.
15797 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
15798 Peak Signal to Noise ratio of the compared frames for the component
15799 specified by the suffix.
15801 @item max_avg, max_y, max_u, max_v
15802 Maximum allowed value for each channel, and average over all
15806 @subsection Examples
15811 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
15812 [main][ref] psnr="stats_file=stats.log" [out]
15815 On this example the input file being processed is compared with the
15816 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
15817 is stored in @file{stats.log}.
15820 Another example with different containers:
15822 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 -
15829 Pulldown reversal (inverse telecine) filter, capable of handling mixed
15830 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
15833 The pullup filter is designed to take advantage of future context in making
15834 its decisions. This filter is stateless in the sense that it does not lock
15835 onto a pattern to follow, but it instead looks forward to the following
15836 fields in order to identify matches and rebuild progressive frames.
15838 To produce content with an even framerate, insert the fps filter after
15839 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
15840 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
15842 The filter accepts the following options:
15849 These options set the amount of "junk" to ignore at the left, right, top, and
15850 bottom of the image, respectively. Left and right are in units of 8 pixels,
15851 while top and bottom are in units of 2 lines.
15852 The default is 8 pixels on each side.
15855 Set the strict breaks. Setting this option to 1 will reduce the chances of
15856 filter generating an occasional mismatched frame, but it may also cause an
15857 excessive number of frames to be dropped during high motion sequences.
15858 Conversely, setting it to -1 will make filter match fields more easily.
15859 This may help processing of video where there is slight blurring between
15860 the fields, but may also cause there to be interlaced frames in the output.
15861 Default value is @code{0}.
15864 Set the metric plane to use. It accepts the following values:
15870 Use chroma blue plane.
15873 Use chroma red plane.
15876 This option may be set to use chroma plane instead of the default luma plane
15877 for doing filter's computations. This may improve accuracy on very clean
15878 source material, but more likely will decrease accuracy, especially if there
15879 is chroma noise (rainbow effect) or any grayscale video.
15880 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
15881 load and make pullup usable in realtime on slow machines.
15884 For best results (without duplicated frames in the output file) it is
15885 necessary to change the output frame rate. For example, to inverse
15886 telecine NTSC input:
15888 ffmpeg -i input -vf pullup -r 24000/1001 ...
15893 Change video quantization parameters (QP).
15895 The filter accepts the following option:
15899 Set expression for quantization parameter.
15902 The expression is evaluated through the eval API and can contain, among others,
15903 the following constants:
15907 1 if index is not 129, 0 otherwise.
15910 Sequential index starting from -129 to 128.
15913 @subsection Examples
15917 Some equation like:
15925 Flush video frames from internal cache of frames into a random order.
15926 No frame is discarded.
15927 Inspired by @ref{frei0r} nervous filter.
15931 Set size in number of frames of internal cache, in range from @code{2} to
15932 @code{512}. Default is @code{30}.
15935 Set seed for random number generator, must be an integer included between
15936 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
15937 less than @code{0}, the filter will try to use a good random seed on a
15941 @section readeia608
15943 Read closed captioning (EIA-608) information from the top lines of a video frame.
15945 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
15946 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
15947 with EIA-608 data (starting from 0). A description of each metadata value follows:
15950 @item lavfi.readeia608.X.cc
15951 The two bytes stored as EIA-608 data (printed in hexadecimal).
15953 @item lavfi.readeia608.X.line
15954 The number of the line on which the EIA-608 data was identified and read.
15957 This filter accepts the following options:
15961 Set the line to start scanning for EIA-608 data. Default is @code{0}.
15964 Set the line to end scanning for EIA-608 data. Default is @code{29}.
15967 Set the ratio of width reserved for sync code detection.
15968 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
15971 Enable checking the parity bit. In the event of a parity error, the filter will output
15972 @code{0x00} for that character. Default is false.
15975 Lowpass lines prior to further processing. Default is enabled.
15978 @subsection Commands
15980 This filter supports the all above options as @ref{commands}.
15982 @subsection Examples
15986 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
15988 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
15994 Read vertical interval timecode (VITC) information from the top lines of a
15997 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
15998 timecode value, if a valid timecode has been detected. Further metadata key
15999 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16000 timecode data has been found or not.
16002 This filter accepts the following options:
16006 Set the maximum number of lines to scan for VITC data. If the value is set to
16007 @code{-1} the full video frame is scanned. Default is @code{45}.
16010 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16011 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16014 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16015 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16018 @subsection Examples
16022 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16023 draw @code{--:--:--:--} as a placeholder:
16025 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16031 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16033 Destination pixel at position (X, Y) will be picked from source (x, y) position
16034 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16035 value for pixel will be used for destination pixel.
16037 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16038 will have Xmap/Ymap video stream dimensions.
16039 Xmap and Ymap input video streams are 16bit depth, single channel.
16043 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16044 Default is @code{color}.
16047 Specify the color of the unmapped pixels. For the syntax of this option,
16048 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16049 manual,ffmpeg-utils}. Default color is @code{black}.
16052 @section removegrain
16054 The removegrain filter is a spatial denoiser for progressive video.
16058 Set mode for the first plane.
16061 Set mode for the second plane.
16064 Set mode for the third plane.
16067 Set mode for the fourth plane.
16070 Range of mode is from 0 to 24. Description of each mode follows:
16074 Leave input plane unchanged. Default.
16077 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16080 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16083 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16086 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16087 This is equivalent to a median filter.
16090 Line-sensitive clipping giving the minimal change.
16093 Line-sensitive clipping, intermediate.
16096 Line-sensitive clipping, intermediate.
16099 Line-sensitive clipping, intermediate.
16102 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16105 Replaces the target pixel with the closest neighbour.
16108 [1 2 1] horizontal and vertical kernel blur.
16114 Bob mode, interpolates top field from the line where the neighbours
16115 pixels are the closest.
16118 Bob mode, interpolates bottom field from the line where the neighbours
16119 pixels are the closest.
16122 Bob mode, interpolates top field. Same as 13 but with a more complicated
16123 interpolation formula.
16126 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16127 interpolation formula.
16130 Clips the pixel with the minimum and maximum of respectively the maximum and
16131 minimum of each pair of opposite neighbour pixels.
16134 Line-sensitive clipping using opposite neighbours whose greatest distance from
16135 the current pixel is minimal.
16138 Replaces the pixel with the average of its 8 neighbours.
16141 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16144 Clips pixels using the averages of opposite neighbour.
16147 Same as mode 21 but simpler and faster.
16150 Small edge and halo removal, but reputed useless.
16156 @section removelogo
16158 Suppress a TV station logo, using an image file to determine which
16159 pixels comprise the logo. It works by filling in the pixels that
16160 comprise the logo with neighboring pixels.
16162 The filter accepts the following options:
16166 Set the filter bitmap file, which can be any image format supported by
16167 libavformat. The width and height of the image file must match those of the
16168 video stream being processed.
16171 Pixels in the provided bitmap image with a value of zero are not
16172 considered part of the logo, non-zero pixels are considered part of
16173 the logo. If you use white (255) for the logo and black (0) for the
16174 rest, you will be safe. For making the filter bitmap, it is
16175 recommended to take a screen capture of a black frame with the logo
16176 visible, and then using a threshold filter followed by the erode
16177 filter once or twice.
16179 If needed, little splotches can be fixed manually. Remember that if
16180 logo pixels are not covered, the filter quality will be much
16181 reduced. Marking too many pixels as part of the logo does not hurt as
16182 much, but it will increase the amount of blurring needed to cover over
16183 the image and will destroy more information than necessary, and extra
16184 pixels will slow things down on a large logo.
16186 @section repeatfields
16188 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16189 fields based on its value.
16193 Reverse a video clip.
16195 Warning: This filter requires memory to buffer the entire clip, so trimming
16198 @subsection Examples
16202 Take the first 5 seconds of a clip, and reverse it.
16209 Shift R/G/B/A pixels horizontally and/or vertically.
16211 The filter accepts the following options:
16214 Set amount to shift red horizontally.
16216 Set amount to shift red vertically.
16218 Set amount to shift green horizontally.
16220 Set amount to shift green vertically.
16222 Set amount to shift blue horizontally.
16224 Set amount to shift blue vertically.
16226 Set amount to shift alpha horizontally.
16228 Set amount to shift alpha vertically.
16230 Set edge mode, can be @var{smear}, default, or @var{warp}.
16233 @subsection Commands
16235 This filter supports the all above options as @ref{commands}.
16238 Apply roberts cross operator to input video stream.
16240 The filter accepts the following option:
16244 Set which planes will be processed, unprocessed planes will be copied.
16245 By default value 0xf, all planes will be processed.
16248 Set value which will be multiplied with filtered result.
16251 Set value which will be added to filtered result.
16256 Rotate video by an arbitrary angle expressed in radians.
16258 The filter accepts the following options:
16260 A description of the optional parameters follows.
16263 Set an expression for the angle by which to rotate the input video
16264 clockwise, expressed as a number of radians. A negative value will
16265 result in a counter-clockwise rotation. By default it is set to "0".
16267 This expression is evaluated for each frame.
16270 Set the output width expression, default value is "iw".
16271 This expression is evaluated just once during configuration.
16274 Set the output height expression, default value is "ih".
16275 This expression is evaluated just once during configuration.
16278 Enable bilinear interpolation if set to 1, a value of 0 disables
16279 it. Default value is 1.
16282 Set the color used to fill the output area not covered by the rotated
16283 image. For the general syntax of this option, check the
16284 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16285 If the special value "none" is selected then no
16286 background is printed (useful for example if the background is never shown).
16288 Default value is "black".
16291 The expressions for the angle and the output size can contain the
16292 following constants and functions:
16296 sequential number of the input frame, starting from 0. It is always NAN
16297 before the first frame is filtered.
16300 time in seconds of the input frame, it is set to 0 when the filter is
16301 configured. It is always NAN before the first frame is filtered.
16305 horizontal and vertical chroma subsample values. For example for the
16306 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16310 the input video width and height
16314 the output width and height, that is the size of the padded area as
16315 specified by the @var{width} and @var{height} expressions
16319 the minimal width/height required for completely containing the input
16320 video rotated by @var{a} radians.
16322 These are only available when computing the @option{out_w} and
16323 @option{out_h} expressions.
16326 @subsection Examples
16330 Rotate the input by PI/6 radians clockwise:
16336 Rotate the input by PI/6 radians counter-clockwise:
16342 Rotate the input by 45 degrees clockwise:
16348 Apply a constant rotation with period T, starting from an angle of PI/3:
16350 rotate=PI/3+2*PI*t/T
16354 Make the input video rotation oscillating with a period of T
16355 seconds and an amplitude of A radians:
16357 rotate=A*sin(2*PI/T*t)
16361 Rotate the video, output size is chosen so that the whole rotating
16362 input video is always completely contained in the output:
16364 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16368 Rotate the video, reduce the output size so that no background is ever
16371 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16375 @subsection Commands
16377 The filter supports the following commands:
16381 Set the angle expression.
16382 The command accepts the same syntax of the corresponding option.
16384 If the specified expression is not valid, it is kept at its current
16390 Apply Shape Adaptive Blur.
16392 The filter accepts the following options:
16395 @item luma_radius, lr
16396 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16397 value is 1.0. A greater value will result in a more blurred image, and
16398 in slower processing.
16400 @item luma_pre_filter_radius, lpfr
16401 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16404 @item luma_strength, ls
16405 Set luma maximum difference between pixels to still be considered, must
16406 be a value in the 0.1-100.0 range, default value is 1.0.
16408 @item chroma_radius, cr
16409 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16410 greater value will result in a more blurred image, and in slower
16413 @item chroma_pre_filter_radius, cpfr
16414 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16416 @item chroma_strength, cs
16417 Set chroma maximum difference between pixels to still be considered,
16418 must be a value in the -0.9-100.0 range.
16421 Each chroma option value, if not explicitly specified, is set to the
16422 corresponding luma option value.
16427 Scale (resize) the input video, using the libswscale library.
16429 The scale filter forces the output display aspect ratio to be the same
16430 of the input, by changing the output sample aspect ratio.
16432 If the input image format is different from the format requested by
16433 the next filter, the scale filter will convert the input to the
16436 @subsection Options
16437 The filter accepts the following options, or any of the options
16438 supported by the libswscale scaler.
16440 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16441 the complete list of scaler options.
16446 Set the output video dimension expression. Default value is the input
16449 If the @var{width} or @var{w} value is 0, the input width is used for
16450 the output. If the @var{height} or @var{h} value is 0, the input height
16451 is used for the output.
16453 If one and only one of the values is -n with n >= 1, the scale filter
16454 will use a value that maintains the aspect ratio of the input image,
16455 calculated from the other specified dimension. After that it will,
16456 however, make sure that the calculated dimension is divisible by n and
16457 adjust the value if necessary.
16459 If both values are -n with n >= 1, the behavior will be identical to
16460 both values being set to 0 as previously detailed.
16462 See below for the list of accepted constants for use in the dimension
16466 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16470 Only evaluate expressions once during the filter initialization or when a command is processed.
16473 Evaluate expressions for each incoming frame.
16477 Default value is @samp{init}.
16481 Set the interlacing mode. It accepts the following values:
16485 Force interlaced aware scaling.
16488 Do not apply interlaced scaling.
16491 Select interlaced aware scaling depending on whether the source frames
16492 are flagged as interlaced or not.
16495 Default value is @samp{0}.
16498 Set libswscale scaling flags. See
16499 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16500 complete list of values. If not explicitly specified the filter applies
16504 @item param0, param1
16505 Set libswscale input parameters for scaling algorithms that need them. See
16506 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16507 complete documentation. If not explicitly specified the filter applies
16513 Set the video size. For the syntax of this option, check the
16514 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16516 @item in_color_matrix
16517 @item out_color_matrix
16518 Set in/output YCbCr color space type.
16520 This allows the autodetected value to be overridden as well as allows forcing
16521 a specific value used for the output and encoder.
16523 If not specified, the color space type depends on the pixel format.
16529 Choose automatically.
16532 Format conforming to International Telecommunication Union (ITU)
16533 Recommendation BT.709.
16536 Set color space conforming to the United States Federal Communications
16537 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16542 Set color space conforming to:
16546 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16549 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16552 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16557 Set color space conforming to SMPTE ST 240:1999.
16560 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16565 Set in/output YCbCr sample range.
16567 This allows the autodetected value to be overridden as well as allows forcing
16568 a specific value used for the output and encoder. If not specified, the
16569 range depends on the pixel format. Possible values:
16573 Choose automatically.
16576 Set full range (0-255 in case of 8-bit luma).
16578 @item mpeg/limited/tv
16579 Set "MPEG" range (16-235 in case of 8-bit luma).
16582 @item force_original_aspect_ratio
16583 Enable decreasing or increasing output video width or height if necessary to
16584 keep the original aspect ratio. Possible values:
16588 Scale the video as specified and disable this feature.
16591 The output video dimensions will automatically be decreased if needed.
16594 The output video dimensions will automatically be increased if needed.
16598 One useful instance of this option is that when you know a specific device's
16599 maximum allowed resolution, you can use this to limit the output video to
16600 that, while retaining the aspect ratio. For example, device A allows
16601 1280x720 playback, and your video is 1920x800. Using this option (set it to
16602 decrease) and specifying 1280x720 to the command line makes the output
16605 Please note that this is a different thing than specifying -1 for @option{w}
16606 or @option{h}, you still need to specify the output resolution for this option
16609 @item force_divisible_by
16610 Ensures that both the output dimensions, width and height, are divisible by the
16611 given integer when used together with @option{force_original_aspect_ratio}. This
16612 works similar to using @code{-n} in the @option{w} and @option{h} options.
16614 This option respects the value set for @option{force_original_aspect_ratio},
16615 increasing or decreasing the resolution accordingly. The video's aspect ratio
16616 may be slightly modified.
16618 This option can be handy if you need to have a video fit within or exceed
16619 a defined resolution using @option{force_original_aspect_ratio} but also have
16620 encoder restrictions on width or height divisibility.
16624 The values of the @option{w} and @option{h} options are expressions
16625 containing the following constants:
16630 The input width and height
16634 These are the same as @var{in_w} and @var{in_h}.
16638 The output (scaled) width and height
16642 These are the same as @var{out_w} and @var{out_h}
16645 The same as @var{iw} / @var{ih}
16648 input sample aspect ratio
16651 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16655 horizontal and vertical input chroma subsample values. For example for the
16656 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16660 horizontal and vertical output chroma subsample values. For example for the
16661 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16664 The (sequential) number of the input frame, starting from 0.
16665 Only available with @code{eval=frame}.
16668 The presentation timestamp of the input frame, expressed as a number of
16669 seconds. Only available with @code{eval=frame}.
16672 The position (byte offset) of the frame in the input stream, or NaN if
16673 this information is unavailable and/or meaningless (for example in case of synthetic video).
16674 Only available with @code{eval=frame}.
16677 @subsection Examples
16681 Scale the input video to a size of 200x100
16686 This is equivalent to:
16697 Specify a size abbreviation for the output size:
16702 which can also be written as:
16708 Scale the input to 2x:
16710 scale=w=2*iw:h=2*ih
16714 The above is the same as:
16716 scale=2*in_w:2*in_h
16720 Scale the input to 2x with forced interlaced scaling:
16722 scale=2*iw:2*ih:interl=1
16726 Scale the input to half size:
16728 scale=w=iw/2:h=ih/2
16732 Increase the width, and set the height to the same size:
16738 Seek Greek harmony:
16745 Increase the height, and set the width to 3/2 of the height:
16747 scale=w=3/2*oh:h=3/5*ih
16751 Increase the size, making the size a multiple of the chroma
16754 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
16758 Increase the width to a maximum of 500 pixels,
16759 keeping the same aspect ratio as the input:
16761 scale=w='min(500\, iw*3/2):h=-1'
16765 Make pixels square by combining scale and setsar:
16767 scale='trunc(ih*dar):ih',setsar=1/1
16771 Make pixels square by combining scale and setsar,
16772 making sure the resulting resolution is even (required by some codecs):
16774 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
16778 @subsection Commands
16780 This filter supports the following commands:
16784 Set the output video dimension expression.
16785 The command accepts the same syntax of the corresponding option.
16787 If the specified expression is not valid, it is kept at its current
16793 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
16794 format conversion on CUDA video frames. Setting the output width and height
16795 works in the same way as for the @var{scale} filter.
16797 The following additional options are accepted:
16800 The pixel format of the output CUDA frames. If set to the string "same" (the
16801 default), the input format will be kept. Note that automatic format negotiation
16802 and conversion is not yet supported for hardware frames
16805 The interpolation algorithm used for resizing. One of the following:
16812 @item cubic2p_bspline
16813 2-parameter cubic (B=1, C=0)
16815 @item cubic2p_catmullrom
16816 2-parameter cubic (B=0, C=1/2)
16818 @item cubic2p_b05c03
16819 2-parameter cubic (B=1/2, C=3/10)
16827 @item force_original_aspect_ratio
16828 Enable decreasing or increasing output video width or height if necessary to
16829 keep the original aspect ratio. Possible values:
16833 Scale the video as specified and disable this feature.
16836 The output video dimensions will automatically be decreased if needed.
16839 The output video dimensions will automatically be increased if needed.
16843 One useful instance of this option is that when you know a specific device's
16844 maximum allowed resolution, you can use this to limit the output video to
16845 that, while retaining the aspect ratio. For example, device A allows
16846 1280x720 playback, and your video is 1920x800. Using this option (set it to
16847 decrease) and specifying 1280x720 to the command line makes the output
16850 Please note that this is a different thing than specifying -1 for @option{w}
16851 or @option{h}, you still need to specify the output resolution for this option
16854 @item force_divisible_by
16855 Ensures that both the output dimensions, width and height, are divisible by the
16856 given integer when used together with @option{force_original_aspect_ratio}. This
16857 works similar to using @code{-n} in the @option{w} and @option{h} options.
16859 This option respects the value set for @option{force_original_aspect_ratio},
16860 increasing or decreasing the resolution accordingly. The video's aspect ratio
16861 may be slightly modified.
16863 This option can be handy if you need to have a video fit within or exceed
16864 a defined resolution using @option{force_original_aspect_ratio} but also have
16865 encoder restrictions on width or height divisibility.
16871 Scale (resize) the input video, based on a reference video.
16873 See the scale filter for available options, scale2ref supports the same but
16874 uses the reference video instead of the main input as basis. scale2ref also
16875 supports the following additional constants for the @option{w} and
16876 @option{h} options:
16881 The main input video's width and height
16884 The same as @var{main_w} / @var{main_h}
16887 The main input video's sample aspect ratio
16889 @item main_dar, mdar
16890 The main input video's display aspect ratio. Calculated from
16891 @code{(main_w / main_h) * main_sar}.
16895 The main input video's horizontal and vertical chroma subsample values.
16896 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
16900 The (sequential) number of the main input frame, starting from 0.
16901 Only available with @code{eval=frame}.
16904 The presentation timestamp of the main input frame, expressed as a number of
16905 seconds. Only available with @code{eval=frame}.
16908 The position (byte offset) of the frame in the main input stream, or NaN if
16909 this information is unavailable and/or meaningless (for example in case of synthetic video).
16910 Only available with @code{eval=frame}.
16913 @subsection Examples
16917 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
16919 'scale2ref[b][a];[a][b]overlay'
16923 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
16925 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
16929 @subsection Commands
16931 This filter supports the following commands:
16935 Set the output video dimension expression.
16936 The command accepts the same syntax of the corresponding option.
16938 If the specified expression is not valid, it is kept at its current
16943 Scroll input video horizontally and/or vertically by constant speed.
16945 The filter accepts the following options:
16947 @item horizontal, h
16948 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
16949 Negative values changes scrolling direction.
16952 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
16953 Negative values changes scrolling direction.
16956 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
16959 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
16962 @subsection Commands
16964 This filter supports the following @ref{commands}:
16966 @item horizontal, h
16967 Set the horizontal scrolling speed.
16969 Set the vertical scrolling speed.
16975 Detect video scene change.
16977 This filter sets frame metadata with mafd between frame, the scene score, and
16978 forward the frame to the next filter, so they can use these metadata to detect
16979 scene change or others.
16981 In addition, this filter logs a message and sets frame metadata when it detects
16982 a scene change by @option{threshold}.
16984 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
16986 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
16987 to detect scene change.
16989 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
16990 detect scene change with @option{threshold}.
16992 The filter accepts the following options:
16996 Set the scene change detection threshold as a percentage of maximum change. Good
16997 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17000 Default value is @code{10.}.
17003 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17004 You can enable it if you want to get snapshot of scene change frames only.
17007 @anchor{selectivecolor}
17008 @section selectivecolor
17010 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17011 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17012 by the "purity" of the color (that is, how saturated it already is).
17014 This filter is similar to the Adobe Photoshop Selective Color tool.
17016 The filter accepts the following options:
17019 @item correction_method
17020 Select color correction method.
17022 Available values are:
17025 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17028 Specified adjustments are relative to the original component value.
17030 Default is @code{absolute}.
17032 Adjustments for red pixels (pixels where the red component is the maximum)
17034 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17036 Adjustments for green pixels (pixels where the green component is the maximum)
17038 Adjustments for cyan pixels (pixels where the red component is the minimum)
17040 Adjustments for blue pixels (pixels where the blue component is the maximum)
17042 Adjustments for magenta pixels (pixels where the green component is the minimum)
17044 Adjustments for white pixels (pixels where all components are greater than 128)
17046 Adjustments for all pixels except pure black and pure white
17048 Adjustments for black pixels (pixels where all components are lesser than 128)
17050 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17053 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17054 4 space separated floating point adjustment values in the [-1,1] range,
17055 respectively to adjust the amount of cyan, magenta, yellow and black for the
17056 pixels of its range.
17058 @subsection Examples
17062 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17063 increase magenta by 27% in blue areas:
17065 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17069 Use a Photoshop selective color preset:
17071 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17075 @anchor{separatefields}
17076 @section separatefields
17078 The @code{separatefields} takes a frame-based video input and splits
17079 each frame into its components fields, producing a new half height clip
17080 with twice the frame rate and twice the frame count.
17082 This filter use field-dominance information in frame to decide which
17083 of each pair of fields to place first in the output.
17084 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17086 @section setdar, setsar
17088 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17091 This is done by changing the specified Sample (aka Pixel) Aspect
17092 Ratio, according to the following equation:
17094 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17097 Keep in mind that the @code{setdar} filter does not modify the pixel
17098 dimensions of the video frame. Also, the display aspect ratio set by
17099 this filter may be changed by later filters in the filterchain,
17100 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17103 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17104 the filter output video.
17106 Note that as a consequence of the application of this filter, the
17107 output display aspect ratio will change according to the equation
17110 Keep in mind that the sample aspect ratio set by the @code{setsar}
17111 filter may be changed by later filters in the filterchain, e.g. if
17112 another "setsar" or a "setdar" filter is applied.
17114 It accepts the following parameters:
17117 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17118 Set the aspect ratio used by the filter.
17120 The parameter can be a floating point number string, an expression, or
17121 a string of the form @var{num}:@var{den}, where @var{num} and
17122 @var{den} are the numerator and denominator of the aspect ratio. If
17123 the parameter is not specified, it is assumed the value "0".
17124 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17128 Set the maximum integer value to use for expressing numerator and
17129 denominator when reducing the expressed aspect ratio to a rational.
17130 Default value is @code{100}.
17134 The parameter @var{sar} is an expression containing
17135 the following constants:
17139 These are approximated values for the mathematical constants e
17140 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17143 The input width and height.
17146 These are the same as @var{w} / @var{h}.
17149 The input sample aspect ratio.
17152 The input display aspect ratio. It is the same as
17153 (@var{w} / @var{h}) * @var{sar}.
17156 Horizontal and vertical chroma subsample values. For example, for the
17157 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17160 @subsection Examples
17165 To change the display aspect ratio to 16:9, specify one of the following:
17172 To change the sample aspect ratio to 10:11, specify:
17178 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17179 1000 in the aspect ratio reduction, use the command:
17181 setdar=ratio=16/9:max=1000
17189 Force field for the output video frame.
17191 The @code{setfield} filter marks the interlace type field for the
17192 output frames. It does not change the input frame, but only sets the
17193 corresponding property, which affects how the frame is treated by
17194 following filters (e.g. @code{fieldorder} or @code{yadif}).
17196 The filter accepts the following options:
17201 Available values are:
17205 Keep the same field property.
17208 Mark the frame as bottom-field-first.
17211 Mark the frame as top-field-first.
17214 Mark the frame as progressive.
17221 Force frame parameter for the output video frame.
17223 The @code{setparams} filter marks interlace and color range for the
17224 output frames. It does not change the input frame, but only sets the
17225 corresponding property, which affects how the frame is treated by
17230 Available values are:
17234 Keep the same field property (default).
17237 Mark the frame as bottom-field-first.
17240 Mark the frame as top-field-first.
17243 Mark the frame as progressive.
17247 Available values are:
17251 Keep the same color range property (default).
17253 @item unspecified, unknown
17254 Mark the frame as unspecified color range.
17256 @item limited, tv, mpeg
17257 Mark the frame as limited range.
17259 @item full, pc, jpeg
17260 Mark the frame as full range.
17263 @item color_primaries
17264 Set the color primaries.
17265 Available values are:
17269 Keep the same color primaries property (default).
17286 Set the color transfer.
17287 Available values are:
17291 Keep the same color trc property (default).
17313 Set the colorspace.
17314 Available values are:
17318 Keep the same colorspace property (default).
17331 @item chroma-derived-nc
17332 @item chroma-derived-c
17339 Show a line containing various information for each input video frame.
17340 The input video is not modified.
17342 This filter supports the following options:
17346 Calculate checksums of each plane. By default enabled.
17349 The shown line contains a sequence of key/value pairs of the form
17350 @var{key}:@var{value}.
17352 The following values are shown in the output:
17356 The (sequential) number of the input frame, starting from 0.
17359 The Presentation TimeStamp of the input frame, expressed as a number of
17360 time base units. The time base unit depends on the filter input pad.
17363 The Presentation TimeStamp of the input frame, expressed as a number of
17367 The position of the frame in the input stream, or -1 if this information is
17368 unavailable and/or meaningless (for example in case of synthetic video).
17371 The pixel format name.
17374 The sample aspect ratio of the input frame, expressed in the form
17375 @var{num}/@var{den}.
17378 The size of the input frame. For the syntax of this option, check the
17379 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17382 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17383 for bottom field first).
17386 This is 1 if the frame is a key frame, 0 otherwise.
17389 The picture type of the input frame ("I" for an I-frame, "P" for a
17390 P-frame, "B" for a B-frame, or "?" for an unknown type).
17391 Also refer to the documentation of the @code{AVPictureType} enum and of
17392 the @code{av_get_picture_type_char} function defined in
17393 @file{libavutil/avutil.h}.
17396 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17398 @item plane_checksum
17399 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17400 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17403 The mean value of pixels in each plane of the input frame, expressed in the form
17404 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17407 The standard deviation of pixel values in each plane of the input frame, expressed
17408 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17412 @section showpalette
17414 Displays the 256 colors palette of each frame. This filter is only relevant for
17415 @var{pal8} pixel format frames.
17417 It accepts the following option:
17421 Set the size of the box used to represent one palette color entry. Default is
17422 @code{30} (for a @code{30x30} pixel box).
17425 @section shuffleframes
17427 Reorder and/or duplicate and/or drop video frames.
17429 It accepts the following parameters:
17433 Set the destination indexes of input frames.
17434 This is space or '|' separated list of indexes that maps input frames to output
17435 frames. Number of indexes also sets maximal value that each index may have.
17436 '-1' index have special meaning and that is to drop frame.
17439 The first frame has the index 0. The default is to keep the input unchanged.
17441 @subsection Examples
17445 Swap second and third frame of every three frames of the input:
17447 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17451 Swap 10th and 1st frame of every ten frames of the input:
17453 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17457 @section shuffleplanes
17459 Reorder and/or duplicate video planes.
17461 It accepts the following parameters:
17466 The index of the input plane to be used as the first output plane.
17469 The index of the input plane to be used as the second output plane.
17472 The index of the input plane to be used as the third output plane.
17475 The index of the input plane to be used as the fourth output plane.
17479 The first plane has the index 0. The default is to keep the input unchanged.
17481 @subsection Examples
17485 Swap the second and third planes of the input:
17487 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17491 @anchor{signalstats}
17492 @section signalstats
17493 Evaluate various visual metrics that assist in determining issues associated
17494 with the digitization of analog video media.
17496 By default the filter will log these metadata values:
17500 Display the minimal Y value contained within the input frame. Expressed in
17504 Display the Y value at the 10% percentile within the input frame. Expressed in
17508 Display the average Y value within the input frame. Expressed in range of
17512 Display the Y value at the 90% percentile within the input frame. Expressed in
17516 Display the maximum Y value contained within the input frame. Expressed in
17520 Display the minimal U value contained within the input frame. Expressed in
17524 Display the U value at the 10% percentile within the input frame. Expressed in
17528 Display the average U value within the input frame. Expressed in range of
17532 Display the U value at the 90% percentile within the input frame. Expressed in
17536 Display the maximum U value contained within the input frame. Expressed in
17540 Display the minimal V value contained within the input frame. Expressed in
17544 Display the V value at the 10% percentile within the input frame. Expressed in
17548 Display the average V value within the input frame. Expressed in range of
17552 Display the V value at the 90% percentile within the input frame. Expressed in
17556 Display the maximum V value contained within the input frame. Expressed in
17560 Display the minimal saturation value contained within the input frame.
17561 Expressed in range of [0-~181.02].
17564 Display the saturation value at the 10% percentile within the input frame.
17565 Expressed in range of [0-~181.02].
17568 Display the average saturation value within the input frame. Expressed in range
17572 Display the saturation value at the 90% percentile within the input frame.
17573 Expressed in range of [0-~181.02].
17576 Display the maximum saturation value contained within the input frame.
17577 Expressed in range of [0-~181.02].
17580 Display the median value for hue within the input frame. Expressed in range of
17584 Display the average value for hue within the input frame. Expressed in range of
17588 Display the average of sample value difference between all values of the Y
17589 plane in the current frame and corresponding values of the previous input frame.
17590 Expressed in range of [0-255].
17593 Display the average of sample value difference between all values of the U
17594 plane in the current frame and corresponding values of the previous input frame.
17595 Expressed in range of [0-255].
17598 Display the average of sample value difference between all values of the V
17599 plane in the current frame and corresponding values of the previous input frame.
17600 Expressed in range of [0-255].
17603 Display bit depth of Y plane in current frame.
17604 Expressed in range of [0-16].
17607 Display bit depth of U plane in current frame.
17608 Expressed in range of [0-16].
17611 Display bit depth of V plane in current frame.
17612 Expressed in range of [0-16].
17615 The filter accepts the following options:
17621 @option{stat} specify an additional form of image analysis.
17622 @option{out} output video with the specified type of pixel highlighted.
17624 Both options accept the following values:
17628 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17629 unlike the neighboring pixels of the same field. Examples of temporal outliers
17630 include the results of video dropouts, head clogs, or tape tracking issues.
17633 Identify @var{vertical line repetition}. Vertical line repetition includes
17634 similar rows of pixels within a frame. In born-digital video vertical line
17635 repetition is common, but this pattern is uncommon in video digitized from an
17636 analog source. When it occurs in video that results from the digitization of an
17637 analog source it can indicate concealment from a dropout compensator.
17640 Identify pixels that fall outside of legal broadcast range.
17644 Set the highlight color for the @option{out} option. The default color is
17648 @subsection Examples
17652 Output data of various video metrics:
17654 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17658 Output specific data about the minimum and maximum values of the Y plane per frame:
17660 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17664 Playback video while highlighting pixels that are outside of broadcast range in red.
17666 ffplay example.mov -vf signalstats="out=brng:color=red"
17670 Playback video with signalstats metadata drawn over the frame.
17672 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17675 The contents of signalstat_drawtext.txt used in the command are:
17678 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17679 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17680 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17681 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17689 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17690 input. In this case the matching between the inputs can be calculated additionally.
17691 The filter always passes through the first input. The signature of each stream can
17692 be written into a file.
17694 It accepts the following options:
17698 Enable or disable the matching process.
17700 Available values are:
17704 Disable the calculation of a matching (default).
17706 Calculate the matching for the whole video and output whether the whole video
17707 matches or only parts.
17709 Calculate only until a matching is found or the video ends. Should be faster in
17714 Set the number of inputs. The option value must be a non negative integer.
17715 Default value is 1.
17718 Set the path to which the output is written. If there is more than one input,
17719 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
17720 integer), that will be replaced with the input number. If no filename is
17721 specified, no output will be written. This is the default.
17724 Choose the output format.
17726 Available values are:
17730 Use the specified binary representation (default).
17732 Use the specified xml representation.
17736 Set threshold to detect one word as similar. The option value must be an integer
17737 greater than zero. The default value is 9000.
17740 Set threshold to detect all words as similar. The option value must be an integer
17741 greater than zero. The default value is 60000.
17744 Set threshold to detect frames as similar. The option value must be an integer
17745 greater than zero. The default value is 116.
17748 Set the minimum length of a sequence in frames to recognize it as matching
17749 sequence. The option value must be a non negative integer value.
17750 The default value is 0.
17753 Set the minimum relation, that matching frames to all frames must have.
17754 The option value must be a double value between 0 and 1. The default value is 0.5.
17757 @subsection Examples
17761 To calculate the signature of an input video and store it in signature.bin:
17763 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
17767 To detect whether two videos match and store the signatures in XML format in
17768 signature0.xml and signature1.xml:
17770 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 -
17778 Blur the input video without impacting the outlines.
17780 It accepts the following options:
17783 @item luma_radius, lr
17784 Set the luma radius. The option value must be a float number in
17785 the range [0.1,5.0] that specifies the variance of the gaussian filter
17786 used to blur the image (slower if larger). Default value is 1.0.
17788 @item luma_strength, ls
17789 Set the luma strength. The option value must be a float number
17790 in the range [-1.0,1.0] that configures the blurring. A value included
17791 in [0.0,1.0] will blur the image whereas a value included in
17792 [-1.0,0.0] will sharpen the image. Default value is 1.0.
17794 @item luma_threshold, lt
17795 Set the luma threshold used as a coefficient to determine
17796 whether a pixel should be blurred or not. The option value must be an
17797 integer in the range [-30,30]. A value of 0 will filter all the image,
17798 a value included in [0,30] will filter flat areas and a value included
17799 in [-30,0] will filter edges. Default value is 0.
17801 @item chroma_radius, cr
17802 Set the chroma radius. The option value must be a float number in
17803 the range [0.1,5.0] that specifies the variance of the gaussian filter
17804 used to blur the image (slower if larger). Default value is @option{luma_radius}.
17806 @item chroma_strength, cs
17807 Set the chroma strength. The option value must be a float number
17808 in the range [-1.0,1.0] that configures the blurring. A value included
17809 in [0.0,1.0] will blur the image whereas a value included in
17810 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
17812 @item chroma_threshold, ct
17813 Set the chroma threshold used as a coefficient to determine
17814 whether a pixel should be blurred or not. The option value must be an
17815 integer in the range [-30,30]. A value of 0 will filter all the image,
17816 a value included in [0,30] will filter flat areas and a value included
17817 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
17820 If a chroma option is not explicitly set, the corresponding luma value
17824 Apply sobel operator to input video stream.
17826 The filter accepts the following option:
17830 Set which planes will be processed, unprocessed planes will be copied.
17831 By default value 0xf, all planes will be processed.
17834 Set value which will be multiplied with filtered result.
17837 Set value which will be added to filtered result.
17843 Apply a simple postprocessing filter that compresses and decompresses the image
17844 at several (or - in the case of @option{quality} level @code{6} - all) shifts
17845 and average the results.
17847 The filter accepts the following options:
17851 Set quality. This option defines the number of levels for averaging. It accepts
17852 an integer in the range 0-6. If set to @code{0}, the filter will have no
17853 effect. A value of @code{6} means the higher quality. For each increment of
17854 that value the speed drops by a factor of approximately 2. Default value is
17858 Force a constant quantization parameter. If not set, the filter will use the QP
17859 from the video stream (if available).
17862 Set thresholding mode. Available modes are:
17866 Set hard thresholding (default).
17868 Set soft thresholding (better de-ringing effect, but likely blurrier).
17871 @item use_bframe_qp
17872 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
17873 option may cause flicker since the B-Frames have often larger QP. Default is
17874 @code{0} (not enabled).
17877 @subsection Commands
17879 This filter supports the following commands:
17881 @item quality, level
17882 Set quality level. The value @code{max} can be used to set the maximum level,
17883 currently @code{6}.
17889 Scale the input by applying one of the super-resolution methods based on
17890 convolutional neural networks. Supported models:
17894 Super-Resolution Convolutional Neural Network model (SRCNN).
17895 See @url{https://arxiv.org/abs/1501.00092}.
17898 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
17899 See @url{https://arxiv.org/abs/1609.05158}.
17902 Training scripts as well as scripts for model file (.pb) saving can be found at
17903 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
17904 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
17906 Native model files (.model) can be generated from TensorFlow model
17907 files (.pb) by using tools/python/convert.py
17909 The filter accepts the following options:
17913 Specify which DNN backend to use for model loading and execution. This option accepts
17914 the following values:
17918 Native implementation of DNN loading and execution.
17921 TensorFlow backend. To enable this backend you
17922 need to install the TensorFlow for C library (see
17923 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
17924 @code{--enable-libtensorflow}
17927 Default value is @samp{native}.
17930 Set path to model file specifying network architecture and its parameters.
17931 Note that different backends use different file formats. TensorFlow backend
17932 can load files for both formats, while native backend can load files for only
17936 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
17937 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
17938 input upscaled using bicubic upscaling with proper scale factor.
17941 This feature can also be finished with @ref{dnn_processing} filter.
17945 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
17947 This filter takes in input two input videos, the first input is
17948 considered the "main" source and is passed unchanged to the
17949 output. The second input is used as a "reference" video for computing
17952 Both video inputs must have the same resolution and pixel format for
17953 this filter to work correctly. Also it assumes that both inputs
17954 have the same number of frames, which are compared one by one.
17956 The filter stores the calculated SSIM of each frame.
17958 The description of the accepted parameters follows.
17961 @item stats_file, f
17962 If specified the filter will use the named file to save the SSIM of
17963 each individual frame. When filename equals "-" the data is sent to
17967 The file printed if @var{stats_file} is selected, contains a sequence of
17968 key/value pairs of the form @var{key}:@var{value} for each compared
17971 A description of each shown parameter follows:
17975 sequential number of the input frame, starting from 1
17977 @item Y, U, V, R, G, B
17978 SSIM of the compared frames for the component specified by the suffix.
17981 SSIM of the compared frames for the whole frame.
17984 Same as above but in dB representation.
17987 This filter also supports the @ref{framesync} options.
17989 @subsection Examples
17994 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
17995 [main][ref] ssim="stats_file=stats.log" [out]
17998 On this example the input file being processed is compared with the
17999 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18000 is stored in @file{stats.log}.
18003 Another example with both psnr and ssim at same time:
18005 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18009 Another example with different containers:
18011 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 -
18017 Convert between different stereoscopic image formats.
18019 The filters accept the following options:
18023 Set stereoscopic image format of input.
18025 Available values for input image formats are:
18028 side by side parallel (left eye left, right eye right)
18031 side by side crosseye (right eye left, left eye right)
18034 side by side parallel with half width resolution
18035 (left eye left, right eye right)
18038 side by side crosseye with half width resolution
18039 (right eye left, left eye right)
18043 above-below (left eye above, right eye below)
18047 above-below (right eye above, left eye below)
18051 above-below with half height resolution
18052 (left eye above, right eye below)
18056 above-below with half height resolution
18057 (right eye above, left eye below)
18060 alternating frames (left eye first, right eye second)
18063 alternating frames (right eye first, left eye second)
18066 interleaved rows (left eye has top row, right eye starts on next row)
18069 interleaved rows (right eye has top row, left eye starts on next row)
18072 interleaved columns, left eye first
18075 interleaved columns, right eye first
18077 Default value is @samp{sbsl}.
18081 Set stereoscopic image format of output.
18085 side by side parallel (left eye left, right eye right)
18088 side by side crosseye (right eye left, left eye right)
18091 side by side parallel with half width resolution
18092 (left eye left, right eye right)
18095 side by side crosseye with half width resolution
18096 (right eye left, left eye right)
18100 above-below (left eye above, right eye below)
18104 above-below (right eye above, left eye below)
18108 above-below with half height resolution
18109 (left eye above, right eye below)
18113 above-below with half height resolution
18114 (right eye above, left eye below)
18117 alternating frames (left eye first, right eye second)
18120 alternating frames (right eye first, left eye second)
18123 interleaved rows (left eye has top row, right eye starts on next row)
18126 interleaved rows (right eye has top row, left eye starts on next row)
18129 anaglyph red/blue gray
18130 (red filter on left eye, blue filter on right eye)
18133 anaglyph red/green gray
18134 (red filter on left eye, green filter on right eye)
18137 anaglyph red/cyan gray
18138 (red filter on left eye, cyan filter on right eye)
18141 anaglyph red/cyan half colored
18142 (red filter on left eye, cyan filter on right eye)
18145 anaglyph red/cyan color
18146 (red filter on left eye, cyan filter on right eye)
18149 anaglyph red/cyan color optimized with the least squares projection of dubois
18150 (red filter on left eye, cyan filter on right eye)
18153 anaglyph green/magenta gray
18154 (green filter on left eye, magenta filter on right eye)
18157 anaglyph green/magenta half colored
18158 (green filter on left eye, magenta filter on right eye)
18161 anaglyph green/magenta colored
18162 (green filter on left eye, magenta filter on right eye)
18165 anaglyph green/magenta color optimized with the least squares projection of dubois
18166 (green filter on left eye, magenta filter on right eye)
18169 anaglyph yellow/blue gray
18170 (yellow filter on left eye, blue filter on right eye)
18173 anaglyph yellow/blue half colored
18174 (yellow filter on left eye, blue filter on right eye)
18177 anaglyph yellow/blue colored
18178 (yellow filter on left eye, blue filter on right eye)
18181 anaglyph yellow/blue color optimized with the least squares projection of dubois
18182 (yellow filter on left eye, blue filter on right eye)
18185 mono output (left eye only)
18188 mono output (right eye only)
18191 checkerboard, left eye first
18194 checkerboard, right eye first
18197 interleaved columns, left eye first
18200 interleaved columns, right eye first
18206 Default value is @samp{arcd}.
18209 @subsection Examples
18213 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18219 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18225 @section streamselect, astreamselect
18226 Select video or audio streams.
18228 The filter accepts the following options:
18232 Set number of inputs. Default is 2.
18235 Set input indexes to remap to outputs.
18238 @subsection Commands
18240 The @code{streamselect} and @code{astreamselect} filter supports the following
18245 Set input indexes to remap to outputs.
18248 @subsection Examples
18252 Select first 5 seconds 1st stream and rest of time 2nd stream:
18254 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18258 Same as above, but for audio:
18260 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18267 Draw subtitles on top of input video using the libass library.
18269 To enable compilation of this filter you need to configure FFmpeg with
18270 @code{--enable-libass}. This filter also requires a build with libavcodec and
18271 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18272 Alpha) subtitles format.
18274 The filter accepts the following options:
18278 Set the filename of the subtitle file to read. It must be specified.
18280 @item original_size
18281 Specify the size of the original video, the video for which the ASS file
18282 was composed. For the syntax of this option, check the
18283 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18284 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18285 correctly scale the fonts if the aspect ratio has been changed.
18288 Set a directory path containing fonts that can be used by the filter.
18289 These fonts will be used in addition to whatever the font provider uses.
18292 Process alpha channel, by default alpha channel is untouched.
18295 Set subtitles input character encoding. @code{subtitles} filter only. Only
18296 useful if not UTF-8.
18298 @item stream_index, si
18299 Set subtitles stream index. @code{subtitles} filter only.
18302 Override default style or script info parameters of the subtitles. It accepts a
18303 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18306 If the first key is not specified, it is assumed that the first value
18307 specifies the @option{filename}.
18309 For example, to render the file @file{sub.srt} on top of the input
18310 video, use the command:
18315 which is equivalent to:
18317 subtitles=filename=sub.srt
18320 To render the default subtitles stream from file @file{video.mkv}, use:
18322 subtitles=video.mkv
18325 To render the second subtitles stream from that file, use:
18327 subtitles=video.mkv:si=1
18330 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18331 @code{DejaVu Serif}, use:
18333 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18336 @section super2xsai
18338 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18339 Interpolate) pixel art scaling algorithm.
18341 Useful for enlarging pixel art images without reducing sharpness.
18345 Swap two rectangular objects in video.
18347 This filter accepts the following options:
18357 Set 1st rect x coordinate.
18360 Set 1st rect y coordinate.
18363 Set 2nd rect x coordinate.
18366 Set 2nd rect y coordinate.
18368 All expressions are evaluated once for each frame.
18371 The all options are expressions containing the following constants:
18376 The input width and height.
18379 same as @var{w} / @var{h}
18382 input sample aspect ratio
18385 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18388 The number of the input frame, starting from 0.
18391 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18394 the position in the file of the input frame, NAN if unknown
18401 Blend successive video frames.
18407 Apply telecine process to the video.
18409 This filter accepts the following options:
18418 The default value is @code{top}.
18422 A string of numbers representing the pulldown pattern you wish to apply.
18423 The default value is @code{23}.
18427 Some typical patterns:
18432 24p: 2332 (preferred)
18439 24p: 222222222223 ("Euro pulldown")
18444 @section thistogram
18446 Compute and draw a color distribution histogram for the input video across time.
18448 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18449 at certain time, this filter shows also past histograms of number of frames defined
18450 by @code{width} option.
18452 The computed histogram is a representation of the color component
18453 distribution in an image.
18455 The filter accepts the following options:
18459 Set width of single color component output. Default value is @code{0}.
18460 Value of @code{0} means width will be picked from input video.
18461 This also set number of passed histograms to keep.
18462 Allowed range is [0, 8192].
18464 @item display_mode, d
18466 It accepts the following values:
18469 Per color component graphs are placed below each other.
18472 Per color component graphs are placed side by side.
18475 Presents information identical to that in the @code{parade}, except
18476 that the graphs representing color components are superimposed directly
18479 Default is @code{stack}.
18481 @item levels_mode, m
18482 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18483 Default is @code{linear}.
18485 @item components, c
18486 Set what color components to display.
18487 Default is @code{7}.
18490 Set background opacity. Default is @code{0.9}.
18493 Show envelope. Default is disabled.
18496 Set envelope color. Default is @code{gold}.
18501 Available values for slide is:
18504 Draw new frame when right border is reached.
18507 Replace old columns with new ones.
18510 Scroll from right to left.
18513 Scroll from left to right.
18516 Draw single picture.
18519 Default is @code{replace}.
18524 Apply threshold effect to video stream.
18526 This filter needs four video streams to perform thresholding.
18527 First stream is stream we are filtering.
18528 Second stream is holding threshold values, third stream is holding min values,
18529 and last, fourth stream is holding max values.
18531 The filter accepts the following option:
18535 Set which planes will be processed, unprocessed planes will be copied.
18536 By default value 0xf, all planes will be processed.
18539 For example if first stream pixel's component value is less then threshold value
18540 of pixel component from 2nd threshold stream, third stream value will picked,
18541 otherwise fourth stream pixel component value will be picked.
18543 Using color source filter one can perform various types of thresholding:
18545 @subsection Examples
18549 Binary threshold, using gray color as threshold:
18551 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18555 Inverted binary threshold, using gray color as threshold:
18557 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18561 Truncate binary threshold, using gray color as threshold:
18563 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18567 Threshold to zero, using gray color as threshold:
18569 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18573 Inverted threshold to zero, using gray color as threshold:
18575 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18580 Select the most representative frame in a given sequence of consecutive frames.
18582 The filter accepts the following options:
18586 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18587 will pick one of them, and then handle the next batch of @var{n} frames until
18588 the end. Default is @code{100}.
18591 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18592 value will result in a higher memory usage, so a high value is not recommended.
18594 @subsection Examples
18598 Extract one picture each 50 frames:
18604 Complete example of a thumbnail creation with @command{ffmpeg}:
18606 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18613 Tile several successive frames together.
18615 The @ref{untile} filter can do the reverse.
18617 The filter accepts the following options:
18622 Set the grid size (i.e. the number of lines and columns). For the syntax of
18623 this option, check the
18624 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18627 Set the maximum number of frames to render in the given area. It must be less
18628 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18629 the area will be used.
18632 Set the outer border margin in pixels.
18635 Set the inner border thickness (i.e. the number of pixels between frames). For
18636 more advanced padding options (such as having different values for the edges),
18637 refer to the pad video filter.
18640 Specify the color of the unused area. For the syntax of this option, check the
18641 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18642 The default value of @var{color} is "black".
18645 Set the number of frames to overlap when tiling several successive frames together.
18646 The value must be between @code{0} and @var{nb_frames - 1}.
18649 Set the number of frames to initially be empty before displaying first output frame.
18650 This controls how soon will one get first output frame.
18651 The value must be between @code{0} and @var{nb_frames - 1}.
18654 @subsection Examples
18658 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18660 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18662 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18663 duplicating each output frame to accommodate the originally detected frame
18667 Display @code{5} pictures in an area of @code{3x2} frames,
18668 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18669 mixed flat and named options:
18671 tile=3x2:nb_frames=5:padding=7:margin=2
18675 @section tinterlace
18677 Perform various types of temporal field interlacing.
18679 Frames are counted starting from 1, so the first input frame is
18682 The filter accepts the following options:
18687 Specify the mode of the interlacing. This option can also be specified
18688 as a value alone. See below for a list of values for this option.
18690 Available values are:
18694 Move odd frames into the upper field, even into the lower field,
18695 generating a double height frame at half frame rate.
18699 Frame 1 Frame 2 Frame 3 Frame 4
18701 11111 22222 33333 44444
18702 11111 22222 33333 44444
18703 11111 22222 33333 44444
18704 11111 22222 33333 44444
18718 Only output odd frames, even frames are dropped, generating a frame with
18719 unchanged height at half frame rate.
18724 Frame 1 Frame 2 Frame 3 Frame 4
18726 11111 22222 33333 44444
18727 11111 22222 33333 44444
18728 11111 22222 33333 44444
18729 11111 22222 33333 44444
18739 Only output even frames, odd frames are dropped, generating a frame with
18740 unchanged height at half frame rate.
18745 Frame 1 Frame 2 Frame 3 Frame 4
18747 11111 22222 33333 44444
18748 11111 22222 33333 44444
18749 11111 22222 33333 44444
18750 11111 22222 33333 44444
18760 Expand each frame to full height, but pad alternate lines with black,
18761 generating a frame with double height at the same input frame rate.
18766 Frame 1 Frame 2 Frame 3 Frame 4
18768 11111 22222 33333 44444
18769 11111 22222 33333 44444
18770 11111 22222 33333 44444
18771 11111 22222 33333 44444
18774 11111 ..... 33333 .....
18775 ..... 22222 ..... 44444
18776 11111 ..... 33333 .....
18777 ..... 22222 ..... 44444
18778 11111 ..... 33333 .....
18779 ..... 22222 ..... 44444
18780 11111 ..... 33333 .....
18781 ..... 22222 ..... 44444
18785 @item interleave_top, 4
18786 Interleave the upper field from odd frames with the lower field from
18787 even frames, generating a frame with unchanged height at half frame rate.
18792 Frame 1 Frame 2 Frame 3 Frame 4
18794 11111<- 22222 33333<- 44444
18795 11111 22222<- 33333 44444<-
18796 11111<- 22222 33333<- 44444
18797 11111 22222<- 33333 44444<-
18807 @item interleave_bottom, 5
18808 Interleave the lower field from odd frames with the upper field from
18809 even frames, generating a frame with unchanged height at half frame rate.
18814 Frame 1 Frame 2 Frame 3 Frame 4
18816 11111 22222<- 33333 44444<-
18817 11111<- 22222 33333<- 44444
18818 11111 22222<- 33333 44444<-
18819 11111<- 22222 33333<- 44444
18829 @item interlacex2, 6
18830 Double frame rate with unchanged height. Frames are inserted each
18831 containing the second temporal field from the previous input frame and
18832 the first temporal field from the next input frame. This mode relies on
18833 the top_field_first flag. Useful for interlaced video displays with no
18834 field synchronisation.
18839 Frame 1 Frame 2 Frame 3 Frame 4
18841 11111 22222 33333 44444
18842 11111 22222 33333 44444
18843 11111 22222 33333 44444
18844 11111 22222 33333 44444
18847 11111 22222 22222 33333 33333 44444 44444
18848 11111 11111 22222 22222 33333 33333 44444
18849 11111 22222 22222 33333 33333 44444 44444
18850 11111 11111 22222 22222 33333 33333 44444
18855 Move odd frames into the upper field, even into the lower field,
18856 generating a double height frame at same frame rate.
18861 Frame 1 Frame 2 Frame 3 Frame 4
18863 11111 22222 33333 44444
18864 11111 22222 33333 44444
18865 11111 22222 33333 44444
18866 11111 22222 33333 44444
18869 11111 33333 33333 55555
18870 22222 22222 44444 44444
18871 11111 33333 33333 55555
18872 22222 22222 44444 44444
18873 11111 33333 33333 55555
18874 22222 22222 44444 44444
18875 11111 33333 33333 55555
18876 22222 22222 44444 44444
18881 Numeric values are deprecated but are accepted for backward
18882 compatibility reasons.
18884 Default mode is @code{merge}.
18887 Specify flags influencing the filter process.
18889 Available value for @var{flags} is:
18892 @item low_pass_filter, vlpf
18893 Enable linear vertical low-pass filtering in the filter.
18894 Vertical low-pass filtering is required when creating an interlaced
18895 destination from a progressive source which contains high-frequency
18896 vertical detail. Filtering will reduce interlace 'twitter' and Moire
18899 @item complex_filter, cvlpf
18900 Enable complex vertical low-pass filtering.
18901 This will slightly less reduce interlace 'twitter' and Moire
18902 patterning but better retain detail and subjective sharpness impression.
18905 Bypass already interlaced frames, only adjust the frame rate.
18908 Vertical low-pass filtering and bypassing already interlaced frames can only be
18909 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
18914 Pick median pixels from several successive input video frames.
18916 The filter accepts the following options:
18920 Set radius of median filter.
18921 Default is 1. Allowed range is from 1 to 127.
18924 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
18927 Set median percentile. Default value is @code{0.5}.
18928 Default value of @code{0.5} will pick always median values, while @code{0} will pick
18929 minimum values, and @code{1} maximum values.
18934 Mix successive video frames.
18936 A description of the accepted options follows.
18940 The number of successive frames to mix. If unspecified, it defaults to 3.
18943 Specify weight of each input video frame.
18944 Each weight is separated by space. If number of weights is smaller than
18945 number of @var{frames} last specified weight will be used for all remaining
18949 Specify scale, if it is set it will be multiplied with sum
18950 of each weight multiplied with pixel values to give final destination
18951 pixel value. By default @var{scale} is auto scaled to sum of weights.
18954 @subsection Examples
18958 Average 7 successive frames:
18960 tmix=frames=7:weights="1 1 1 1 1 1 1"
18964 Apply simple temporal convolution:
18966 tmix=frames=3:weights="-1 3 -1"
18970 Similar as above but only showing temporal differences:
18972 tmix=frames=3:weights="-1 2 -1":scale=1
18978 Tone map colors from different dynamic ranges.
18980 This filter expects data in single precision floating point, as it needs to
18981 operate on (and can output) out-of-range values. Another filter, such as
18982 @ref{zscale}, is needed to convert the resulting frame to a usable format.
18984 The tonemapping algorithms implemented only work on linear light, so input
18985 data should be linearized beforehand (and possibly correctly tagged).
18988 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
18991 @subsection Options
18992 The filter accepts the following options.
18996 Set the tone map algorithm to use.
18998 Possible values are:
19001 Do not apply any tone map, only desaturate overbright pixels.
19004 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19005 in-range values, while distorting out-of-range values.
19008 Stretch the entire reference gamut to a linear multiple of the display.
19011 Fit a logarithmic transfer between the tone curves.
19014 Preserve overall image brightness with a simple curve, using nonlinear
19015 contrast, which results in flattening details and degrading color accuracy.
19018 Preserve both dark and bright details better than @var{reinhard}, at the cost
19019 of slightly darkening everything. Use it when detail preservation is more
19020 important than color and brightness accuracy.
19023 Smoothly map out-of-range values, while retaining contrast and colors for
19024 in-range material as much as possible. Use it when color accuracy is more
19025 important than detail preservation.
19031 Tune the tone mapping algorithm.
19033 This affects the following algorithms:
19039 Specifies the scale factor to use while stretching.
19043 Specifies the exponent of the function.
19047 Specify an extra linear coefficient to multiply into the signal before clipping.
19051 Specify the local contrast coefficient at the display peak.
19052 Default to 0.5, which means that in-gamut values will be about half as bright
19059 Specify the transition point from linear to mobius transform. Every value
19060 below this point is guaranteed to be mapped 1:1. The higher the value, the
19061 more accurate the result will be, at the cost of losing bright details.
19062 Default to 0.3, which due to the steep initial slope still preserves in-range
19063 colors fairly accurately.
19067 Apply desaturation for highlights that exceed this level of brightness. The
19068 higher the parameter, the more color information will be preserved. This
19069 setting helps prevent unnaturally blown-out colors for super-highlights, by
19070 (smoothly) turning into white instead. This makes images feel more natural,
19071 at the cost of reducing information about out-of-range colors.
19073 The default of 2.0 is somewhat conservative and will mostly just apply to
19074 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19076 This option works only if the input frame has a supported color tag.
19079 Override signal/nominal/reference peak with this value. Useful when the
19080 embedded peak information in display metadata is not reliable or when tone
19081 mapping from a lower range to a higher range.
19086 Temporarily pad video frames.
19088 The filter accepts the following options:
19092 Specify number of delay frames before input video stream. Default is 0.
19095 Specify number of padding frames after input video stream.
19096 Set to -1 to pad indefinitely. Default is 0.
19099 Set kind of frames added to beginning of stream.
19100 Can be either @var{add} or @var{clone}.
19101 With @var{add} frames of solid-color are added.
19102 With @var{clone} frames are clones of first frame.
19103 Default is @var{add}.
19106 Set kind of frames added to end of stream.
19107 Can be either @var{add} or @var{clone}.
19108 With @var{add} frames of solid-color are added.
19109 With @var{clone} frames are clones of last frame.
19110 Default is @var{add}.
19112 @item start_duration, stop_duration
19113 Specify the duration of the start/stop delay. See
19114 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19115 for the accepted syntax.
19116 These options override @var{start} and @var{stop}. Default is 0.
19119 Specify the color of the padded area. For the syntax of this option,
19120 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19121 manual,ffmpeg-utils}.
19123 The default value of @var{color} is "black".
19129 Transpose rows with columns in the input video and optionally flip it.
19131 It accepts the following parameters:
19136 Specify the transposition direction.
19138 Can assume the following values:
19140 @item 0, 4, cclock_flip
19141 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19149 Rotate by 90 degrees clockwise, that is:
19157 Rotate by 90 degrees counterclockwise, that is:
19164 @item 3, 7, clock_flip
19165 Rotate by 90 degrees clockwise and vertically flip, that is:
19173 For values between 4-7, the transposition is only done if the input
19174 video geometry is portrait and not landscape. These values are
19175 deprecated, the @code{passthrough} option should be used instead.
19177 Numerical values are deprecated, and should be dropped in favor of
19178 symbolic constants.
19181 Do not apply the transposition if the input geometry matches the one
19182 specified by the specified value. It accepts the following values:
19185 Always apply transposition.
19187 Preserve portrait geometry (when @var{height} >= @var{width}).
19189 Preserve landscape geometry (when @var{width} >= @var{height}).
19192 Default value is @code{none}.
19195 For example to rotate by 90 degrees clockwise and preserve portrait
19198 transpose=dir=1:passthrough=portrait
19201 The command above can also be specified as:
19203 transpose=1:portrait
19206 @section transpose_npp
19208 Transpose rows with columns in the input video and optionally flip it.
19209 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19211 It accepts the following parameters:
19216 Specify the transposition direction.
19218 Can assume the following values:
19221 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19224 Rotate by 90 degrees clockwise.
19227 Rotate by 90 degrees counterclockwise.
19230 Rotate by 90 degrees clockwise and vertically flip.
19234 Do not apply the transposition if the input geometry matches the one
19235 specified by the specified value. It accepts the following values:
19238 Always apply transposition. (default)
19240 Preserve portrait geometry (when @var{height} >= @var{width}).
19242 Preserve landscape geometry (when @var{width} >= @var{height}).
19248 Trim the input so that the output contains one continuous subpart of the input.
19250 It accepts the following parameters:
19253 Specify the time of the start of the kept section, i.e. the frame with the
19254 timestamp @var{start} will be the first frame in the output.
19257 Specify the time of the first frame that will be dropped, i.e. the frame
19258 immediately preceding the one with the timestamp @var{end} will be the last
19259 frame in the output.
19262 This is the same as @var{start}, except this option sets the start timestamp
19263 in timebase units instead of seconds.
19266 This is the same as @var{end}, except this option sets the end timestamp
19267 in timebase units instead of seconds.
19270 The maximum duration of the output in seconds.
19273 The number of the first frame that should be passed to the output.
19276 The number of the first frame that should be dropped.
19279 @option{start}, @option{end}, and @option{duration} are expressed as time
19280 duration specifications; see
19281 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19282 for the accepted syntax.
19284 Note that the first two sets of the start/end options and the @option{duration}
19285 option look at the frame timestamp, while the _frame variants simply count the
19286 frames that pass through the filter. Also note that this filter does not modify
19287 the timestamps. If you wish for the output timestamps to start at zero, insert a
19288 setpts filter after the trim filter.
19290 If multiple start or end options are set, this filter tries to be greedy and
19291 keep all the frames that match at least one of the specified constraints. To keep
19292 only the part that matches all the constraints at once, chain multiple trim
19295 The defaults are such that all the input is kept. So it is possible to set e.g.
19296 just the end values to keep everything before the specified time.
19301 Drop everything except the second minute of input:
19303 ffmpeg -i INPUT -vf trim=60:120
19307 Keep only the first second:
19309 ffmpeg -i INPUT -vf trim=duration=1
19314 @section unpremultiply
19315 Apply alpha unpremultiply effect to input video stream using first plane
19316 of second stream as alpha.
19318 Both streams must have same dimensions and same pixel format.
19320 The filter accepts the following option:
19324 Set which planes will be processed, unprocessed planes will be copied.
19325 By default value 0xf, all planes will be processed.
19327 If the format has 1 or 2 components, then luma is bit 0.
19328 If the format has 3 or 4 components:
19329 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19330 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19331 If present, the alpha channel is always the last bit.
19334 Do not require 2nd input for processing, instead use alpha plane from input stream.
19340 Sharpen or blur the input video.
19342 It accepts the following parameters:
19345 @item luma_msize_x, lx
19346 Set the luma matrix horizontal size. It must be an odd integer between
19347 3 and 23. The default value is 5.
19349 @item luma_msize_y, ly
19350 Set the luma matrix vertical size. It must be an odd integer between 3
19351 and 23. The default value is 5.
19353 @item luma_amount, la
19354 Set the luma effect strength. It must be a floating point number, reasonable
19355 values lay between -1.5 and 1.5.
19357 Negative values will blur the input video, while positive values will
19358 sharpen it, a value of zero will disable the effect.
19360 Default value is 1.0.
19362 @item chroma_msize_x, cx
19363 Set the chroma matrix horizontal size. It must be an odd integer
19364 between 3 and 23. The default value is 5.
19366 @item chroma_msize_y, cy
19367 Set the chroma matrix vertical size. It must be an odd integer
19368 between 3 and 23. The default value is 5.
19370 @item chroma_amount, ca
19371 Set the chroma effect strength. It must be a floating point number, reasonable
19372 values lay between -1.5 and 1.5.
19374 Negative values will blur the input video, while positive values will
19375 sharpen it, a value of zero will disable the effect.
19377 Default value is 0.0.
19381 All parameters are optional and default to the equivalent of the
19382 string '5:5:1.0:5:5:0.0'.
19384 @subsection Examples
19388 Apply strong luma sharpen effect:
19390 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19394 Apply a strong blur of both luma and chroma parameters:
19396 unsharp=7:7:-2:7:7:-2
19403 Decompose a video made of tiled images into the individual images.
19405 The frame rate of the output video is the frame rate of the input video
19406 multiplied by the number of tiles.
19408 This filter does the reverse of @ref{tile}.
19410 The filter accepts the following options:
19415 Set the grid size (i.e. the number of lines and columns). For the syntax of
19416 this option, check the
19417 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19420 @subsection Examples
19424 Produce a 1-second video from a still image file made of 25 frames stacked
19425 vertically, like an analogic film reel:
19427 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19433 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19434 the image at several (or - in the case of @option{quality} level @code{8} - all)
19435 shifts and average the results.
19437 The way this differs from the behavior of spp is that uspp actually encodes &
19438 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19439 DCT similar to MJPEG.
19441 The filter accepts the following options:
19445 Set quality. This option defines the number of levels for averaging. It accepts
19446 an integer in the range 0-8. If set to @code{0}, the filter will have no
19447 effect. A value of @code{8} means the higher quality. For each increment of
19448 that value the speed drops by a factor of approximately 2. Default value is
19452 Force a constant quantization parameter. If not set, the filter will use the QP
19453 from the video stream (if available).
19458 Convert 360 videos between various formats.
19460 The filter accepts the following options:
19466 Set format of the input/output video.
19474 Equirectangular projection.
19479 Cubemap with 3x2/6x1/1x6 layout.
19481 Format specific options:
19486 Set padding proportion for the input/output cubemap. Values in decimals.
19493 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)
19496 Default value is @b{@samp{0}}.
19497 Maximum value is @b{@samp{0.1}}.
19501 Set fixed padding for the input/output cubemap. Values in pixels.
19503 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19507 Set order of faces for the input/output cubemap. Choose one direction for each position.
19509 Designation of directions:
19525 Default value is @b{@samp{rludfb}}.
19529 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19531 Designation of angles:
19534 0 degrees clockwise
19536 90 degrees clockwise
19538 180 degrees clockwise
19540 270 degrees clockwise
19543 Default value is @b{@samp{000000}}.
19547 Equi-Angular Cubemap.
19554 Format specific options:
19559 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19561 If diagonal field of view is set it overrides horizontal and vertical field of view.
19566 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19568 If diagonal field of view is set it overrides horizontal and vertical field of view.
19574 Format specific options:
19579 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19581 If diagonal field of view is set it overrides horizontal and vertical field of view.
19586 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19588 If diagonal field of view is set it overrides horizontal and vertical field of view.
19594 Facebook's 360 formats.
19597 Stereographic format.
19599 Format specific options:
19604 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19606 If diagonal field of view is set it overrides horizontal and vertical field of view.
19611 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19613 If diagonal field of view is set it overrides horizontal and vertical field of view.
19620 Ball format, gives significant distortion toward the back.
19623 Hammer-Aitoff map projection format.
19626 Sinusoidal map projection format.
19629 Fisheye projection.
19631 Format specific options:
19636 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19638 If diagonal field of view is set it overrides horizontal and vertical field of view.
19643 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19645 If diagonal field of view is set it overrides horizontal and vertical field of view.
19649 Pannini projection.
19651 Format specific options:
19654 Set output pannini parameter.
19657 Set input pannini parameter.
19661 Cylindrical projection.
19663 Format specific options:
19668 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19670 If diagonal field of view is set it overrides horizontal and vertical field of view.
19675 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19677 If diagonal field of view is set it overrides horizontal and vertical field of view.
19681 Perspective projection. @i{(output only)}
19683 Format specific options:
19686 Set perspective parameter.
19690 Tetrahedron projection.
19693 Truncated square pyramid projection.
19697 Half equirectangular projection.
19702 Format specific options:
19707 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19709 If diagonal field of view is set it overrides horizontal and vertical field of view.
19714 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19716 If diagonal field of view is set it overrides horizontal and vertical field of view.
19720 Orthographic format.
19722 Format specific options:
19727 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19729 If diagonal field of view is set it overrides horizontal and vertical field of view.
19734 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19736 If diagonal field of view is set it overrides horizontal and vertical field of view.
19740 Octahedron projection.
19744 Set interpolation method.@*
19745 @i{Note: more complex interpolation methods require much more memory to run.}
19755 Bilinear interpolation.
19757 Lagrange9 interpolation.
19760 Bicubic interpolation.
19763 Lanczos interpolation.
19766 Spline16 interpolation.
19769 Gaussian interpolation.
19771 Mitchell interpolation.
19774 Default value is @b{@samp{line}}.
19778 Set the output video resolution.
19780 Default resolution depends on formats.
19784 Set the input/output stereo format.
19795 Default value is @b{@samp{2d}} for input and output format.
19800 Set rotation for the output video. Values in degrees.
19803 Set rotation order for the output video. Choose one item for each position.
19814 Default value is @b{@samp{ypr}}.
19819 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
19823 Set if input video is flipped horizontally/vertically. Boolean values.
19826 Set if input video is transposed. Boolean value, by default disabled.
19829 Set if output video needs to be transposed. Boolean value, by default disabled.
19832 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
19835 @subsection Examples
19839 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
19841 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
19844 Extract back view of Equi-Angular Cubemap:
19846 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
19849 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
19851 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
19855 @subsection Commands
19857 This filter supports subset of above options as @ref{commands}.
19859 @section vaguedenoiser
19861 Apply a wavelet based denoiser.
19863 It transforms each frame from the video input into the wavelet domain,
19864 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
19865 the obtained coefficients. It does an inverse wavelet transform after.
19866 Due to wavelet properties, it should give a nice smoothed result, and
19867 reduced noise, without blurring picture features.
19869 This filter accepts the following options:
19873 The filtering strength. The higher, the more filtered the video will be.
19874 Hard thresholding can use a higher threshold than soft thresholding
19875 before the video looks overfiltered. Default value is 2.
19878 The filtering method the filter will use.
19880 It accepts the following values:
19883 All values under the threshold will be zeroed.
19886 All values under the threshold will be zeroed. All values above will be
19887 reduced by the threshold.
19890 Scales or nullifies coefficients - intermediary between (more) soft and
19891 (less) hard thresholding.
19894 Default is garrote.
19897 Number of times, the wavelet will decompose the picture. Picture can't
19898 be decomposed beyond a particular point (typically, 8 for a 640x480
19899 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
19902 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
19905 A list of the planes to process. By default all planes are processed.
19908 The threshold type the filter will use.
19910 It accepts the following values:
19913 Threshold used is same for all decompositions.
19916 Threshold used depends also on each decomposition coefficients.
19919 Default is universal.
19922 @section vectorscope
19924 Display 2 color component values in the two dimensional graph (which is called
19927 This filter accepts the following options:
19931 Set vectorscope mode.
19933 It accepts the following values:
19937 Gray values are displayed on graph, higher brightness means more pixels have
19938 same component color value on location in graph. This is the default mode.
19941 Gray values are displayed on graph. Surrounding pixels values which are not
19942 present in video frame are drawn in gradient of 2 color components which are
19943 set by option @code{x} and @code{y}. The 3rd color component is static.
19946 Actual color components values present in video frame are displayed on graph.
19949 Similar as color2 but higher frequency of same values @code{x} and @code{y}
19950 on graph increases value of another color component, which is luminance by
19951 default values of @code{x} and @code{y}.
19954 Actual colors present in video frame are displayed on graph. If two different
19955 colors map to same position on graph then color with higher value of component
19956 not present in graph is picked.
19959 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
19960 component picked from radial gradient.
19964 Set which color component will be represented on X-axis. Default is @code{1}.
19967 Set which color component will be represented on Y-axis. Default is @code{2}.
19970 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
19971 of color component which represents frequency of (X, Y) location in graph.
19976 No envelope, this is default.
19979 Instant envelope, even darkest single pixel will be clearly highlighted.
19982 Hold maximum and minimum values presented in graph over time. This way you
19983 can still spot out of range values without constantly looking at vectorscope.
19986 Peak and instant envelope combined together.
19990 Set what kind of graticule to draw.
19999 Set graticule opacity.
20002 Set graticule flags.
20006 Draw graticule for white point.
20009 Draw graticule for black point.
20012 Draw color points short names.
20016 Set background opacity.
20018 @item lthreshold, l
20019 Set low threshold for color component not represented on X or Y axis.
20020 Values lower than this value will be ignored. Default is 0.
20021 Note this value is multiplied with actual max possible value one pixel component
20022 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20025 @item hthreshold, h
20026 Set high threshold for color component not represented on X or Y axis.
20027 Values higher than this value will be ignored. Default is 1.
20028 Note this value is multiplied with actual max possible value one pixel component
20029 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20030 is 0.9 * 255 = 230.
20032 @item colorspace, c
20033 Set what kind of colorspace to use when drawing graticule.
20043 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20044 This means no tint, and output will remain gray.
20047 @anchor{vidstabdetect}
20048 @section vidstabdetect
20050 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20051 @ref{vidstabtransform} for pass 2.
20053 This filter generates a file with relative translation and rotation
20054 transform information about subsequent frames, which is then used by
20055 the @ref{vidstabtransform} filter.
20057 To enable compilation of this filter you need to configure FFmpeg with
20058 @code{--enable-libvidstab}.
20060 This filter accepts the following options:
20064 Set the path to the file used to write the transforms information.
20065 Default value is @file{transforms.trf}.
20068 Set how shaky the video is and how quick the camera is. It accepts an
20069 integer in the range 1-10, a value of 1 means little shakiness, a
20070 value of 10 means strong shakiness. Default value is 5.
20073 Set the accuracy of the detection process. It must be a value in the
20074 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20075 accuracy. Default value is 15.
20078 Set stepsize of the search process. The region around minimum is
20079 scanned with 1 pixel resolution. Default value is 6.
20082 Set minimum contrast. Below this value a local measurement field is
20083 discarded. Must be a floating point value in the range 0-1. Default
20087 Set reference frame number for tripod mode.
20089 If enabled, the motion of the frames is compared to a reference frame
20090 in the filtered stream, identified by the specified number. The idea
20091 is to compensate all movements in a more-or-less static scene and keep
20092 the camera view absolutely still.
20094 If set to 0, it is disabled. The frames are counted starting from 1.
20097 Show fields and transforms in the resulting frames. It accepts an
20098 integer in the range 0-2. Default value is 0, which disables any
20102 @subsection Examples
20106 Use default values:
20112 Analyze strongly shaky movie and put the results in file
20113 @file{mytransforms.trf}:
20115 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20119 Visualize the result of internal transformations in the resulting
20122 vidstabdetect=show=1
20126 Analyze a video with medium shakiness using @command{ffmpeg}:
20128 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20132 @anchor{vidstabtransform}
20133 @section vidstabtransform
20135 Video stabilization/deshaking: pass 2 of 2,
20136 see @ref{vidstabdetect} for pass 1.
20138 Read a file with transform information for each frame and
20139 apply/compensate them. Together with the @ref{vidstabdetect}
20140 filter this can be used to deshake videos. See also
20141 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20142 the @ref{unsharp} filter, see below.
20144 To enable compilation of this filter you need to configure FFmpeg with
20145 @code{--enable-libvidstab}.
20147 @subsection Options
20151 Set path to the file used to read the transforms. Default value is
20152 @file{transforms.trf}.
20155 Set the number of frames (value*2 + 1) used for lowpass filtering the
20156 camera movements. Default value is 10.
20158 For example a number of 10 means that 21 frames are used (10 in the
20159 past and 10 in the future) to smoothen the motion in the video. A
20160 larger value leads to a smoother video, but limits the acceleration of
20161 the camera (pan/tilt movements). 0 is a special case where a static
20162 camera is simulated.
20165 Set the camera path optimization algorithm.
20167 Accepted values are:
20170 gaussian kernel low-pass filter on camera motion (default)
20172 averaging on transformations
20176 Set maximal number of pixels to translate frames. Default value is -1,
20180 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20181 value is -1, meaning no limit.
20184 Specify how to deal with borders that may be visible due to movement
20187 Available values are:
20190 keep image information from previous frame (default)
20192 fill the border black
20196 Invert transforms if set to 1. Default value is 0.
20199 Consider transforms as relative to previous frame if set to 1,
20200 absolute if set to 0. Default value is 0.
20203 Set percentage to zoom. A positive value will result in a zoom-in
20204 effect, a negative value in a zoom-out effect. Default value is 0 (no
20208 Set optimal zooming to avoid borders.
20210 Accepted values are:
20215 optimal static zoom value is determined (only very strong movements
20216 will lead to visible borders) (default)
20218 optimal adaptive zoom value is determined (no borders will be
20219 visible), see @option{zoomspeed}
20222 Note that the value given at zoom is added to the one calculated here.
20225 Set percent to zoom maximally each frame (enabled when
20226 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20230 Specify type of interpolation.
20232 Available values are:
20237 linear only horizontal
20239 linear in both directions (default)
20241 cubic in both directions (slow)
20245 Enable virtual tripod mode if set to 1, which is equivalent to
20246 @code{relative=0:smoothing=0}. Default value is 0.
20248 Use also @code{tripod} option of @ref{vidstabdetect}.
20251 Increase log verbosity if set to 1. Also the detected global motions
20252 are written to the temporary file @file{global_motions.trf}. Default
20256 @subsection Examples
20260 Use @command{ffmpeg} for a typical stabilization with default values:
20262 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20265 Note the use of the @ref{unsharp} filter which is always recommended.
20268 Zoom in a bit more and load transform data from a given file:
20270 vidstabtransform=zoom=5:input="mytransforms.trf"
20274 Smoothen the video even more:
20276 vidstabtransform=smoothing=30
20282 Flip the input video vertically.
20284 For example, to vertically flip a video with @command{ffmpeg}:
20286 ffmpeg -i in.avi -vf "vflip" out.avi
20291 Detect variable frame rate video.
20293 This filter tries to detect if the input is variable or constant frame rate.
20295 At end it will output number of frames detected as having variable delta pts,
20296 and ones with constant delta pts.
20297 If there was frames with variable delta, than it will also show min, max and
20298 average delta encountered.
20302 Boost or alter saturation.
20304 The filter accepts the following options:
20307 Set strength of boost if positive value or strength of alter if negative value.
20308 Default is 0. Allowed range is from -2 to 2.
20311 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20314 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20317 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20320 Set the red luma coefficient.
20323 Set the green luma coefficient.
20326 Set the blue luma coefficient.
20329 If @code{intensity} is negative and this is set to 1, colors will change,
20330 otherwise colors will be less saturated, more towards gray.
20333 @subsection Commands
20335 This filter supports the all above options as @ref{commands}.
20340 Make or reverse a natural vignetting effect.
20342 The filter accepts the following options:
20346 Set lens angle expression as a number of radians.
20348 The value is clipped in the @code{[0,PI/2]} range.
20350 Default value: @code{"PI/5"}
20354 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20358 Set forward/backward mode.
20360 Available modes are:
20363 The larger the distance from the central point, the darker the image becomes.
20366 The larger the distance from the central point, the brighter the image becomes.
20367 This can be used to reverse a vignette effect, though there is no automatic
20368 detection to extract the lens @option{angle} and other settings (yet). It can
20369 also be used to create a burning effect.
20372 Default value is @samp{forward}.
20375 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20377 It accepts the following values:
20380 Evaluate expressions only once during the filter initialization.
20383 Evaluate expressions for each incoming frame. This is way slower than the
20384 @samp{init} mode since it requires all the scalers to be re-computed, but it
20385 allows advanced dynamic expressions.
20388 Default value is @samp{init}.
20391 Set dithering to reduce the circular banding effects. Default is @code{1}
20395 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20396 Setting this value to the SAR of the input will make a rectangular vignetting
20397 following the dimensions of the video.
20399 Default is @code{1/1}.
20402 @subsection Expressions
20404 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20405 following parameters.
20410 input width and height
20413 the number of input frame, starting from 0
20416 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20417 @var{TB} units, NAN if undefined
20420 frame rate of the input video, NAN if the input frame rate is unknown
20423 the PTS (Presentation TimeStamp) of the filtered video frame,
20424 expressed in seconds, NAN if undefined
20427 time base of the input video
20431 @subsection Examples
20435 Apply simple strong vignetting effect:
20441 Make a flickering vignetting:
20443 vignette='PI/4+random(1)*PI/50':eval=frame
20448 @section vmafmotion
20450 Obtain the average VMAF motion score of a video.
20451 It is one of the component metrics of VMAF.
20453 The obtained average motion score is printed through the logging system.
20455 The filter accepts the following options:
20459 If specified, the filter will use the named file to save the motion score of
20460 each frame with respect to the previous frame.
20461 When filename equals "-" the data is sent to standard output.
20466 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20470 Stack input videos vertically.
20472 All streams must be of same pixel format and of same width.
20474 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20475 to create same output.
20477 The filter accepts the following options:
20481 Set number of input streams. Default is 2.
20484 If set to 1, force the output to terminate when the shortest input
20485 terminates. Default value is 0.
20490 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20491 Deinterlacing Filter").
20493 Based on the process described by Martin Weston for BBC R&D, and
20494 implemented based on the de-interlace algorithm written by Jim
20495 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20496 uses filter coefficients calculated by BBC R&D.
20498 This filter uses field-dominance information in frame to decide which
20499 of each pair of fields to place first in the output.
20500 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20502 There are two sets of filter coefficients, so called "simple"
20503 and "complex". Which set of filter coefficients is used can
20504 be set by passing an optional parameter:
20508 Set the interlacing filter coefficients. Accepts one of the following values:
20512 Simple filter coefficient set.
20514 More-complex filter coefficient set.
20516 Default value is @samp{complex}.
20519 Specify which frames to deinterlace. Accepts one of the following values:
20523 Deinterlace all frames,
20525 Only deinterlace frames marked as interlaced.
20528 Default value is @samp{all}.
20532 Video waveform monitor.
20534 The waveform monitor plots color component intensity. By default luminance
20535 only. Each column of the waveform corresponds to a column of pixels in the
20538 It accepts the following options:
20542 Can be either @code{row}, or @code{column}. Default is @code{column}.
20543 In row mode, the graph on the left side represents color component value 0 and
20544 the right side represents value = 255. In column mode, the top side represents
20545 color component value = 0 and bottom side represents value = 255.
20548 Set intensity. Smaller values are useful to find out how many values of the same
20549 luminance are distributed across input rows/columns.
20550 Default value is @code{0.04}. Allowed range is [0, 1].
20553 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20554 In mirrored mode, higher values will be represented on the left
20555 side for @code{row} mode and at the top for @code{column} mode. Default is
20556 @code{1} (mirrored).
20560 It accepts the following values:
20563 Presents information identical to that in the @code{parade}, except
20564 that the graphs representing color components are superimposed directly
20567 This display mode makes it easier to spot relative differences or similarities
20568 in overlapping areas of the color components that are supposed to be identical,
20569 such as neutral whites, grays, or blacks.
20572 Display separate graph for the color components side by side in
20573 @code{row} mode or one below the other in @code{column} mode.
20576 Display separate graph for the color components side by side in
20577 @code{column} mode or one below the other in @code{row} mode.
20579 Using this display mode makes it easy to spot color casts in the highlights
20580 and shadows of an image, by comparing the contours of the top and the bottom
20581 graphs of each waveform. Since whites, grays, and blacks are characterized
20582 by exactly equal amounts of red, green, and blue, neutral areas of the picture
20583 should display three waveforms of roughly equal width/height. If not, the
20584 correction is easy to perform by making level adjustments the three waveforms.
20586 Default is @code{stack}.
20588 @item components, c
20589 Set which color components to display. Default is 1, which means only luminance
20590 or red color component if input is in RGB colorspace. If is set for example to
20591 7 it will display all 3 (if) available color components.
20596 No envelope, this is default.
20599 Instant envelope, minimum and maximum values presented in graph will be easily
20600 visible even with small @code{step} value.
20603 Hold minimum and maximum values presented in graph across time. This way you
20604 can still spot out of range values without constantly looking at waveforms.
20607 Peak and instant envelope combined together.
20613 No filtering, this is default.
20616 Luma and chroma combined together.
20619 Similar as above, but shows difference between blue and red chroma.
20622 Similar as above, but use different colors.
20625 Similar as above, but again with different colors.
20628 Displays only chroma.
20631 Displays actual color value on waveform.
20634 Similar as above, but with luma showing frequency of chroma values.
20638 Set which graticule to display.
20642 Do not display graticule.
20645 Display green graticule showing legal broadcast ranges.
20648 Display orange graticule showing legal broadcast ranges.
20651 Display invert graticule showing legal broadcast ranges.
20655 Set graticule opacity.
20658 Set graticule flags.
20662 Draw numbers above lines. By default enabled.
20665 Draw dots instead of lines.
20669 Set scale used for displaying graticule.
20676 Default is digital.
20679 Set background opacity.
20683 Set tint for output.
20684 Only used with lowpass filter and when display is not overlay and input
20685 pixel formats are not RGB.
20688 @section weave, doubleweave
20690 The @code{weave} takes a field-based video input and join
20691 each two sequential fields into single frame, producing a new double
20692 height clip with half the frame rate and half the frame count.
20694 The @code{doubleweave} works same as @code{weave} but without
20695 halving frame rate and frame count.
20697 It accepts the following option:
20701 Set first field. Available values are:
20705 Set the frame as top-field-first.
20708 Set the frame as bottom-field-first.
20712 @subsection Examples
20716 Interlace video using @ref{select} and @ref{separatefields} filter:
20718 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
20723 Apply the xBR high-quality magnification filter which is designed for pixel
20724 art. It follows a set of edge-detection rules, see
20725 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
20727 It accepts the following option:
20731 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
20732 @code{3xBR} and @code{4} for @code{4xBR}.
20733 Default is @code{3}.
20738 Apply cross fade from one input video stream to another input video stream.
20739 The cross fade is applied for specified duration.
20741 The filter accepts the following options:
20745 Set one of available transition effects:
20793 Default transition effect is fade.
20796 Set cross fade duration in seconds.
20797 Default duration is 1 second.
20800 Set cross fade start relative to first input stream in seconds.
20801 Default offset is 0.
20804 Set expression for custom transition effect.
20806 The expressions can use the following variables and functions:
20811 The coordinates of the current sample.
20815 The width and height of the image.
20818 Progress of transition effect.
20821 Currently processed plane.
20824 Return value of first input at current location and plane.
20827 Return value of second input at current location and plane.
20833 Return the value of the pixel at location (@var{x},@var{y}) of the
20834 first/second/third/fourth component of first input.
20840 Return the value of the pixel at location (@var{x},@var{y}) of the
20841 first/second/third/fourth component of second input.
20845 @subsection Examples
20849 Cross fade from one input video to another input video, with fade transition and duration of transition
20850 of 2 seconds starting at offset of 5 seconds:
20852 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
20857 Pick median pixels from several input videos.
20859 The filter accepts the following options:
20863 Set number of inputs.
20864 Default is 3. Allowed range is from 3 to 255.
20865 If number of inputs is even number, than result will be mean value between two median values.
20868 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
20871 Set median percentile. Default value is @code{0.5}.
20872 Default value of @code{0.5} will pick always median values, while @code{0} will pick
20873 minimum values, and @code{1} maximum values.
20877 Stack video inputs into custom layout.
20879 All streams must be of same pixel format.
20881 The filter accepts the following options:
20885 Set number of input streams. Default is 2.
20888 Specify layout of inputs.
20889 This option requires the desired layout configuration to be explicitly set by the user.
20890 This sets position of each video input in output. Each input
20891 is separated by '|'.
20892 The first number represents the column, and the second number represents the row.
20893 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
20894 where X is video input from which to take width or height.
20895 Multiple values can be used when separated by '+'. In such
20896 case values are summed together.
20898 Note that if inputs are of different sizes gaps may appear, as not all of
20899 the output video frame will be filled. Similarly, videos can overlap each
20900 other if their position doesn't leave enough space for the full frame of
20903 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
20904 a layout must be set by the user.
20907 If set to 1, force the output to terminate when the shortest input
20908 terminates. Default value is 0.
20911 If set to valid color, all unused pixels will be filled with that color.
20912 By default fill is set to none, so it is disabled.
20915 @subsection Examples
20919 Display 4 inputs into 2x2 grid.
20923 input1(0, 0) | input3(w0, 0)
20924 input2(0, h0) | input4(w0, h0)
20928 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
20931 Note that if inputs are of different sizes, gaps or overlaps may occur.
20934 Display 4 inputs into 1x4 grid.
20941 input4(0, h0+h1+h2)
20945 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
20948 Note that if inputs are of different widths, unused space will appear.
20951 Display 9 inputs into 3x3 grid.
20955 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
20956 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
20957 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
20961 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
20964 Note that if inputs are of different sizes, gaps or overlaps may occur.
20967 Display 16 inputs into 4x4 grid.
20971 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
20972 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
20973 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
20974 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
20978 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|
20979 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
20982 Note that if inputs are of different sizes, gaps or overlaps may occur.
20989 Deinterlace the input video ("yadif" means "yet another deinterlacing
20992 It accepts the following parameters:
20998 The interlacing mode to adopt. It accepts one of the following values:
21001 @item 0, send_frame
21002 Output one frame for each frame.
21003 @item 1, send_field
21004 Output one frame for each field.
21005 @item 2, send_frame_nospatial
21006 Like @code{send_frame}, but it skips the spatial interlacing check.
21007 @item 3, send_field_nospatial
21008 Like @code{send_field}, but it skips the spatial interlacing check.
21011 The default value is @code{send_frame}.
21014 The picture field parity assumed for the input interlaced video. It accepts one
21015 of the following values:
21019 Assume the top field is first.
21021 Assume the bottom field is first.
21023 Enable automatic detection of field parity.
21026 The default value is @code{auto}.
21027 If the interlacing is unknown or the decoder does not export this information,
21028 top field first will be assumed.
21031 Specify which frames to deinterlace. Accepts one of the following
21036 Deinterlace all frames.
21037 @item 1, interlaced
21038 Only deinterlace frames marked as interlaced.
21041 The default value is @code{all}.
21044 @section yadif_cuda
21046 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21047 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21050 It accepts the following parameters:
21056 The interlacing mode to adopt. It accepts one of the following values:
21059 @item 0, send_frame
21060 Output one frame for each frame.
21061 @item 1, send_field
21062 Output one frame for each field.
21063 @item 2, send_frame_nospatial
21064 Like @code{send_frame}, but it skips the spatial interlacing check.
21065 @item 3, send_field_nospatial
21066 Like @code{send_field}, but it skips the spatial interlacing check.
21069 The default value is @code{send_frame}.
21072 The picture field parity assumed for the input interlaced video. It accepts one
21073 of the following values:
21077 Assume the top field is first.
21079 Assume the bottom field is first.
21081 Enable automatic detection of field parity.
21084 The default value is @code{auto}.
21085 If the interlacing is unknown or the decoder does not export this information,
21086 top field first will be assumed.
21089 Specify which frames to deinterlace. Accepts one of the following
21094 Deinterlace all frames.
21095 @item 1, interlaced
21096 Only deinterlace frames marked as interlaced.
21099 The default value is @code{all}.
21104 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21105 The algorithm is described in
21106 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21108 It accepts the following parameters:
21112 Set the window radius. Default value is 3.
21115 Set which planes to filter. Default is only the first plane.
21118 Set blur strength. Default value is 128.
21121 @subsection Commands
21122 This filter supports same @ref{commands} as options.
21126 Apply Zoom & Pan effect.
21128 This filter accepts the following options:
21132 Set the zoom expression. Range is 1-10. Default is 1.
21136 Set the x and y expression. Default is 0.
21139 Set the duration expression in number of frames.
21140 This sets for how many number of frames effect will last for
21141 single input image.
21144 Set the output image size, default is 'hd720'.
21147 Set the output frame rate, default is '25'.
21150 Each expression can contain the following constants:
21169 Output frame count.
21172 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21174 @item out_time, time, ot
21175 The output timestamp expressed in seconds.
21179 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21180 for current input frame.
21184 'x' and 'y' of last output frame of previous input frame or 0 when there was
21185 not yet such frame (first input frame).
21188 Last calculated zoom from 'z' expression for current input frame.
21191 Last calculated zoom of last output frame of previous input frame.
21194 Number of output frames for current input frame. Calculated from 'd' expression
21195 for each input frame.
21198 number of output frames created for previous input frame
21201 Rational number: input width / input height
21204 sample aspect ratio
21207 display aspect ratio
21211 @subsection Examples
21215 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21217 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
21221 Zoom in up to 1.5x and pan always at center of picture:
21223 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21227 Same as above but without pausing:
21229 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21233 Zoom in 2x into center of picture only for the first second of the input video:
21235 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21242 Scale (resize) the input video, using the z.lib library:
21243 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21244 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21246 The zscale filter forces the output display aspect ratio to be the same
21247 as the input, by changing the output sample aspect ratio.
21249 If the input image format is different from the format requested by
21250 the next filter, the zscale filter will convert the input to the
21253 @subsection Options
21254 The filter accepts the following options.
21259 Set the output video dimension expression. Default value is the input
21262 If the @var{width} or @var{w} value is 0, the input width is used for
21263 the output. If the @var{height} or @var{h} value is 0, the input height
21264 is used for the output.
21266 If one and only one of the values is -n with n >= 1, the zscale filter
21267 will use a value that maintains the aspect ratio of the input image,
21268 calculated from the other specified dimension. After that it will,
21269 however, make sure that the calculated dimension is divisible by n and
21270 adjust the value if necessary.
21272 If both values are -n with n >= 1, the behavior will be identical to
21273 both values being set to 0 as previously detailed.
21275 See below for the list of accepted constants for use in the dimension
21279 Set the video size. For the syntax of this option, check the
21280 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21283 Set the dither type.
21285 Possible values are:
21290 @item error_diffusion
21296 Set the resize filter type.
21298 Possible values are:
21308 Default is bilinear.
21311 Set the color range.
21313 Possible values are:
21320 Default is same as input.
21323 Set the color primaries.
21325 Possible values are:
21335 Default is same as input.
21338 Set the transfer characteristics.
21340 Possible values are:
21354 Default is same as input.
21357 Set the colorspace matrix.
21359 Possible value are:
21370 Default is same as input.
21373 Set the input color range.
21375 Possible values are:
21382 Default is same as input.
21384 @item primariesin, pin
21385 Set the input color primaries.
21387 Possible values are:
21397 Default is same as input.
21399 @item transferin, tin
21400 Set the input transfer characteristics.
21402 Possible values are:
21413 Default is same as input.
21415 @item matrixin, min
21416 Set the input colorspace matrix.
21418 Possible value are:
21430 Set the output chroma location.
21432 Possible values are:
21443 @item chromalin, cin
21444 Set the input chroma location.
21446 Possible values are:
21458 Set the nominal peak luminance.
21461 The values of the @option{w} and @option{h} options are expressions
21462 containing the following constants:
21467 The input width and height
21471 These are the same as @var{in_w} and @var{in_h}.
21475 The output (scaled) width and height
21479 These are the same as @var{out_w} and @var{out_h}
21482 The same as @var{iw} / @var{ih}
21485 input sample aspect ratio
21488 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21492 horizontal and vertical input chroma subsample values. For example for the
21493 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21497 horizontal and vertical output chroma subsample values. For example for the
21498 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21501 @subsection Commands
21503 This filter supports the following commands:
21507 Set the output video dimension expression.
21508 The command accepts the same syntax of the corresponding option.
21510 If the specified expression is not valid, it is kept at its current
21514 @c man end VIDEO FILTERS
21516 @chapter OpenCL Video Filters
21517 @c man begin OPENCL VIDEO FILTERS
21519 Below is a description of the currently available OpenCL video filters.
21521 To enable compilation of these filters you need to configure FFmpeg with
21522 @code{--enable-opencl}.
21524 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21527 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21528 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21529 given device parameters.
21531 @item -filter_hw_device @var{name}
21532 Pass the hardware device called @var{name} to all filters in any filter graph.
21536 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21540 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21542 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21546 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.
21548 @section avgblur_opencl
21550 Apply average blur filter.
21552 The filter accepts the following options:
21556 Set horizontal radius size.
21557 Range is @code{[1, 1024]} and default value is @code{1}.
21560 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21563 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21566 @subsection Example
21570 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.
21572 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21576 @section boxblur_opencl
21578 Apply a boxblur algorithm to the input video.
21580 It accepts the following parameters:
21584 @item luma_radius, lr
21585 @item luma_power, lp
21586 @item chroma_radius, cr
21587 @item chroma_power, cp
21588 @item alpha_radius, ar
21589 @item alpha_power, ap
21593 A description of the accepted options follows.
21596 @item luma_radius, lr
21597 @item chroma_radius, cr
21598 @item alpha_radius, ar
21599 Set an expression for the box radius in pixels used for blurring the
21600 corresponding input plane.
21602 The radius value must be a non-negative number, and must not be
21603 greater than the value of the expression @code{min(w,h)/2} for the
21604 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
21607 Default value for @option{luma_radius} is "2". If not specified,
21608 @option{chroma_radius} and @option{alpha_radius} default to the
21609 corresponding value set for @option{luma_radius}.
21611 The expressions can contain the following constants:
21615 The input width and height in pixels.
21619 The input chroma image width and height in pixels.
21623 The horizontal and vertical chroma subsample values. For example, for the
21624 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
21627 @item luma_power, lp
21628 @item chroma_power, cp
21629 @item alpha_power, ap
21630 Specify how many times the boxblur filter is applied to the
21631 corresponding plane.
21633 Default value for @option{luma_power} is 2. If not specified,
21634 @option{chroma_power} and @option{alpha_power} default to the
21635 corresponding value set for @option{luma_power}.
21637 A value of 0 will disable the effect.
21640 @subsection Examples
21642 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.
21646 Apply a boxblur filter with the luma, chroma, and alpha radius
21647 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.
21649 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
21650 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
21654 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.
21656 For the luma plane, a 2x2 box radius will be run once.
21658 For the chroma plane, a 4x4 box radius will be run 5 times.
21660 For the alpha plane, a 3x3 box radius will be run 7 times.
21662 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
21666 @section colorkey_opencl
21667 RGB colorspace color keying.
21669 The filter accepts the following options:
21673 The color which will be replaced with transparency.
21676 Similarity percentage with the key color.
21678 0.01 matches only the exact key color, while 1.0 matches everything.
21683 0.0 makes pixels either fully transparent, or not transparent at all.
21685 Higher values result in semi-transparent pixels, with a higher transparency
21686 the more similar the pixels color is to the key color.
21689 @subsection Examples
21693 Make every semi-green pixel in the input transparent with some slight blending:
21695 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
21699 @section convolution_opencl
21701 Apply convolution of 3x3, 5x5, 7x7 matrix.
21703 The filter accepts the following options:
21710 Set matrix for each plane.
21711 Matrix is sequence of 9, 25 or 49 signed numbers.
21712 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
21718 Set multiplier for calculated value for each plane.
21719 If unset or 0, it will be sum of all matrix elements.
21720 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
21726 Set bias for each plane. This value is added to the result of the multiplication.
21727 Useful for making the overall image brighter or darker.
21728 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
21732 @subsection Examples
21738 -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
21744 -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
21748 Apply edge enhance:
21750 -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
21756 -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
21760 Apply laplacian edge detector which includes diagonals:
21762 -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
21768 -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
21772 @section erosion_opencl
21774 Apply erosion effect to the video.
21776 This filter replaces the pixel by the local(3x3) minimum.
21778 It accepts the following options:
21785 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21786 If @code{0}, plane will remain unchanged.
21789 Flag which specifies the pixel to refer to.
21790 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21792 Flags to local 3x3 coordinates region centered on @code{x}:
21801 @subsection Example
21805 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.
21807 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21811 @section deshake_opencl
21812 Feature-point based video stabilization filter.
21814 The filter accepts the following options:
21818 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
21821 Whether or not additional debug info should be displayed, both in the processed output and in the console.
21823 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
21825 Viewing point matches in the output video is only supported for RGB input.
21827 Defaults to @code{0}.
21829 @item adaptive_crop
21830 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
21832 Defaults to @code{1}.
21834 @item refine_features
21835 Whether or not feature points should be refined at a sub-pixel level.
21837 This can be turned off for a slight performance gain at the cost of precision.
21839 Defaults to @code{1}.
21841 @item smooth_strength
21842 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
21844 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
21846 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
21848 Defaults to @code{0.0}.
21850 @item smooth_window_multiplier
21851 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
21853 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
21855 Acceptable values range from @code{0.1} to @code{10.0}.
21857 Larger values increase the amount of motion data available for determining how to smooth the camera path,
21858 potentially improving smoothness, but also increase latency and memory usage.
21860 Defaults to @code{2.0}.
21864 @subsection Examples
21868 Stabilize a video with a fixed, medium smoothing strength:
21870 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
21874 Stabilize a video with debugging (both in console and in rendered video):
21876 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
21880 @section dilation_opencl
21882 Apply dilation effect to the video.
21884 This filter replaces the pixel by the local(3x3) maximum.
21886 It accepts the following options:
21893 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21894 If @code{0}, plane will remain unchanged.
21897 Flag which specifies the pixel to refer to.
21898 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21900 Flags to local 3x3 coordinates region centered on @code{x}:
21909 @subsection Example
21913 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.
21915 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21919 @section nlmeans_opencl
21921 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
21923 @section overlay_opencl
21925 Overlay one video on top of another.
21927 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
21928 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
21930 The filter accepts the following options:
21935 Set the x coordinate of the overlaid video on the main video.
21936 Default value is @code{0}.
21939 Set the y coordinate of the overlaid video on the main video.
21940 Default value is @code{0}.
21944 @subsection Examples
21948 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
21950 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21953 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
21955 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21960 @section pad_opencl
21962 Add paddings to the input image, and place the original input at the
21963 provided @var{x}, @var{y} coordinates.
21965 It accepts the following options:
21970 Specify an expression for the size of the output image with the
21971 paddings added. If the value for @var{width} or @var{height} is 0, the
21972 corresponding input size is used for the output.
21974 The @var{width} expression can reference the value set by the
21975 @var{height} expression, and vice versa.
21977 The default value of @var{width} and @var{height} is 0.
21981 Specify the offsets to place the input image at within the padded area,
21982 with respect to the top/left border of the output image.
21984 The @var{x} expression can reference the value set by the @var{y}
21985 expression, and vice versa.
21987 The default value of @var{x} and @var{y} is 0.
21989 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
21990 so the input image is centered on the padded area.
21993 Specify the color of the padded area. For the syntax of this option,
21994 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
21995 manual,ffmpeg-utils}.
21998 Pad to an aspect instead to a resolution.
22001 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22002 options are expressions containing the following constants:
22007 The input video width and height.
22011 These are the same as @var{in_w} and @var{in_h}.
22015 The output width and height (the size of the padded area), as
22016 specified by the @var{width} and @var{height} expressions.
22020 These are the same as @var{out_w} and @var{out_h}.
22024 The x and y offsets as specified by the @var{x} and @var{y}
22025 expressions, or NAN if not yet specified.
22028 same as @var{iw} / @var{ih}
22031 input sample aspect ratio
22034 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22037 @section prewitt_opencl
22039 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22041 The filter accepts the following option:
22045 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22048 Set value which will be multiplied with filtered result.
22049 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22052 Set value which will be added to filtered result.
22053 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22056 @subsection Example
22060 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22062 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22066 @anchor{program_opencl}
22067 @section program_opencl
22069 Filter video using an OpenCL program.
22074 OpenCL program source file.
22077 Kernel name in program.
22080 Number of inputs to the filter. Defaults to 1.
22083 Size of output frames. Defaults to the same as the first input.
22087 The @code{program_opencl} filter also supports the @ref{framesync} options.
22089 The program source file must contain a kernel function with the given name,
22090 which will be run once for each plane of the output. Each run on a plane
22091 gets enqueued as a separate 2D global NDRange with one work-item for each
22092 pixel to be generated. The global ID offset for each work-item is therefore
22093 the coordinates of a pixel in the destination image.
22095 The kernel function needs to take the following arguments:
22098 Destination image, @var{__write_only image2d_t}.
22100 This image will become the output; the kernel should write all of it.
22102 Frame index, @var{unsigned int}.
22104 This is a counter starting from zero and increasing by one for each frame.
22106 Source images, @var{__read_only image2d_t}.
22108 These are the most recent images on each input. The kernel may read from
22109 them to generate the output, but they can't be written to.
22116 Copy the input to the output (output must be the same size as the input).
22118 __kernel void copy(__write_only image2d_t destination,
22119 unsigned int index,
22120 __read_only image2d_t source)
22122 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22124 int2 location = (int2)(get_global_id(0), get_global_id(1));
22126 float4 value = read_imagef(source, sampler, location);
22128 write_imagef(destination, location, value);
22133 Apply a simple transformation, rotating the input by an amount increasing
22134 with the index counter. Pixel values are linearly interpolated by the
22135 sampler, and the output need not have the same dimensions as the input.
22137 __kernel void rotate_image(__write_only image2d_t dst,
22138 unsigned int index,
22139 __read_only image2d_t src)
22141 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22142 CLK_FILTER_LINEAR);
22144 float angle = (float)index / 100.0f;
22146 float2 dst_dim = convert_float2(get_image_dim(dst));
22147 float2 src_dim = convert_float2(get_image_dim(src));
22149 float2 dst_cen = dst_dim / 2.0f;
22150 float2 src_cen = src_dim / 2.0f;
22152 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22154 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22156 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22157 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22159 src_pos = src_pos * src_dim / dst_dim;
22161 float2 src_loc = src_pos + src_cen;
22163 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22164 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22165 write_imagef(dst, dst_loc, 0.5f);
22167 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22172 Blend two inputs together, with the amount of each input used varying
22173 with the index counter.
22175 __kernel void blend_images(__write_only image2d_t dst,
22176 unsigned int index,
22177 __read_only image2d_t src1,
22178 __read_only image2d_t src2)
22180 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22181 CLK_FILTER_LINEAR);
22183 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22185 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22186 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22187 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22189 float4 val1 = read_imagef(src1, sampler, src1_loc);
22190 float4 val2 = read_imagef(src2, sampler, src2_loc);
22192 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22198 @section roberts_opencl
22199 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22201 The filter accepts the following option:
22205 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22208 Set value which will be multiplied with filtered result.
22209 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22212 Set value which will be added to filtered result.
22213 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22216 @subsection Example
22220 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22222 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22226 @section sobel_opencl
22228 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22230 The filter accepts the following option:
22234 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22237 Set value which will be multiplied with filtered result.
22238 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22241 Set value which will be added to filtered result.
22242 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22245 @subsection Example
22249 Apply sobel operator with scale set to 2 and delta set to 10
22251 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22255 @section tonemap_opencl
22257 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22259 It accepts the following parameters:
22263 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22266 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22269 Apply desaturation for highlights that exceed this level of brightness. The
22270 higher the parameter, the more color information will be preserved. This
22271 setting helps prevent unnaturally blown-out colors for super-highlights, by
22272 (smoothly) turning into white instead. This makes images feel more natural,
22273 at the cost of reducing information about out-of-range colors.
22275 The default value is 0.5, and the algorithm here is a little different from
22276 the cpu version tonemap currently. A setting of 0.0 disables this option.
22279 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22280 is used to detect whether the scene has changed or not. If the distance between
22281 the current frame average brightness and the current running average exceeds
22282 a threshold value, we would re-calculate scene average and peak brightness.
22283 The default value is 0.2.
22286 Specify the output pixel format.
22288 Currently supported formats are:
22295 Set the output color range.
22297 Possible values are:
22303 Default is same as input.
22306 Set the output color primaries.
22308 Possible values are:
22314 Default is same as input.
22317 Set the output transfer characteristics.
22319 Possible values are:
22328 Set the output colorspace matrix.
22330 Possible value are:
22336 Default is same as input.
22340 @subsection Example
22344 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22346 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22350 @section unsharp_opencl
22352 Sharpen or blur the input video.
22354 It accepts the following parameters:
22357 @item luma_msize_x, lx
22358 Set the luma matrix horizontal size.
22359 Range is @code{[1, 23]} and default value is @code{5}.
22361 @item luma_msize_y, ly
22362 Set the luma matrix vertical size.
22363 Range is @code{[1, 23]} and default value is @code{5}.
22365 @item luma_amount, la
22366 Set the luma effect strength.
22367 Range is @code{[-10, 10]} and default value is @code{1.0}.
22369 Negative values will blur the input video, while positive values will
22370 sharpen it, a value of zero will disable the effect.
22372 @item chroma_msize_x, cx
22373 Set the chroma matrix horizontal size.
22374 Range is @code{[1, 23]} and default value is @code{5}.
22376 @item chroma_msize_y, cy
22377 Set the chroma matrix vertical size.
22378 Range is @code{[1, 23]} and default value is @code{5}.
22380 @item chroma_amount, ca
22381 Set the chroma effect strength.
22382 Range is @code{[-10, 10]} and default value is @code{0.0}.
22384 Negative values will blur the input video, while positive values will
22385 sharpen it, a value of zero will disable the effect.
22389 All parameters are optional and default to the equivalent of the
22390 string '5:5:1.0:5:5:0.0'.
22392 @subsection Examples
22396 Apply strong luma sharpen effect:
22398 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22402 Apply a strong blur of both luma and chroma parameters:
22404 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22408 @section xfade_opencl
22410 Cross fade two videos with custom transition effect by using OpenCL.
22412 It accepts the following options:
22416 Set one of possible transition effects.
22420 Select custom transition effect, the actual transition description
22421 will be picked from source and kernel options.
22433 Default transition is fade.
22437 OpenCL program source file for custom transition.
22440 Set name of kernel to use for custom transition from program source file.
22443 Set duration of video transition.
22446 Set time of start of transition relative to first video.
22449 The program source file must contain a kernel function with the given name,
22450 which will be run once for each plane of the output. Each run on a plane
22451 gets enqueued as a separate 2D global NDRange with one work-item for each
22452 pixel to be generated. The global ID offset for each work-item is therefore
22453 the coordinates of a pixel in the destination image.
22455 The kernel function needs to take the following arguments:
22458 Destination image, @var{__write_only image2d_t}.
22460 This image will become the output; the kernel should write all of it.
22463 First Source image, @var{__read_only image2d_t}.
22464 Second Source image, @var{__read_only image2d_t}.
22466 These are the most recent images on each input. The kernel may read from
22467 them to generate the output, but they can't be written to.
22470 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22477 Apply dots curtain transition effect:
22479 __kernel void blend_images(__write_only image2d_t dst,
22480 __read_only image2d_t src1,
22481 __read_only image2d_t src2,
22484 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22485 CLK_FILTER_LINEAR);
22486 int2 p = (int2)(get_global_id(0), get_global_id(1));
22487 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22488 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22491 float2 dots = (float2)(20.0, 20.0);
22492 float2 center = (float2)(0,0);
22495 float4 val1 = read_imagef(src1, sampler, p);
22496 float4 val2 = read_imagef(src2, sampler, p);
22497 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22499 write_imagef(dst, p, next ? val1 : val2);
22505 @c man end OPENCL VIDEO FILTERS
22507 @chapter VAAPI Video Filters
22508 @c man begin VAAPI VIDEO FILTERS
22510 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22512 To enable compilation of these filters you need to configure FFmpeg with
22513 @code{--enable-vaapi}.
22515 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}
22517 @section tonemap_vaapi
22519 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22520 It maps the dynamic range of HDR10 content to the SDR content.
22521 It currently only accepts HDR10 as input.
22523 It accepts the following parameters:
22527 Specify the output pixel format.
22529 Currently supported formats are:
22538 Set the output color primaries.
22540 Default is same as input.
22543 Set the output transfer characteristics.
22548 Set the output colorspace matrix.
22550 Default is same as input.
22554 @subsection Example
22558 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22560 tonemap_vaapi=format=p010:t=bt2020-10
22564 @c man end VAAPI VIDEO FILTERS
22566 @chapter Video Sources
22567 @c man begin VIDEO SOURCES
22569 Below is a description of the currently available video sources.
22573 Buffer video frames, and make them available to the filter chain.
22575 This source is mainly intended for a programmatic use, in particular
22576 through the interface defined in @file{libavfilter/buffersrc.h}.
22578 It accepts the following parameters:
22583 Specify the size (width and height) of the buffered video frames. For the
22584 syntax of this option, check the
22585 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22588 The input video width.
22591 The input video height.
22594 A string representing the pixel format of the buffered video frames.
22595 It may be a number corresponding to a pixel format, or a pixel format
22599 Specify the timebase assumed by the timestamps of the buffered frames.
22602 Specify the frame rate expected for the video stream.
22604 @item pixel_aspect, sar
22605 The sample (pixel) aspect ratio of the input video.
22608 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
22609 to the filtergraph description to specify swscale flags for automatically
22610 inserted scalers. See @ref{Filtergraph syntax}.
22612 @item hw_frames_ctx
22613 When using a hardware pixel format, this should be a reference to an
22614 AVHWFramesContext describing input frames.
22619 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
22622 will instruct the source to accept video frames with size 320x240 and
22623 with format "yuv410p", assuming 1/24 as the timestamps timebase and
22624 square pixels (1:1 sample aspect ratio).
22625 Since the pixel format with name "yuv410p" corresponds to the number 6
22626 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
22627 this example corresponds to:
22629 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
22632 Alternatively, the options can be specified as a flat string, but this
22633 syntax is deprecated:
22635 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
22639 Create a pattern generated by an elementary cellular automaton.
22641 The initial state of the cellular automaton can be defined through the
22642 @option{filename} and @option{pattern} options. If such options are
22643 not specified an initial state is created randomly.
22645 At each new frame a new row in the video is filled with the result of
22646 the cellular automaton next generation. The behavior when the whole
22647 frame is filled is defined by the @option{scroll} option.
22649 This source accepts the following options:
22653 Read the initial cellular automaton state, i.e. the starting row, from
22654 the specified file.
22655 In the file, each non-whitespace character is considered an alive
22656 cell, a newline will terminate the row, and further characters in the
22657 file will be ignored.
22660 Read the initial cellular automaton state, i.e. the starting row, from
22661 the specified string.
22663 Each non-whitespace character in the string is considered an alive
22664 cell, a newline will terminate the row, and further characters in the
22665 string will be ignored.
22668 Set the video rate, that is the number of frames generated per second.
22671 @item random_fill_ratio, ratio
22672 Set the random fill ratio for the initial cellular automaton row. It
22673 is a floating point number value ranging from 0 to 1, defaults to
22676 This option is ignored when a file or a pattern is specified.
22678 @item random_seed, seed
22679 Set the seed for filling randomly the initial row, must be an integer
22680 included between 0 and UINT32_MAX. If not specified, or if explicitly
22681 set to -1, the filter will try to use a good random seed on a best
22685 Set the cellular automaton rule, it is a number ranging from 0 to 255.
22686 Default value is 110.
22689 Set the size of the output video. For the syntax of this option, check the
22690 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22692 If @option{filename} or @option{pattern} is specified, the size is set
22693 by default to the width of the specified initial state row, and the
22694 height is set to @var{width} * PHI.
22696 If @option{size} is set, it must contain the width of the specified
22697 pattern string, and the specified pattern will be centered in the
22700 If a filename or a pattern string is not specified, the size value
22701 defaults to "320x518" (used for a randomly generated initial state).
22704 If set to 1, scroll the output upward when all the rows in the output
22705 have been already filled. If set to 0, the new generated row will be
22706 written over the top row just after the bottom row is filled.
22709 @item start_full, full
22710 If set to 1, completely fill the output with generated rows before
22711 outputting the first frame.
22712 This is the default behavior, for disabling set the value to 0.
22715 If set to 1, stitch the left and right row edges together.
22716 This is the default behavior, for disabling set the value to 0.
22719 @subsection Examples
22723 Read the initial state from @file{pattern}, and specify an output of
22726 cellauto=f=pattern:s=200x400
22730 Generate a random initial row with a width of 200 cells, with a fill
22733 cellauto=ratio=2/3:s=200x200
22737 Create a pattern generated by rule 18 starting by a single alive cell
22738 centered on an initial row with width 100:
22740 cellauto=p=@@:s=100x400:full=0:rule=18
22744 Specify a more elaborated initial pattern:
22746 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
22751 @anchor{coreimagesrc}
22752 @section coreimagesrc
22753 Video source generated on GPU using Apple's CoreImage API on OSX.
22755 This video source is a specialized version of the @ref{coreimage} video filter.
22756 Use a core image generator at the beginning of the applied filterchain to
22757 generate the content.
22759 The coreimagesrc video source accepts the following options:
22761 @item list_generators
22762 List all available generators along with all their respective options as well as
22763 possible minimum and maximum values along with the default values.
22765 list_generators=true
22769 Specify the size of the sourced video. For the syntax of this option, check the
22770 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22771 The default value is @code{320x240}.
22774 Specify the frame rate of the sourced video, as the number of frames
22775 generated per second. It has to be a string in the format
22776 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22777 number or a valid video frame rate abbreviation. The default value is
22781 Set the sample aspect ratio of the sourced video.
22784 Set the duration of the sourced video. See
22785 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22786 for the accepted syntax.
22788 If not specified, or the expressed duration is negative, the video is
22789 supposed to be generated forever.
22792 Additionally, all options of the @ref{coreimage} video filter are accepted.
22793 A complete filterchain can be used for further processing of the
22794 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
22795 and examples for details.
22797 @subsection Examples
22802 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
22803 given as complete and escaped command-line for Apple's standard bash shell:
22805 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
22807 This example is equivalent to the QRCode example of @ref{coreimage} without the
22808 need for a nullsrc video source.
22813 Generate several gradients.
22817 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22818 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22821 Set frame rate, expressed as number of frames per second. Default
22824 @item c0, c1, c2, c3, c4, c5, c6, c7
22825 Set 8 colors. Default values for colors is to pick random one.
22827 @item x0, y0, y0, y1
22828 Set gradient line source and destination points. If negative or out of range, random ones
22832 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
22835 Set seed for picking gradient line points.
22838 Set the duration of the sourced video. See
22839 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22840 for the accepted syntax.
22842 If not specified, or the expressed duration is negative, the video is
22843 supposed to be generated forever.
22846 Set speed of gradients rotation.
22850 @section mandelbrot
22852 Generate a Mandelbrot set fractal, and progressively zoom towards the
22853 point specified with @var{start_x} and @var{start_y}.
22855 This source accepts the following options:
22860 Set the terminal pts value. Default value is 400.
22863 Set the terminal scale value.
22864 Must be a floating point value. Default value is 0.3.
22867 Set the inner coloring mode, that is the algorithm used to draw the
22868 Mandelbrot fractal internal region.
22870 It shall assume one of the following values:
22875 Show time until convergence.
22877 Set color based on point closest to the origin of the iterations.
22882 Default value is @var{mincol}.
22885 Set the bailout value. Default value is 10.0.
22888 Set the maximum of iterations performed by the rendering
22889 algorithm. Default value is 7189.
22892 Set outer coloring mode.
22893 It shall assume one of following values:
22895 @item iteration_count
22896 Set iteration count mode.
22897 @item normalized_iteration_count
22898 set normalized iteration count mode.
22900 Default value is @var{normalized_iteration_count}.
22903 Set frame rate, expressed as number of frames per second. Default
22907 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22908 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22911 Set the initial scale value. Default value is 3.0.
22914 Set the initial x position. Must be a floating point value between
22915 -100 and 100. Default value is -0.743643887037158704752191506114774.
22918 Set the initial y position. Must be a floating point value between
22919 -100 and 100. Default value is -0.131825904205311970493132056385139.
22924 Generate various test patterns, as generated by the MPlayer test filter.
22926 The size of the generated video is fixed, and is 256x256.
22927 This source is useful in particular for testing encoding features.
22929 This source accepts the following options:
22934 Specify the frame rate of the sourced video, as the number of frames
22935 generated per second. It has to be a string in the format
22936 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22937 number or a valid video frame rate abbreviation. The default value is
22941 Set the duration of the sourced video. See
22942 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22943 for the accepted syntax.
22945 If not specified, or the expressed duration is negative, the video is
22946 supposed to be generated forever.
22950 Set the number or the name of the test to perform. Supported tests are:
22964 @item max_frames, m
22965 Set the maximum number of frames generated for each test, default value is 30.
22969 Default value is "all", which will cycle through the list of all tests.
22974 mptestsrc=t=dc_luma
22977 will generate a "dc_luma" test pattern.
22979 @section frei0r_src
22981 Provide a frei0r source.
22983 To enable compilation of this filter you need to install the frei0r
22984 header and configure FFmpeg with @code{--enable-frei0r}.
22986 This source accepts the following parameters:
22991 The size of the video to generate. For the syntax of this option, check the
22992 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22995 The framerate of the generated video. It may be a string of the form
22996 @var{num}/@var{den} or a frame rate abbreviation.
22999 The name to the frei0r source to load. For more information regarding frei0r and
23000 how to set the parameters, read the @ref{frei0r} section in the video filters
23003 @item filter_params
23004 A '|'-separated list of parameters to pass to the frei0r source.
23008 For example, to generate a frei0r partik0l source with size 200x200
23009 and frame rate 10 which is overlaid on the overlay filter main input:
23011 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23016 Generate a life pattern.
23018 This source is based on a generalization of John Conway's life game.
23020 The sourced input represents a life grid, each pixel represents a cell
23021 which can be in one of two possible states, alive or dead. Every cell
23022 interacts with its eight neighbours, which are the cells that are
23023 horizontally, vertically, or diagonally adjacent.
23025 At each interaction the grid evolves according to the adopted rule,
23026 which specifies the number of neighbor alive cells which will make a
23027 cell stay alive or born. The @option{rule} option allows one to specify
23030 This source accepts the following options:
23034 Set the file from which to read the initial grid state. In the file,
23035 each non-whitespace character is considered an alive cell, and newline
23036 is used to delimit the end of each row.
23038 If this option is not specified, the initial grid is generated
23042 Set the video rate, that is the number of frames generated per second.
23045 @item random_fill_ratio, ratio
23046 Set the random fill ratio for the initial random grid. It is a
23047 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23048 It is ignored when a file is specified.
23050 @item random_seed, seed
23051 Set the seed for filling the initial random grid, must be an integer
23052 included between 0 and UINT32_MAX. If not specified, or if explicitly
23053 set to -1, the filter will try to use a good random seed on a best
23059 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23060 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23061 @var{NS} specifies the number of alive neighbor cells which make a
23062 live cell stay alive, and @var{NB} the number of alive neighbor cells
23063 which make a dead cell to become alive (i.e. to "born").
23064 "s" and "b" can be used in place of "S" and "B", respectively.
23066 Alternatively a rule can be specified by an 18-bits integer. The 9
23067 high order bits are used to encode the next cell state if it is alive
23068 for each number of neighbor alive cells, the low order bits specify
23069 the rule for "borning" new cells. Higher order bits encode for an
23070 higher number of neighbor cells.
23071 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23072 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23074 Default value is "S23/B3", which is the original Conway's game of life
23075 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23076 cells, and will born a new cell if there are three alive cells around
23080 Set the size of the output video. For the syntax of this option, check the
23081 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23083 If @option{filename} is specified, the size is set by default to the
23084 same size of the input file. If @option{size} is set, it must contain
23085 the size specified in the input file, and the initial grid defined in
23086 that file is centered in the larger resulting area.
23088 If a filename is not specified, the size value defaults to "320x240"
23089 (used for a randomly generated initial grid).
23092 If set to 1, stitch the left and right grid edges together, and the
23093 top and bottom edges also. Defaults to 1.
23096 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23097 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23098 value from 0 to 255.
23101 Set the color of living (or new born) cells.
23104 Set the color of dead cells. If @option{mold} is set, this is the first color
23105 used to represent a dead cell.
23108 Set mold color, for definitely dead and moldy cells.
23110 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23111 ffmpeg-utils manual,ffmpeg-utils}.
23114 @subsection Examples
23118 Read a grid from @file{pattern}, and center it on a grid of size
23121 life=f=pattern:s=300x300
23125 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23127 life=ratio=2/3:s=200x200
23131 Specify a custom rule for evolving a randomly generated grid:
23137 Full example with slow death effect (mold) using @command{ffplay}:
23139 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23146 @anchor{haldclutsrc}
23149 @anchor{pal100bars}
23150 @anchor{rgbtestsrc}
23152 @anchor{smptehdbars}
23155 @anchor{yuvtestsrc}
23156 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23158 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23160 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23162 The @code{color} source provides an uniformly colored input.
23164 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23165 @ref{haldclut} filter.
23167 The @code{nullsrc} source returns unprocessed video frames. It is
23168 mainly useful to be employed in analysis / debugging tools, or as the
23169 source for filters which ignore the input data.
23171 The @code{pal75bars} source generates a color bars pattern, based on
23172 EBU PAL recommendations with 75% color levels.
23174 The @code{pal100bars} source generates a color bars pattern, based on
23175 EBU PAL recommendations with 100% color levels.
23177 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23178 detecting RGB vs BGR issues. You should see a red, green and blue
23179 stripe from top to bottom.
23181 The @code{smptebars} source generates a color bars pattern, based on
23182 the SMPTE Engineering Guideline EG 1-1990.
23184 The @code{smptehdbars} source generates a color bars pattern, based on
23185 the SMPTE RP 219-2002.
23187 The @code{testsrc} source generates a test video pattern, showing a
23188 color pattern, a scrolling gradient and a timestamp. This is mainly
23189 intended for testing purposes.
23191 The @code{testsrc2} source is similar to testsrc, but supports more
23192 pixel formats instead of just @code{rgb24}. This allows using it as an
23193 input for other tests without requiring a format conversion.
23195 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23196 see a y, cb and cr stripe from top to bottom.
23198 The sources accept the following parameters:
23203 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23204 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23205 pixels to be used as identity matrix for 3D lookup tables. Each component is
23206 coded on a @code{1/(N*N)} scale.
23209 Specify the color of the source, only available in the @code{color}
23210 source. For the syntax of this option, check the
23211 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23214 Specify the size of the sourced video. For the syntax of this option, check the
23215 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23216 The default value is @code{320x240}.
23218 This option is not available with the @code{allrgb}, @code{allyuv}, and
23219 @code{haldclutsrc} filters.
23222 Specify the frame rate of the sourced video, as the number of frames
23223 generated per second. It has to be a string in the format
23224 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23225 number or a valid video frame rate abbreviation. The default value is
23229 Set the duration of the sourced video. See
23230 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23231 for the accepted syntax.
23233 If not specified, or the expressed duration is negative, the video is
23234 supposed to be generated forever.
23236 Since the frame rate is used as time base, all frames including the last one
23237 will have their full duration. If the specified duration is not a multiple
23238 of the frame duration, it will be rounded up.
23241 Set the sample aspect ratio of the sourced video.
23244 Specify the alpha (opacity) of the background, only available in the
23245 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23246 255 (fully opaque, the default).
23249 Set the number of decimals to show in the timestamp, only available in the
23250 @code{testsrc} source.
23252 The displayed timestamp value will correspond to the original
23253 timestamp value multiplied by the power of 10 of the specified
23254 value. Default value is 0.
23257 @subsection Examples
23261 Generate a video with a duration of 5.3 seconds, with size
23262 176x144 and a frame rate of 10 frames per second:
23264 testsrc=duration=5.3:size=qcif:rate=10
23268 The following graph description will generate a red source
23269 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23272 color=c=red@@0.2:s=qcif:r=10
23276 If the input content is to be ignored, @code{nullsrc} can be used. The
23277 following command generates noise in the luminance plane by employing
23278 the @code{geq} filter:
23280 nullsrc=s=256x256, geq=random(1)*255:128:128
23284 @subsection Commands
23286 The @code{color} source supports the following commands:
23290 Set the color of the created image. Accepts the same syntax of the
23291 corresponding @option{color} option.
23296 Generate video using an OpenCL program.
23301 OpenCL program source file.
23304 Kernel name in program.
23307 Size of frames to generate. This must be set.
23310 Pixel format to use for the generated frames. This must be set.
23313 Number of frames generated every second. Default value is '25'.
23317 For details of how the program loading works, see the @ref{program_opencl}
23324 Generate a colour ramp by setting pixel values from the position of the pixel
23325 in the output image. (Note that this will work with all pixel formats, but
23326 the generated output will not be the same.)
23328 __kernel void ramp(__write_only image2d_t dst,
23329 unsigned int index)
23331 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23334 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23336 write_imagef(dst, loc, val);
23341 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23343 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23344 unsigned int index)
23346 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23348 float4 value = 0.0f;
23349 int x = loc.x + index;
23350 int y = loc.y + index;
23351 while (x > 0 || y > 0) {
23352 if (x % 3 == 1 && y % 3 == 1) {
23360 write_imagef(dst, loc, value);
23366 @section sierpinski
23368 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23370 This source accepts the following options:
23374 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23375 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23378 Set frame rate, expressed as number of frames per second. Default
23382 Set seed which is used for random panning.
23385 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23388 Set fractal type, can be default @code{carpet} or @code{triangle}.
23391 @c man end VIDEO SOURCES
23393 @chapter Video Sinks
23394 @c man begin VIDEO SINKS
23396 Below is a description of the currently available video sinks.
23398 @section buffersink
23400 Buffer video frames, and make them available to the end of the filter
23403 This sink is mainly intended for programmatic use, in particular
23404 through the interface defined in @file{libavfilter/buffersink.h}
23405 or the options system.
23407 It accepts a pointer to an AVBufferSinkContext structure, which
23408 defines the incoming buffers' formats, to be passed as the opaque
23409 parameter to @code{avfilter_init_filter} for initialization.
23413 Null video sink: do absolutely nothing with the input video. It is
23414 mainly useful as a template and for use in analysis / debugging
23417 @c man end VIDEO SINKS
23419 @chapter Multimedia Filters
23420 @c man begin MULTIMEDIA FILTERS
23422 Below is a description of the currently available multimedia filters.
23426 Convert input audio to a video output, displaying the audio bit scope.
23428 The filter accepts the following options:
23432 Set frame rate, expressed as number of frames per second. Default
23436 Specify the video size for the output. For the syntax of this option, check the
23437 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23438 Default value is @code{1024x256}.
23441 Specify list of colors separated by space or by '|' which will be used to
23442 draw channels. Unrecognized or missing colors will be replaced
23446 @section adrawgraph
23447 Draw a graph using input audio metadata.
23449 See @ref{drawgraph}
23451 @section agraphmonitor
23453 See @ref{graphmonitor}.
23455 @section ahistogram
23457 Convert input audio to a video output, displaying the volume histogram.
23459 The filter accepts the following options:
23463 Specify how histogram is calculated.
23465 It accepts the following values:
23468 Use single histogram for all channels.
23470 Use separate histogram for each channel.
23472 Default is @code{single}.
23475 Set frame rate, expressed as number of frames per second. Default
23479 Specify the video size for the output. For the syntax of this option, check the
23480 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23481 Default value is @code{hd720}.
23486 It accepts the following values:
23497 reverse logarithmic
23499 Default is @code{log}.
23502 Set amplitude scale.
23504 It accepts the following values:
23511 Default is @code{log}.
23514 Set how much frames to accumulate in histogram.
23515 Default is 1. Setting this to -1 accumulates all frames.
23518 Set histogram ratio of window height.
23521 Set sonogram sliding.
23523 It accepts the following values:
23526 replace old rows with new ones.
23528 scroll from top to bottom.
23530 Default is @code{replace}.
23533 @section aphasemeter
23535 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23536 representing mean phase of current audio frame. A video output can also be produced and is
23537 enabled by default. The audio is passed through as first output.
23539 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23540 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23541 and @code{1} means channels are in phase.
23543 The filter accepts the following options, all related to its video output:
23547 Set the output frame rate. Default value is @code{25}.
23550 Set the video size for the output. For the syntax of this option, check the
23551 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23552 Default value is @code{800x400}.
23557 Specify the red, green, blue contrast. Default values are @code{2},
23558 @code{7} and @code{1}.
23559 Allowed range is @code{[0, 255]}.
23562 Set color which will be used for drawing median phase. If color is
23563 @code{none} which is default, no median phase value will be drawn.
23566 Enable video output. Default is enabled.
23569 @subsection phasing detection
23571 The filter also detects out of phase and mono sequences in stereo streams.
23572 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
23574 The filter accepts the following options for this detection:
23578 Enable mono and out of phase detection. Default is disabled.
23581 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
23582 Allowed range is @code{[0, 1]}.
23585 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
23586 Allowed range is @code{[90, 180]}.
23589 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
23592 @subsection Examples
23596 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
23598 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
23602 @section avectorscope
23604 Convert input audio to a video output, representing the audio vector
23607 The filter is used to measure the difference between channels of stereo
23608 audio stream. A monaural signal, consisting of identical left and right
23609 signal, results in straight vertical line. Any stereo separation is visible
23610 as a deviation from this line, creating a Lissajous figure.
23611 If the straight (or deviation from it) but horizontal line appears this
23612 indicates that the left and right channels are out of phase.
23614 The filter accepts the following options:
23618 Set the vectorscope mode.
23620 Available values are:
23623 Lissajous rotated by 45 degrees.
23626 Same as above but not rotated.
23629 Shape resembling half of circle.
23632 Default value is @samp{lissajous}.
23635 Set the video size for the output. For the syntax of this option, check the
23636 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23637 Default value is @code{400x400}.
23640 Set the output frame rate. Default value is @code{25}.
23646 Specify the red, green, blue and alpha contrast. Default values are @code{40},
23647 @code{160}, @code{80} and @code{255}.
23648 Allowed range is @code{[0, 255]}.
23654 Specify the red, green, blue and alpha fade. Default values are @code{15},
23655 @code{10}, @code{5} and @code{5}.
23656 Allowed range is @code{[0, 255]}.
23659 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
23660 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
23663 Set the vectorscope drawing mode.
23665 Available values are:
23668 Draw dot for each sample.
23671 Draw line between previous and current sample.
23674 Default value is @samp{dot}.
23677 Specify amplitude scale of audio samples.
23679 Available values are:
23695 Swap left channel axis with right channel axis.
23705 Mirror only x axis.
23708 Mirror only y axis.
23716 @subsection Examples
23720 Complete example using @command{ffplay}:
23722 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
23723 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
23727 @section bench, abench
23729 Benchmark part of a filtergraph.
23731 The filter accepts the following options:
23735 Start or stop a timer.
23737 Available values are:
23740 Get the current time, set it as frame metadata (using the key
23741 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
23744 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
23745 the input frame metadata to get the time difference. Time difference, average,
23746 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
23747 @code{min}) are then printed. The timestamps are expressed in seconds.
23751 @subsection Examples
23755 Benchmark @ref{selectivecolor} filter:
23757 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
23763 Concatenate audio and video streams, joining them together one after the
23766 The filter works on segments of synchronized video and audio streams. All
23767 segments must have the same number of streams of each type, and that will
23768 also be the number of streams at output.
23770 The filter accepts the following options:
23775 Set the number of segments. Default is 2.
23778 Set the number of output video streams, that is also the number of video
23779 streams in each segment. Default is 1.
23782 Set the number of output audio streams, that is also the number of audio
23783 streams in each segment. Default is 0.
23786 Activate unsafe mode: do not fail if segments have a different format.
23790 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
23791 @var{a} audio outputs.
23793 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
23794 segment, in the same order as the outputs, then the inputs for the second
23797 Related streams do not always have exactly the same duration, for various
23798 reasons including codec frame size or sloppy authoring. For that reason,
23799 related synchronized streams (e.g. a video and its audio track) should be
23800 concatenated at once. The concat filter will use the duration of the longest
23801 stream in each segment (except the last one), and if necessary pad shorter
23802 audio streams with silence.
23804 For this filter to work correctly, all segments must start at timestamp 0.
23806 All corresponding streams must have the same parameters in all segments; the
23807 filtering system will automatically select a common pixel format for video
23808 streams, and a common sample format, sample rate and channel layout for
23809 audio streams, but other settings, such as resolution, must be converted
23810 explicitly by the user.
23812 Different frame rates are acceptable but will result in variable frame rate
23813 at output; be sure to configure the output file to handle it.
23815 @subsection Examples
23819 Concatenate an opening, an episode and an ending, all in bilingual version
23820 (video in stream 0, audio in streams 1 and 2):
23822 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
23823 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
23824 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
23825 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
23829 Concatenate two parts, handling audio and video separately, using the
23830 (a)movie sources, and adjusting the resolution:
23832 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
23833 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
23834 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
23836 Note that a desync will happen at the stitch if the audio and video streams
23837 do not have exactly the same duration in the first file.
23841 @subsection Commands
23843 This filter supports the following commands:
23846 Close the current segment and step to the next one
23852 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
23853 level. By default, it logs a message at a frequency of 10Hz with the
23854 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
23855 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
23857 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
23858 sample format is double-precision floating point. The input stream will be converted to
23859 this specification, if needed. Users may need to insert aformat and/or aresample filters
23860 after this filter to obtain the original parameters.
23862 The filter also has a video output (see the @var{video} option) with a real
23863 time graph to observe the loudness evolution. The graphic contains the logged
23864 message mentioned above, so it is not printed anymore when this option is set,
23865 unless the verbose logging is set. The main graphing area contains the
23866 short-term loudness (3 seconds of analysis), and the gauge on the right is for
23867 the momentary loudness (400 milliseconds), but can optionally be configured
23868 to instead display short-term loudness (see @var{gauge}).
23870 The green area marks a +/- 1LU target range around the target loudness
23871 (-23LUFS by default, unless modified through @var{target}).
23873 More information about the Loudness Recommendation EBU R128 on
23874 @url{http://tech.ebu.ch/loudness}.
23876 The filter accepts the following options:
23881 Activate the video output. The audio stream is passed unchanged whether this
23882 option is set or no. The video stream will be the first output stream if
23883 activated. Default is @code{0}.
23886 Set the video size. This option is for video only. For the syntax of this
23888 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23889 Default and minimum resolution is @code{640x480}.
23892 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
23893 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
23894 other integer value between this range is allowed.
23897 Set metadata injection. If set to @code{1}, the audio input will be segmented
23898 into 100ms output frames, each of them containing various loudness information
23899 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
23901 Default is @code{0}.
23904 Force the frame logging level.
23906 Available values are:
23909 information logging level
23911 verbose logging level
23914 By default, the logging level is set to @var{info}. If the @option{video} or
23915 the @option{metadata} options are set, it switches to @var{verbose}.
23920 Available modes can be cumulated (the option is a @code{flag} type). Possible
23924 Disable any peak mode (default).
23926 Enable sample-peak mode.
23928 Simple peak mode looking for the higher sample value. It logs a message
23929 for sample-peak (identified by @code{SPK}).
23931 Enable true-peak mode.
23933 If enabled, the peak lookup is done on an over-sampled version of the input
23934 stream for better peak accuracy. It logs a message for true-peak.
23935 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
23936 This mode requires a build with @code{libswresample}.
23940 Treat mono input files as "dual mono". If a mono file is intended for playback
23941 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
23942 If set to @code{true}, this option will compensate for this effect.
23943 Multi-channel input files are not affected by this option.
23946 Set a specific pan law to be used for the measurement of dual mono files.
23947 This parameter is optional, and has a default value of -3.01dB.
23950 Set a specific target level (in LUFS) used as relative zero in the visualization.
23951 This parameter is optional and has a default value of -23LUFS as specified
23952 by EBU R128. However, material published online may prefer a level of -16LUFS
23953 (e.g. for use with podcasts or video platforms).
23956 Set the value displayed by the gauge. Valid values are @code{momentary} and s
23957 @code{shortterm}. By default the momentary value will be used, but in certain
23958 scenarios it may be more useful to observe the short term value instead (e.g.
23962 Sets the display scale for the loudness. Valid parameters are @code{absolute}
23963 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
23964 video output, not the summary or continuous log output.
23967 @subsection Examples
23971 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
23973 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
23977 Run an analysis with @command{ffmpeg}:
23979 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
23983 @section interleave, ainterleave
23985 Temporally interleave frames from several inputs.
23987 @code{interleave} works with video inputs, @code{ainterleave} with audio.
23989 These filters read frames from several inputs and send the oldest
23990 queued frame to the output.
23992 Input streams must have well defined, monotonically increasing frame
23995 In order to submit one frame to output, these filters need to enqueue
23996 at least one frame for each input, so they cannot work in case one
23997 input is not yet terminated and will not receive incoming frames.
23999 For example consider the case when one input is a @code{select} filter
24000 which always drops input frames. The @code{interleave} filter will keep
24001 reading from that input, but it will never be able to send new frames
24002 to output until the input sends an end-of-stream signal.
24004 Also, depending on inputs synchronization, the filters will drop
24005 frames in case one input receives more frames than the other ones, and
24006 the queue is already filled.
24008 These filters accept the following options:
24012 Set the number of different inputs, it is 2 by default.
24015 How to determine the end-of-stream.
24019 The duration of the longest input. (default)
24022 The duration of the shortest input.
24025 The duration of the first input.
24030 @subsection Examples
24034 Interleave frames belonging to different streams using @command{ffmpeg}:
24036 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24040 Add flickering blur effect:
24042 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24046 @section metadata, ametadata
24048 Manipulate frame metadata.
24050 This filter accepts the following options:
24054 Set mode of operation of the filter.
24056 Can be one of the following:
24060 If both @code{value} and @code{key} is set, select frames
24061 which have such metadata. If only @code{key} is set, select
24062 every frame that has such key in metadata.
24065 Add new metadata @code{key} and @code{value}. If key is already available
24069 Modify value of already present key.
24072 If @code{value} is set, delete only keys that have such value.
24073 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24077 Print key and its value if metadata was found. If @code{key} is not set print all
24078 metadata values available in frame.
24082 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24085 Set metadata value which will be used. This option is mandatory for
24086 @code{modify} and @code{add} mode.
24089 Which function to use when comparing metadata value and @code{value}.
24091 Can be one of following:
24095 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24098 Values are interpreted as strings, returns true if metadata value starts with
24099 the @code{value} option string.
24102 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24105 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24108 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24111 Values are interpreted as floats, returns true if expression from option @code{expr}
24115 Values are interpreted as strings, returns true if metadata value ends with
24116 the @code{value} option string.
24120 Set expression which is used when @code{function} is set to @code{expr}.
24121 The expression is evaluated through the eval API and can contain the following
24126 Float representation of @code{value} from metadata key.
24129 Float representation of @code{value} as supplied by user in @code{value} option.
24133 If specified in @code{print} mode, output is written to the named file. Instead of
24134 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24135 for standard output. If @code{file} option is not set, output is written to the log
24136 with AV_LOG_INFO loglevel.
24139 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24143 @subsection Examples
24147 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24150 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24153 Print silencedetect output to file @file{metadata.txt}.
24155 silencedetect,ametadata=mode=print:file=metadata.txt
24158 Direct all metadata to a pipe with file descriptor 4.
24160 metadata=mode=print:file='pipe\:4'
24164 @section perms, aperms
24166 Set read/write permissions for the output frames.
24168 These filters are mainly aimed at developers to test direct path in the
24169 following filter in the filtergraph.
24171 The filters accept the following options:
24175 Select the permissions mode.
24177 It accepts the following values:
24180 Do nothing. This is the default.
24182 Set all the output frames read-only.
24184 Set all the output frames directly writable.
24186 Make the frame read-only if writable, and writable if read-only.
24188 Set each output frame read-only or writable randomly.
24192 Set the seed for the @var{random} mode, must be an integer included between
24193 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24194 @code{-1}, the filter will try to use a good random seed on a best effort
24198 Note: in case of auto-inserted filter between the permission filter and the
24199 following one, the permission might not be received as expected in that
24200 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24201 perms/aperms filter can avoid this problem.
24203 @section realtime, arealtime
24205 Slow down filtering to match real time approximately.
24207 These filters will pause the filtering for a variable amount of time to
24208 match the output rate with the input timestamps.
24209 They are similar to the @option{re} option to @code{ffmpeg}.
24211 They accept the following options:
24215 Time limit for the pauses. Any pause longer than that will be considered
24216 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24218 Speed factor for processing. The value must be a float larger than zero.
24219 Values larger than 1.0 will result in faster than realtime processing,
24220 smaller will slow processing down. The @var{limit} is automatically adapted
24221 accordingly. Default is 1.0.
24223 A processing speed faster than what is possible without these filters cannot
24228 @section select, aselect
24230 Select frames to pass in output.
24232 This filter accepts the following options:
24237 Set expression, which is evaluated for each input frame.
24239 If the expression is evaluated to zero, the frame is discarded.
24241 If the evaluation result is negative or NaN, the frame is sent to the
24242 first output; otherwise it is sent to the output with index
24243 @code{ceil(val)-1}, assuming that the input index starts from 0.
24245 For example a value of @code{1.2} corresponds to the output with index
24246 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24249 Set the number of outputs. The output to which to send the selected
24250 frame is based on the result of the evaluation. Default value is 1.
24253 The expression can contain the following constants:
24257 The (sequential) number of the filtered frame, starting from 0.
24260 The (sequential) number of the selected frame, starting from 0.
24262 @item prev_selected_n
24263 The sequential number of the last selected frame. It's NAN if undefined.
24266 The timebase of the input timestamps.
24269 The PTS (Presentation TimeStamp) of the filtered video frame,
24270 expressed in @var{TB} units. It's NAN if undefined.
24273 The PTS of the filtered video frame,
24274 expressed in seconds. It's NAN if undefined.
24277 The PTS of the previously filtered video frame. It's NAN if undefined.
24279 @item prev_selected_pts
24280 The PTS of the last previously filtered video frame. It's NAN if undefined.
24282 @item prev_selected_t
24283 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24286 The PTS of the first video frame in the video. It's NAN if undefined.
24289 The time of the first video frame in the video. It's NAN if undefined.
24291 @item pict_type @emph{(video only)}
24292 The type of the filtered frame. It can assume one of the following
24304 @item interlace_type @emph{(video only)}
24305 The frame interlace type. It can assume one of the following values:
24308 The frame is progressive (not interlaced).
24310 The frame is top-field-first.
24312 The frame is bottom-field-first.
24315 @item consumed_sample_n @emph{(audio only)}
24316 the number of selected samples before the current frame
24318 @item samples_n @emph{(audio only)}
24319 the number of samples in the current frame
24321 @item sample_rate @emph{(audio only)}
24322 the input sample rate
24325 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24328 the position in the file of the filtered frame, -1 if the information
24329 is not available (e.g. for synthetic video)
24331 @item scene @emph{(video only)}
24332 value between 0 and 1 to indicate a new scene; a low value reflects a low
24333 probability for the current frame to introduce a new scene, while a higher
24334 value means the current frame is more likely to be one (see the example below)
24336 @item concatdec_select
24337 The concat demuxer can select only part of a concat input file by setting an
24338 inpoint and an outpoint, but the output packets may not be entirely contained
24339 in the selected interval. By using this variable, it is possible to skip frames
24340 generated by the concat demuxer which are not exactly contained in the selected
24343 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24344 and the @var{lavf.concat.duration} packet metadata values which are also
24345 present in the decoded frames.
24347 The @var{concatdec_select} variable is -1 if the frame pts is at least
24348 start_time and either the duration metadata is missing or the frame pts is less
24349 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24352 That basically means that an input frame is selected if its pts is within the
24353 interval set by the concat demuxer.
24357 The default value of the select expression is "1".
24359 @subsection Examples
24363 Select all frames in input:
24368 The example above is the same as:
24380 Select only I-frames:
24382 select='eq(pict_type\,I)'
24386 Select one frame every 100:
24388 select='not(mod(n\,100))'
24392 Select only frames contained in the 10-20 time interval:
24394 select=between(t\,10\,20)
24398 Select only I-frames contained in the 10-20 time interval:
24400 select=between(t\,10\,20)*eq(pict_type\,I)
24404 Select frames with a minimum distance of 10 seconds:
24406 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24410 Use aselect to select only audio frames with samples number > 100:
24412 aselect='gt(samples_n\,100)'
24416 Create a mosaic of the first scenes:
24418 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24421 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24425 Send even and odd frames to separate outputs, and compose them:
24427 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24431 Select useful frames from an ffconcat file which is using inpoints and
24432 outpoints but where the source files are not intra frame only.
24434 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24438 @section sendcmd, asendcmd
24440 Send commands to filters in the filtergraph.
24442 These filters read commands to be sent to other filters in the
24445 @code{sendcmd} must be inserted between two video filters,
24446 @code{asendcmd} must be inserted between two audio filters, but apart
24447 from that they act the same way.
24449 The specification of commands can be provided in the filter arguments
24450 with the @var{commands} option, or in a file specified by the
24451 @var{filename} option.
24453 These filters accept the following options:
24456 Set the commands to be read and sent to the other filters.
24458 Set the filename of the commands to be read and sent to the other
24462 @subsection Commands syntax
24464 A commands description consists of a sequence of interval
24465 specifications, comprising a list of commands to be executed when a
24466 particular event related to that interval occurs. The occurring event
24467 is typically the current frame time entering or leaving a given time
24470 An interval is specified by the following syntax:
24472 @var{START}[-@var{END}] @var{COMMANDS};
24475 The time interval is specified by the @var{START} and @var{END} times.
24476 @var{END} is optional and defaults to the maximum time.
24478 The current frame time is considered within the specified interval if
24479 it is included in the interval [@var{START}, @var{END}), that is when
24480 the time is greater or equal to @var{START} and is lesser than
24483 @var{COMMANDS} consists of a sequence of one or more command
24484 specifications, separated by ",", relating to that interval. The
24485 syntax of a command specification is given by:
24487 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24490 @var{FLAGS} is optional and specifies the type of events relating to
24491 the time interval which enable sending the specified command, and must
24492 be a non-null sequence of identifier flags separated by "+" or "|" and
24493 enclosed between "[" and "]".
24495 The following flags are recognized:
24498 The command is sent when the current frame timestamp enters the
24499 specified interval. In other words, the command is sent when the
24500 previous frame timestamp was not in the given interval, and the
24504 The command is sent when the current frame timestamp leaves the
24505 specified interval. In other words, the command is sent when the
24506 previous frame timestamp was in the given interval, and the
24510 The command @var{ARG} is interpreted as expression and result of
24511 expression is passed as @var{ARG}.
24513 The expression is evaluated through the eval API and can contain the following
24518 Original position in the file of the frame, or undefined if undefined
24519 for the current frame.
24522 The presentation timestamp in input.
24525 The count of the input frame for video or audio, starting from 0.
24528 The time in seconds of the current frame.
24531 The start time in seconds of the current command interval.
24534 The end time in seconds of the current command interval.
24537 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24542 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24545 @var{TARGET} specifies the target of the command, usually the name of
24546 the filter class or a specific filter instance name.
24548 @var{COMMAND} specifies the name of the command for the target filter.
24550 @var{ARG} is optional and specifies the optional list of argument for
24551 the given @var{COMMAND}.
24553 Between one interval specification and another, whitespaces, or
24554 sequences of characters starting with @code{#} until the end of line,
24555 are ignored and can be used to annotate comments.
24557 A simplified BNF description of the commands specification syntax
24560 @var{COMMAND_FLAG} ::= "enter" | "leave"
24561 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24562 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24563 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24564 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24565 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24568 @subsection Examples
24572 Specify audio tempo change at second 4:
24574 asendcmd=c='4.0 atempo tempo 1.5',atempo
24578 Target a specific filter instance:
24580 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
24584 Specify a list of drawtext and hue commands in a file.
24586 # show text in the interval 5-10
24587 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
24588 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
24590 # desaturate the image in the interval 15-20
24591 15.0-20.0 [enter] hue s 0,
24592 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
24594 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
24596 # apply an exponential saturation fade-out effect, starting from time 25
24597 25 [enter] hue s exp(25-t)
24600 A filtergraph allowing to read and process the above command list
24601 stored in a file @file{test.cmd}, can be specified with:
24603 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
24608 @section setpts, asetpts
24610 Change the PTS (presentation timestamp) of the input frames.
24612 @code{setpts} works on video frames, @code{asetpts} on audio frames.
24614 This filter accepts the following options:
24619 The expression which is evaluated for each frame to construct its timestamp.
24623 The expression is evaluated through the eval API and can contain the following
24627 @item FRAME_RATE, FR
24628 frame rate, only defined for constant frame-rate video
24631 The presentation timestamp in input
24634 The count of the input frame for video or the number of consumed samples,
24635 not including the current frame for audio, starting from 0.
24637 @item NB_CONSUMED_SAMPLES
24638 The number of consumed samples, not including the current frame (only
24641 @item NB_SAMPLES, S
24642 The number of samples in the current frame (only audio)
24644 @item SAMPLE_RATE, SR
24645 The audio sample rate.
24648 The PTS of the first frame.
24651 the time in seconds of the first frame
24654 State whether the current frame is interlaced.
24657 the time in seconds of the current frame
24660 original position in the file of the frame, or undefined if undefined
24661 for the current frame
24664 The previous input PTS.
24667 previous input time in seconds
24670 The previous output PTS.
24673 previous output time in seconds
24676 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
24680 The wallclock (RTC) time at the start of the movie in microseconds.
24683 The timebase of the input timestamps.
24687 @subsection Examples
24691 Start counting PTS from zero
24693 setpts=PTS-STARTPTS
24697 Apply fast motion effect:
24703 Apply slow motion effect:
24709 Set fixed rate of 25 frames per second:
24715 Set fixed rate 25 fps with some jitter:
24717 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
24721 Apply an offset of 10 seconds to the input PTS:
24727 Generate timestamps from a "live source" and rebase onto the current timebase:
24729 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
24733 Generate timestamps by counting samples:
24742 Force color range for the output video frame.
24744 The @code{setrange} filter marks the color range property for the
24745 output frames. It does not change the input frame, but only sets the
24746 corresponding property, which affects how the frame is treated by
24749 The filter accepts the following options:
24754 Available values are:
24758 Keep the same color range property.
24760 @item unspecified, unknown
24761 Set the color range as unspecified.
24763 @item limited, tv, mpeg
24764 Set the color range as limited.
24766 @item full, pc, jpeg
24767 Set the color range as full.
24771 @section settb, asettb
24773 Set the timebase to use for the output frames timestamps.
24774 It is mainly useful for testing timebase configuration.
24776 It accepts the following parameters:
24781 The expression which is evaluated into the output timebase.
24785 The value for @option{tb} is an arithmetic expression representing a
24786 rational. The expression can contain the constants "AVTB" (the default
24787 timebase), "intb" (the input timebase) and "sr" (the sample rate,
24788 audio only). Default value is "intb".
24790 @subsection Examples
24794 Set the timebase to 1/25:
24800 Set the timebase to 1/10:
24806 Set the timebase to 1001/1000:
24812 Set the timebase to 2*intb:
24818 Set the default timebase value:
24825 Convert input audio to a video output representing frequency spectrum
24826 logarithmically using Brown-Puckette constant Q transform algorithm with
24827 direct frequency domain coefficient calculation (but the transform itself
24828 is not really constant Q, instead the Q factor is actually variable/clamped),
24829 with musical tone scale, from E0 to D#10.
24831 The filter accepts the following options:
24835 Specify the video size for the output. It must be even. For the syntax of this option,
24836 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24837 Default value is @code{1920x1080}.
24840 Set the output frame rate. Default value is @code{25}.
24843 Set the bargraph height. It must be even. Default value is @code{-1} which
24844 computes the bargraph height automatically.
24847 Set the axis height. It must be even. Default value is @code{-1} which computes
24848 the axis height automatically.
24851 Set the sonogram height. It must be even. Default value is @code{-1} which
24852 computes the sonogram height automatically.
24855 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
24856 instead. Default value is @code{1}.
24858 @item sono_v, volume
24859 Specify the sonogram volume expression. It can contain variables:
24862 the @var{bar_v} evaluated expression
24863 @item frequency, freq, f
24864 the frequency where it is evaluated
24865 @item timeclamp, tc
24866 the value of @var{timeclamp} option
24870 @item a_weighting(f)
24871 A-weighting of equal loudness
24872 @item b_weighting(f)
24873 B-weighting of equal loudness
24874 @item c_weighting(f)
24875 C-weighting of equal loudness.
24877 Default value is @code{16}.
24879 @item bar_v, volume2
24880 Specify the bargraph volume expression. It can contain variables:
24883 the @var{sono_v} evaluated expression
24884 @item frequency, freq, f
24885 the frequency where it is evaluated
24886 @item timeclamp, tc
24887 the value of @var{timeclamp} option
24891 @item a_weighting(f)
24892 A-weighting of equal loudness
24893 @item b_weighting(f)
24894 B-weighting of equal loudness
24895 @item c_weighting(f)
24896 C-weighting of equal loudness.
24898 Default value is @code{sono_v}.
24900 @item sono_g, gamma
24901 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
24902 higher gamma makes the spectrum having more range. Default value is @code{3}.
24903 Acceptable range is @code{[1, 7]}.
24905 @item bar_g, gamma2
24906 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
24910 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
24911 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
24913 @item timeclamp, tc
24914 Specify the transform timeclamp. At low frequency, there is trade-off between
24915 accuracy in time domain and frequency domain. If timeclamp is lower,
24916 event in time domain is represented more accurately (such as fast bass drum),
24917 otherwise event in frequency domain is represented more accurately
24918 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
24921 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
24922 limits future samples by applying asymmetric windowing in time domain, useful
24923 when low latency is required. Accepted range is @code{[0, 1]}.
24926 Specify the transform base frequency. Default value is @code{20.01523126408007475},
24927 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
24930 Specify the transform end frequency. Default value is @code{20495.59681441799654},
24931 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
24934 This option is deprecated and ignored.
24937 Specify the transform length in time domain. Use this option to control accuracy
24938 trade-off between time domain and frequency domain at every frequency sample.
24939 It can contain variables:
24941 @item frequency, freq, f
24942 the frequency where it is evaluated
24943 @item timeclamp, tc
24944 the value of @var{timeclamp} option.
24946 Default value is @code{384*tc/(384+tc*f)}.
24949 Specify the transform count for every video frame. Default value is @code{6}.
24950 Acceptable range is @code{[1, 30]}.
24953 Specify the transform count for every single pixel. Default value is @code{0},
24954 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
24957 Specify font file for use with freetype to draw the axis. If not specified,
24958 use embedded font. Note that drawing with font file or embedded font is not
24959 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
24963 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
24964 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
24968 Specify font color expression. This is arithmetic expression that should return
24969 integer value 0xRRGGBB. It can contain variables:
24971 @item frequency, freq, f
24972 the frequency where it is evaluated
24973 @item timeclamp, tc
24974 the value of @var{timeclamp} option
24979 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
24980 @item r(x), g(x), b(x)
24981 red, green, and blue value of intensity x.
24983 Default value is @code{st(0, (midi(f)-59.5)/12);
24984 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
24985 r(1-ld(1)) + b(ld(1))}.
24988 Specify image file to draw the axis. This option override @var{fontfile} and
24989 @var{fontcolor} option.
24992 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
24993 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
24994 Default value is @code{1}.
24997 Set colorspace. The accepted values are:
25000 Unspecified (default)
25009 BT.470BG or BT.601-6 625
25012 SMPTE-170M or BT.601-6 525
25018 BT.2020 with non-constant luminance
25023 Set spectrogram color scheme. This is list of floating point values with format
25024 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25025 The default is @code{1|0.5|0|0|0.5|1}.
25029 @subsection Examples
25033 Playing audio while showing the spectrum:
25035 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25039 Same as above, but with frame rate 30 fps:
25041 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25045 Playing at 1280x720:
25047 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25051 Disable sonogram display:
25057 A1 and its harmonics: A1, A2, (near)E3, A3:
25059 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),
25060 asplit[a][out1]; [a] showcqt [out0]'
25064 Same as above, but with more accuracy in frequency domain:
25066 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),
25067 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25073 bar_v=10:sono_v=bar_v*a_weighting(f)
25077 Custom gamma, now spectrum is linear to the amplitude.
25083 Custom tlength equation:
25085 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)))'
25089 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25091 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25095 Custom font using fontconfig:
25097 font='Courier New,Monospace,mono|bold'
25101 Custom frequency range with custom axis using image file:
25103 axisfile=myaxis.png:basefreq=40:endfreq=10000
25109 Convert input audio to video output representing the audio power spectrum.
25110 Audio amplitude is on Y-axis while frequency is on X-axis.
25112 The filter accepts the following options:
25116 Specify size of video. For the syntax of this option, check the
25117 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25118 Default is @code{1024x512}.
25122 This set how each frequency bin will be represented.
25124 It accepts the following values:
25130 Default is @code{bar}.
25133 Set amplitude scale.
25135 It accepts the following values:
25149 Default is @code{log}.
25152 Set frequency scale.
25154 It accepts the following values:
25163 Reverse logarithmic scale.
25165 Default is @code{lin}.
25168 Set window size. Allowed range is from 16 to 65536.
25170 Default is @code{2048}
25173 Set windowing function.
25175 It accepts the following values:
25198 Default is @code{hanning}.
25201 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25202 which means optimal overlap for selected window function will be picked.
25205 Set time averaging. Setting this to 0 will display current maximal peaks.
25206 Default is @code{1}, which means time averaging is disabled.
25209 Specify list of colors separated by space or by '|' which will be used to
25210 draw channel frequencies. Unrecognized or missing colors will be replaced
25214 Set channel display mode.
25216 It accepts the following values:
25221 Default is @code{combined}.
25224 Set minimum amplitude used in @code{log} amplitude scaler.
25227 Set data display mode.
25229 It accepts the following values:
25235 Default is @code{magnitude}.
25238 @section showspatial
25240 Convert stereo input audio to a video output, representing the spatial relationship
25241 between two channels.
25243 The filter accepts the following options:
25247 Specify the video size for the output. For the syntax of this option, check the
25248 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25249 Default value is @code{512x512}.
25252 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25255 Set window function.
25257 It accepts the following values:
25282 Default value is @code{hann}.
25285 Set ratio of overlap window. Default value is @code{0.5}.
25286 When value is @code{1} overlap is set to recommended size for specific
25287 window function currently used.
25290 @anchor{showspectrum}
25291 @section showspectrum
25293 Convert input audio to a video output, representing the audio frequency
25296 The filter accepts the following options:
25300 Specify the video size for the output. For the syntax of this option, check the
25301 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25302 Default value is @code{640x512}.
25305 Specify how the spectrum should slide along the window.
25307 It accepts the following values:
25310 the samples start again on the left when they reach the right
25312 the samples scroll from right to left
25314 frames are only produced when the samples reach the right
25316 the samples scroll from left to right
25319 Default value is @code{replace}.
25322 Specify display mode.
25324 It accepts the following values:
25327 all channels are displayed in the same row
25329 all channels are displayed in separate rows
25332 Default value is @samp{combined}.
25335 Specify display color mode.
25337 It accepts the following values:
25340 each channel is displayed in a separate color
25342 each channel is displayed using the same color scheme
25344 each channel is displayed using the rainbow color scheme
25346 each channel is displayed using the moreland color scheme
25348 each channel is displayed using the nebulae color scheme
25350 each channel is displayed using the fire color scheme
25352 each channel is displayed using the fiery color scheme
25354 each channel is displayed using the fruit color scheme
25356 each channel is displayed using the cool color scheme
25358 each channel is displayed using the magma color scheme
25360 each channel is displayed using the green color scheme
25362 each channel is displayed using the viridis color scheme
25364 each channel is displayed using the plasma color scheme
25366 each channel is displayed using the cividis color scheme
25368 each channel is displayed using the terrain color scheme
25371 Default value is @samp{channel}.
25374 Specify scale used for calculating intensity color values.
25376 It accepts the following values:
25381 square root, default
25392 Default value is @samp{sqrt}.
25395 Specify frequency scale.
25397 It accepts the following values:
25405 Default value is @samp{lin}.
25408 Set saturation modifier for displayed colors. Negative values provide
25409 alternative color scheme. @code{0} is no saturation at all.
25410 Saturation must be in [-10.0, 10.0] range.
25411 Default value is @code{1}.
25414 Set window function.
25416 It accepts the following values:
25441 Default value is @code{hann}.
25444 Set orientation of time vs frequency axis. Can be @code{vertical} or
25445 @code{horizontal}. Default is @code{vertical}.
25448 Set ratio of overlap window. Default value is @code{0}.
25449 When value is @code{1} overlap is set to recommended size for specific
25450 window function currently used.
25453 Set scale gain for calculating intensity color values.
25454 Default value is @code{1}.
25457 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25460 Set color rotation, must be in [-1.0, 1.0] range.
25461 Default value is @code{0}.
25464 Set start frequency from which to display spectrogram. Default is @code{0}.
25467 Set stop frequency to which to display spectrogram. Default is @code{0}.
25470 Set upper frame rate limit. Default is @code{auto}, unlimited.
25473 Draw time and frequency axes and legends. Default is disabled.
25476 The usage is very similar to the showwaves filter; see the examples in that
25479 @subsection Examples
25483 Large window with logarithmic color scaling:
25485 showspectrum=s=1280x480:scale=log
25489 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25491 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25492 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25496 @section showspectrumpic
25498 Convert input audio to a single video frame, representing the audio frequency
25501 The filter accepts the following options:
25505 Specify the video size for the output. For the syntax of this option, check the
25506 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25507 Default value is @code{4096x2048}.
25510 Specify display mode.
25512 It accepts the following values:
25515 all channels are displayed in the same row
25517 all channels are displayed in separate rows
25519 Default value is @samp{combined}.
25522 Specify display color mode.
25524 It accepts the following values:
25527 each channel is displayed in a separate color
25529 each channel is displayed using the same color scheme
25531 each channel is displayed using the rainbow color scheme
25533 each channel is displayed using the moreland color scheme
25535 each channel is displayed using the nebulae color scheme
25537 each channel is displayed using the fire color scheme
25539 each channel is displayed using the fiery color scheme
25541 each channel is displayed using the fruit color scheme
25543 each channel is displayed using the cool color scheme
25545 each channel is displayed using the magma color scheme
25547 each channel is displayed using the green color scheme
25549 each channel is displayed using the viridis color scheme
25551 each channel is displayed using the plasma color scheme
25553 each channel is displayed using the cividis color scheme
25555 each channel is displayed using the terrain color scheme
25557 Default value is @samp{intensity}.
25560 Specify scale used for calculating intensity color values.
25562 It accepts the following values:
25567 square root, default
25577 Default value is @samp{log}.
25580 Specify frequency scale.
25582 It accepts the following values:
25590 Default value is @samp{lin}.
25593 Set saturation modifier for displayed colors. Negative values provide
25594 alternative color scheme. @code{0} is no saturation at all.
25595 Saturation must be in [-10.0, 10.0] range.
25596 Default value is @code{1}.
25599 Set window function.
25601 It accepts the following values:
25625 Default value is @code{hann}.
25628 Set orientation of time vs frequency axis. Can be @code{vertical} or
25629 @code{horizontal}. Default is @code{vertical}.
25632 Set scale gain for calculating intensity color values.
25633 Default value is @code{1}.
25636 Draw time and frequency axes and legends. Default is enabled.
25639 Set color rotation, must be in [-1.0, 1.0] range.
25640 Default value is @code{0}.
25643 Set start frequency from which to display spectrogram. Default is @code{0}.
25646 Set stop frequency to which to display spectrogram. Default is @code{0}.
25649 @subsection Examples
25653 Extract an audio spectrogram of a whole audio track
25654 in a 1024x1024 picture using @command{ffmpeg}:
25656 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
25660 @section showvolume
25662 Convert input audio volume to a video output.
25664 The filter accepts the following options:
25671 Set border width, allowed range is [0, 5]. Default is 1.
25674 Set channel width, allowed range is [80, 8192]. Default is 400.
25677 Set channel height, allowed range is [1, 900]. Default is 20.
25680 Set fade, allowed range is [0, 1]. Default is 0.95.
25683 Set volume color expression.
25685 The expression can use the following variables:
25689 Current max volume of channel in dB.
25695 Current channel number, starting from 0.
25699 If set, displays channel names. Default is enabled.
25702 If set, displays volume values. Default is enabled.
25705 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
25706 default is @code{h}.
25709 Set step size, allowed range is [0, 5]. Default is 0, which means
25713 Set background opacity, allowed range is [0, 1]. Default is 0.
25716 Set metering mode, can be peak: @code{p} or rms: @code{r},
25717 default is @code{p}.
25720 Set display scale, can be linear: @code{lin} or log: @code{log},
25721 default is @code{lin}.
25725 If set to > 0., display a line for the max level
25726 in the previous seconds.
25727 default is disabled: @code{0.}
25730 The color of the max line. Use when @code{dm} option is set to > 0.
25731 default is: @code{orange}
25736 Convert input audio to a video output, representing the samples waves.
25738 The filter accepts the following options:
25742 Specify the video size for the output. For the syntax of this option, check the
25743 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25744 Default value is @code{600x240}.
25749 Available values are:
25752 Draw a point for each sample.
25755 Draw a vertical line for each sample.
25758 Draw a point for each sample and a line between them.
25761 Draw a centered vertical line for each sample.
25764 Default value is @code{point}.
25767 Set the number of samples which are printed on the same column. A
25768 larger value will decrease the frame rate. Must be a positive
25769 integer. This option can be set only if the value for @var{rate}
25770 is not explicitly specified.
25773 Set the (approximate) output frame rate. This is done by setting the
25774 option @var{n}. Default value is "25".
25776 @item split_channels
25777 Set if channels should be drawn separately or overlap. Default value is 0.
25780 Set colors separated by '|' which are going to be used for drawing of each channel.
25783 Set amplitude scale.
25785 Available values are:
25803 Set the draw mode. This is mostly useful to set for high @var{n}.
25805 Available values are:
25808 Scale pixel values for each drawn sample.
25811 Draw every sample directly.
25814 Default value is @code{scale}.
25817 @subsection Examples
25821 Output the input file audio and the corresponding video representation
25824 amovie=a.mp3,asplit[out0],showwaves[out1]
25828 Create a synthetic signal and show it with showwaves, forcing a
25829 frame rate of 30 frames per second:
25831 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
25835 @section showwavespic
25837 Convert input audio to a single video frame, representing the samples waves.
25839 The filter accepts the following options:
25843 Specify the video size for the output. For the syntax of this option, check the
25844 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25845 Default value is @code{600x240}.
25847 @item split_channels
25848 Set if channels should be drawn separately or overlap. Default value is 0.
25851 Set colors separated by '|' which are going to be used for drawing of each channel.
25854 Set amplitude scale.
25856 Available values are:
25876 Available values are:
25879 Scale pixel values for each drawn sample.
25882 Draw every sample directly.
25885 Default value is @code{scale}.
25888 Set the filter mode.
25890 Available values are:
25893 Use average samples values for each drawn sample.
25896 Use peak samples values for each drawn sample.
25899 Default value is @code{average}.
25902 @subsection Examples
25906 Extract a channel split representation of the wave form of a whole audio track
25907 in a 1024x800 picture using @command{ffmpeg}:
25909 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
25913 @section sidedata, asidedata
25915 Delete frame side data, or select frames based on it.
25917 This filter accepts the following options:
25921 Set mode of operation of the filter.
25923 Can be one of the following:
25927 Select every frame with side data of @code{type}.
25930 Delete side data of @code{type}. If @code{type} is not set, delete all side
25936 Set side data type used with all modes. Must be set for @code{select} mode. For
25937 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
25938 in @file{libavutil/frame.h}. For example, to choose
25939 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
25943 @section spectrumsynth
25945 Synthesize audio from 2 input video spectrums, first input stream represents
25946 magnitude across time and second represents phase across time.
25947 The filter will transform from frequency domain as displayed in videos back
25948 to time domain as presented in audio output.
25950 This filter is primarily created for reversing processed @ref{showspectrum}
25951 filter outputs, but can synthesize sound from other spectrograms too.
25952 But in such case results are going to be poor if the phase data is not
25953 available, because in such cases phase data need to be recreated, usually
25954 it's just recreated from random noise.
25955 For best results use gray only output (@code{channel} color mode in
25956 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
25957 @code{lin} scale for phase video. To produce phase, for 2nd video, use
25958 @code{data} option. Inputs videos should generally use @code{fullframe}
25959 slide mode as that saves resources needed for decoding video.
25961 The filter accepts the following options:
25965 Specify sample rate of output audio, the sample rate of audio from which
25966 spectrum was generated may differ.
25969 Set number of channels represented in input video spectrums.
25972 Set scale which was used when generating magnitude input spectrum.
25973 Can be @code{lin} or @code{log}. Default is @code{log}.
25976 Set slide which was used when generating inputs spectrums.
25977 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
25978 Default is @code{fullframe}.
25981 Set window function used for resynthesis.
25984 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25985 which means optimal overlap for selected window function will be picked.
25988 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
25989 Default is @code{vertical}.
25992 @subsection Examples
25996 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
25997 then resynthesize videos back to audio with spectrumsynth:
25999 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
26000 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
26001 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26005 @section split, asplit
26007 Split input into several identical outputs.
26009 @code{asplit} works with audio input, @code{split} with video.
26011 The filter accepts a single parameter which specifies the number of outputs. If
26012 unspecified, it defaults to 2.
26014 @subsection Examples
26018 Create two separate outputs from the same input:
26020 [in] split [out0][out1]
26024 To create 3 or more outputs, you need to specify the number of
26027 [in] asplit=3 [out0][out1][out2]
26031 Create two separate outputs from the same input, one cropped and
26034 [in] split [splitout1][splitout2];
26035 [splitout1] crop=100:100:0:0 [cropout];
26036 [splitout2] pad=200:200:100:100 [padout];
26040 Create 5 copies of the input audio with @command{ffmpeg}:
26042 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26048 Receive commands sent through a libzmq client, and forward them to
26049 filters in the filtergraph.
26051 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26052 must be inserted between two video filters, @code{azmq} between two
26053 audio filters. Both are capable to send messages to any filter type.
26055 To enable these filters you need to install the libzmq library and
26056 headers and configure FFmpeg with @code{--enable-libzmq}.
26058 For more information about libzmq see:
26059 @url{http://www.zeromq.org/}
26061 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26062 receives messages sent through a network interface defined by the
26063 @option{bind_address} (or the abbreviation "@option{b}") option.
26064 Default value of this option is @file{tcp://localhost:5555}. You may
26065 want to alter this value to your needs, but do not forget to escape any
26066 ':' signs (see @ref{filtergraph escaping}).
26068 The received message must be in the form:
26070 @var{TARGET} @var{COMMAND} [@var{ARG}]
26073 @var{TARGET} specifies the target of the command, usually the name of
26074 the filter class or a specific filter instance name. The default
26075 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26076 but you can override this by using the @samp{filter_name@@id} syntax
26077 (see @ref{Filtergraph syntax}).
26079 @var{COMMAND} specifies the name of the command for the target filter.
26081 @var{ARG} is optional and specifies the optional argument list for the
26082 given @var{COMMAND}.
26084 Upon reception, the message is processed and the corresponding command
26085 is injected into the filtergraph. Depending on the result, the filter
26086 will send a reply to the client, adopting the format:
26088 @var{ERROR_CODE} @var{ERROR_REASON}
26092 @var{MESSAGE} is optional.
26094 @subsection Examples
26096 Look at @file{tools/zmqsend} for an example of a zmq client which can
26097 be used to send commands processed by these filters.
26099 Consider the following filtergraph generated by @command{ffplay}.
26100 In this example the last overlay filter has an instance name. All other
26101 filters will have default instance names.
26104 ffplay -dumpgraph 1 -f lavfi "
26105 color=s=100x100:c=red [l];
26106 color=s=100x100:c=blue [r];
26107 nullsrc=s=200x100, zmq [bg];
26108 [bg][l] overlay [bg+l];
26109 [bg+l][r] overlay@@my=x=100 "
26112 To change the color of the left side of the video, the following
26113 command can be used:
26115 echo Parsed_color_0 c yellow | tools/zmqsend
26118 To change the right side:
26120 echo Parsed_color_1 c pink | tools/zmqsend
26123 To change the position of the right side:
26125 echo overlay@@my x 150 | tools/zmqsend
26129 @c man end MULTIMEDIA FILTERS
26131 @chapter Multimedia Sources
26132 @c man begin MULTIMEDIA SOURCES
26134 Below is a description of the currently available multimedia sources.
26138 This is the same as @ref{movie} source, except it selects an audio
26144 Read audio and/or video stream(s) from a movie container.
26146 It accepts the following parameters:
26150 The name of the resource to read (not necessarily a file; it can also be a
26151 device or a stream accessed through some protocol).
26153 @item format_name, f
26154 Specifies the format assumed for the movie to read, and can be either
26155 the name of a container or an input device. If not specified, the
26156 format is guessed from @var{movie_name} or by probing.
26158 @item seek_point, sp
26159 Specifies the seek point in seconds. The frames will be output
26160 starting from this seek point. The parameter is evaluated with
26161 @code{av_strtod}, so the numerical value may be suffixed by an IS
26162 postfix. The default value is "0".
26165 Specifies the streams to read. Several streams can be specified,
26166 separated by "+". The source will then have as many outputs, in the
26167 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26168 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26169 respectively the default (best suited) video and audio stream. Default
26170 is "dv", or "da" if the filter is called as "amovie".
26172 @item stream_index, si
26173 Specifies the index of the video stream to read. If the value is -1,
26174 the most suitable video stream will be automatically selected. The default
26175 value is "-1". Deprecated. If the filter is called "amovie", it will select
26176 audio instead of video.
26179 Specifies how many times to read the stream in sequence.
26180 If the value is 0, the stream will be looped infinitely.
26181 Default value is "1".
26183 Note that when the movie is looped the source timestamps are not
26184 changed, so it will generate non monotonically increasing timestamps.
26186 @item discontinuity
26187 Specifies the time difference between frames above which the point is
26188 considered a timestamp discontinuity which is removed by adjusting the later
26192 It allows overlaying a second video on top of the main input of
26193 a filtergraph, as shown in this graph:
26195 input -----------> deltapts0 --> overlay --> output
26198 movie --> scale--> deltapts1 -------+
26200 @subsection Examples
26204 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26205 on top of the input labelled "in":
26207 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26208 [in] setpts=PTS-STARTPTS [main];
26209 [main][over] overlay=16:16 [out]
26213 Read from a video4linux2 device, and overlay it on top of the input
26216 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26217 [in] setpts=PTS-STARTPTS [main];
26218 [main][over] overlay=16:16 [out]
26222 Read the first video stream and the audio stream with id 0x81 from
26223 dvd.vob; the video is connected to the pad named "video" and the audio is
26224 connected to the pad named "audio":
26226 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26230 @subsection Commands
26232 Both movie and amovie support the following commands:
26235 Perform seek using "av_seek_frame".
26236 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26239 @var{stream_index}: If stream_index is -1, a default
26240 stream is selected, and @var{timestamp} is automatically converted
26241 from AV_TIME_BASE units to the stream specific time_base.
26243 @var{timestamp}: Timestamp in AVStream.time_base units
26244 or, if no stream is specified, in AV_TIME_BASE units.
26246 @var{flags}: Flags which select direction and seeking mode.
26250 Get movie duration in AV_TIME_BASE units.
26254 @c man end MULTIMEDIA SOURCES