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, can be @var{2nd}, @var{4th} or @var{8th}.
527 Default is @var{4th}.
532 Reduce audio bit resolution.
534 This filter is bit crusher with enhanced functionality. A bit crusher
535 is used to audibly reduce number of bits an audio signal is sampled
536 with. This doesn't change the bit depth at all, it just produces the
537 effect. Material reduced in bit depth sounds more harsh and "digital".
538 This filter is able to even round to continuous values instead of discrete
540 Additionally it has a D/C offset which results in different crushing of
541 the lower and the upper half of the signal.
542 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
544 Another feature of this filter is the logarithmic mode.
545 This setting switches from linear distances between bits to logarithmic ones.
546 The result is a much more "natural" sounding crusher which doesn't gate low
547 signals for example. The human ear has a logarithmic perception,
548 so this kind of crushing is much more pleasant.
549 Logarithmic crushing is also able to get anti-aliased.
551 The filter accepts the following options:
567 Can be linear: @code{lin} or logarithmic: @code{log}.
576 Set sample reduction.
579 Enable LFO. By default disabled.
590 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
594 Remove impulsive noise from input audio.
596 Samples detected as impulsive noise are replaced by interpolated samples using
597 autoregressive modelling.
601 Set window size, in milliseconds. Allowed range is from @code{10} to
602 @code{100}. Default value is @code{55} milliseconds.
603 This sets size of window which will be processed at once.
606 Set window overlap, in percentage of window size. Allowed range is from
607 @code{50} to @code{95}. Default value is @code{75} percent.
608 Setting this to a very high value increases impulsive noise removal but makes
609 whole process much slower.
612 Set autoregression order, in percentage of window size. Allowed range is from
613 @code{0} to @code{25}. Default value is @code{2} percent. This option also
614 controls quality of interpolated samples using neighbour good samples.
617 Set threshold value. Allowed range is from @code{1} to @code{100}.
618 Default value is @code{2}.
619 This controls the strength of impulsive noise which is going to be removed.
620 The lower value, the more samples will be detected as impulsive noise.
623 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
624 @code{10}. Default value is @code{2}.
625 If any two samples detected as noise are spaced less than this value then any
626 sample between those two samples will be also detected as noise.
631 It accepts the following values:
634 Select overlap-add method. Even not interpolated samples are slightly
635 changed with this method.
638 Select overlap-save method. Not interpolated samples remain unchanged.
641 Default value is @code{a}.
645 Remove clipped samples from input audio.
647 Samples detected as clipped are replaced by interpolated samples using
648 autoregressive modelling.
652 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
653 Default value is @code{55} milliseconds.
654 This sets size of window which will be processed at once.
657 Set window overlap, in percentage of window size. Allowed range is from @code{50}
658 to @code{95}. Default value is @code{75} percent.
661 Set autoregression order, in percentage of window size. Allowed range is from
662 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
663 quality of interpolated samples using neighbour good samples.
666 Set threshold value. Allowed range is from @code{1} to @code{100}.
667 Default value is @code{10}. Higher values make clip detection less aggressive.
670 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
671 Default value is @code{1000}. Higher values make clip detection less aggressive.
676 It accepts the following values:
679 Select overlap-add method. Even not interpolated samples are slightly changed
683 Select overlap-save method. Not interpolated samples remain unchanged.
686 Default value is @code{a}.
691 Delay one or more audio channels.
693 Samples in delayed channel are filled with silence.
695 The filter accepts the following option:
699 Set list of delays in milliseconds for each channel separated by '|'.
700 Unused delays will be silently ignored. If number of given delays is
701 smaller than number of channels all remaining channels will not be delayed.
702 If you want to delay exact number of samples, append 'S' to number.
703 If you want instead to delay in seconds, append 's' to number.
706 Use last set delay for all remaining channels. By default is disabled.
707 This option if enabled changes how option @code{delays} is interpreted.
714 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
715 the second channel (and any other channels that may be present) unchanged.
721 Delay second channel by 500 samples, the third channel by 700 samples and leave
722 the first channel (and any other channels that may be present) unchanged.
728 Delay all channels by same number of samples:
730 adelay=delays=64S:all=1
734 @section aderivative, aintegral
736 Compute derivative/integral of audio stream.
738 Applying both filters one after another produces original audio.
742 Apply echoing to the input audio.
744 Echoes are reflected sound and can occur naturally amongst mountains
745 (and sometimes large buildings) when talking or shouting; digital echo
746 effects emulate this behaviour and are often used to help fill out the
747 sound of a single instrument or vocal. The time difference between the
748 original signal and the reflection is the @code{delay}, and the
749 loudness of the reflected signal is the @code{decay}.
750 Multiple echoes can have different delays and decays.
752 A description of the accepted parameters follows.
756 Set input gain of reflected signal. Default is @code{0.6}.
759 Set output gain of reflected signal. Default is @code{0.3}.
762 Set list of time intervals in milliseconds between original signal and reflections
763 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
764 Default is @code{1000}.
767 Set list of loudness of reflected signals separated by '|'.
768 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
769 Default is @code{0.5}.
776 Make it sound as if there are twice as many instruments as are actually playing:
778 aecho=0.8:0.88:60:0.4
782 If delay is very short, then it sounds like a (metallic) robot playing music:
788 A longer delay will sound like an open air concert in the mountains:
790 aecho=0.8:0.9:1000:0.3
794 Same as above but with one more mountain:
796 aecho=0.8:0.9:1000|1800:0.3|0.25
801 Audio emphasis filter creates or restores material directly taken from LPs or
802 emphased CDs with different filter curves. E.g. to store music on vinyl the
803 signal has to be altered by a filter first to even out the disadvantages of
804 this recording medium.
805 Once the material is played back the inverse filter has to be applied to
806 restore the distortion of the frequency response.
808 The filter accepts the following options:
818 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
819 use @code{production} mode. Default is @code{reproduction} mode.
822 Set filter type. Selects medium. Can be one of the following:
834 select Compact Disc (CD).
840 select 50µs (FM-KF).
842 select 75µs (FM-KF).
848 Modify an audio signal according to the specified expressions.
850 This filter accepts one or more expressions (one for each channel),
851 which are evaluated and used to modify a corresponding audio signal.
853 It accepts the following parameters:
857 Set the '|'-separated expressions list for each separate channel. If
858 the number of input channels is greater than the number of
859 expressions, the last specified expression is used for the remaining
862 @item channel_layout, c
863 Set output channel layout. If not specified, the channel layout is
864 specified by the number of expressions. If set to @samp{same}, it will
865 use by default the same input channel layout.
868 Each expression in @var{exprs} can contain the following constants and functions:
872 channel number of the current expression
875 number of the evaluated sample, starting from 0
881 time of the evaluated sample expressed in seconds
884 @item nb_out_channels
885 input and output number of channels
888 the value of input channel with number @var{CH}
891 Note: this filter is slow. For faster processing you should use a
900 aeval=val(ch)/2:c=same
904 Invert phase of the second channel:
913 Apply fade-in/out effect to input audio.
915 A description of the accepted parameters follows.
919 Specify the effect type, can be either @code{in} for fade-in, or
920 @code{out} for a fade-out effect. Default is @code{in}.
922 @item start_sample, ss
923 Specify the number of the start sample for starting to apply the fade
924 effect. Default is 0.
927 Specify the number of samples for which the fade effect has to last. At
928 the end of the fade-in effect the output audio will have the same
929 volume as the input audio, at the end of the fade-out transition
930 the output audio will be silence. Default is 44100.
933 Specify the start time of the fade effect. Default is 0.
934 The value must be specified as a time duration; see
935 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
936 for the accepted syntax.
937 If set this option is used instead of @var{start_sample}.
940 Specify the duration of the fade effect. See
941 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
942 for the accepted syntax.
943 At the end of the fade-in effect the output audio will have the same
944 volume as the input audio, at the end of the fade-out transition
945 the output audio will be silence.
946 By default the duration is determined by @var{nb_samples}.
947 If set this option is used instead of @var{nb_samples}.
950 Set curve for fade transition.
952 It accepts the following values:
955 select triangular, linear slope (default)
957 select quarter of sine wave
959 select half of sine wave
961 select exponential sine wave
965 select inverted parabola
979 select inverted quarter of sine wave
981 select inverted half of sine wave
983 select double-exponential seat
985 select double-exponential sigmoid
987 select logistic sigmoid
997 Fade in first 15 seconds of audio:
1003 Fade out last 25 seconds of a 900 seconds audio:
1005 afade=t=out:st=875:d=25
1010 Denoise audio samples with FFT.
1012 A description of the accepted parameters follows.
1016 Set the noise reduction in dB, allowed range is 0.01 to 97.
1017 Default value is 12 dB.
1020 Set the noise floor in dB, allowed range is -80 to -20.
1021 Default value is -50 dB.
1026 It accepts the following values:
1035 Select shellac noise.
1038 Select custom noise, defined in @code{bn} option.
1040 Default value is white noise.
1044 Set custom band noise for every one of 15 bands.
1045 Bands are separated by ' ' or '|'.
1048 Set the residual floor in dB, allowed range is -80 to -20.
1049 Default value is -38 dB.
1052 Enable noise tracking. By default is disabled.
1053 With this enabled, noise floor is automatically adjusted.
1056 Enable residual tracking. By default is disabled.
1059 Set the output mode.
1061 It accepts the following values:
1064 Pass input unchanged.
1067 Pass noise filtered out.
1072 Default value is @var{o}.
1076 @subsection Commands
1078 This filter supports the following commands:
1080 @item sample_noise, sn
1081 Start or stop measuring noise profile.
1082 Syntax for the command is : "start" or "stop" string.
1083 After measuring noise profile is stopped it will be
1084 automatically applied in filtering.
1086 @item noise_reduction, nr
1087 Change noise reduction. Argument is single float number.
1088 Syntax for the command is : "@var{noise_reduction}"
1090 @item noise_floor, nf
1091 Change noise floor. Argument is single float number.
1092 Syntax for the command is : "@var{noise_floor}"
1094 @item output_mode, om
1095 Change output mode operation.
1096 Syntax for the command is : "i", "o" or "n" string.
1100 Apply arbitrary expressions to samples in frequency domain.
1104 Set frequency domain real expression for each separate channel separated
1105 by '|'. Default is "re".
1106 If the number of input channels is greater than the number of
1107 expressions, the last specified expression is used for the remaining
1111 Set frequency domain imaginary expression for each separate channel
1112 separated by '|'. Default is "im".
1114 Each expression in @var{real} and @var{imag} can contain the following
1115 constants and functions:
1122 current frequency bin number
1125 number of available bins
1128 channel number of the current expression
1137 current real part of frequency bin of current channel
1140 current imaginary part of frequency bin of current channel
1143 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1146 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1150 Set window size. Allowed range is from 16 to 131072.
1151 Default is @code{4096}
1154 Set window function. Default is @code{hann}.
1157 Set window overlap. If set to 1, the recommended overlap for selected
1158 window function will be picked. Default is @code{0.75}.
1161 @subsection Examples
1165 Leave almost only low frequencies in audio:
1167 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1171 Apply robotize effect:
1173 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1177 Apply whisper effect:
1179 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"
1186 Apply an arbitrary Finite Impulse Response filter.
1188 This filter is designed for applying long FIR filters,
1189 up to 60 seconds long.
1191 It can be used as component for digital crossover filters,
1192 room equalization, cross talk cancellation, wavefield synthesis,
1193 auralization, ambiophonics, ambisonics and spatialization.
1195 This filter uses the streams higher than first one as FIR coefficients.
1196 If the non-first stream holds a single channel, it will be used
1197 for all input channels in the first stream, otherwise
1198 the number of channels in the non-first stream must be same as
1199 the number of channels in the first stream.
1201 It accepts the following parameters:
1205 Set dry gain. This sets input gain.
1208 Set wet gain. This sets final output gain.
1211 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1214 Enable applying gain measured from power of IR.
1216 Set which approach to use for auto gain measurement.
1220 Do not apply any gain.
1223 select peak gain, very conservative approach. This is default value.
1226 select DC gain, limited application.
1229 select gain to noise approach, this is most popular one.
1233 Set gain to be applied to IR coefficients before filtering.
1234 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1237 Set format of IR stream. Can be @code{mono} or @code{input}.
1238 Default is @code{input}.
1241 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1242 Allowed range is 0.1 to 60 seconds.
1245 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1246 By default it is disabled.
1249 Set for which IR channel to display frequency response. By default is first channel
1250 displayed. This option is used only when @var{response} is enabled.
1253 Set video stream size. This option is used only when @var{response} is enabled.
1256 Set video stream frame rate. This option is used only when @var{response} is enabled.
1259 Set minimal partition size used for convolution. Default is @var{8192}.
1260 Allowed range is from @var{1} to @var{32768}.
1261 Lower values decreases latency at cost of higher CPU usage.
1264 Set maximal partition size used for convolution. Default is @var{8192}.
1265 Allowed range is from @var{8} to @var{32768}.
1266 Lower values may increase CPU usage.
1269 Set number of input impulse responses streams which will be switchable at runtime.
1270 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1273 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1274 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1275 This option can be changed at runtime via @ref{commands}.
1278 @subsection Examples
1282 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1284 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1291 Set output format constraints for the input audio. The framework will
1292 negotiate the most appropriate format to minimize conversions.
1294 It accepts the following parameters:
1297 @item sample_fmts, f
1298 A '|'-separated list of requested sample formats.
1300 @item sample_rates, r
1301 A '|'-separated list of requested sample rates.
1303 @item channel_layouts, cl
1304 A '|'-separated list of requested channel layouts.
1306 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1307 for the required syntax.
1310 If a parameter is omitted, all values are allowed.
1312 Force the output to either unsigned 8-bit or signed 16-bit stereo
1314 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1319 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1320 processing reduces disturbing noise between useful signals.
1322 Gating is done by detecting the volume below a chosen level @var{threshold}
1323 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1324 floor is set via @var{range}. Because an exact manipulation of the signal
1325 would cause distortion of the waveform the reduction can be levelled over
1326 time. This is done by setting @var{attack} and @var{release}.
1328 @var{attack} determines how long the signal has to fall below the threshold
1329 before any reduction will occur and @var{release} sets the time the signal
1330 has to rise above the threshold to reduce the reduction again.
1331 Shorter signals than the chosen attack time will be left untouched.
1335 Set input level before filtering.
1336 Default is 1. Allowed range is from 0.015625 to 64.
1339 Set the mode of operation. Can be @code{upward} or @code{downward}.
1340 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1341 will be amplified, expanding dynamic range in upward direction.
1342 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1345 Set the level of gain reduction when the signal is below the threshold.
1346 Default is 0.06125. Allowed range is from 0 to 1.
1347 Setting this to 0 disables reduction and then filter behaves like expander.
1350 If a signal rises above this level the gain reduction is released.
1351 Default is 0.125. Allowed range is from 0 to 1.
1354 Set a ratio by which the signal is reduced.
1355 Default is 2. Allowed range is from 1 to 9000.
1358 Amount of milliseconds the signal has to rise above the threshold before gain
1360 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1363 Amount of milliseconds the signal has to fall below the threshold before the
1364 reduction is increased again. Default is 250 milliseconds.
1365 Allowed range is from 0.01 to 9000.
1368 Set amount of amplification of signal after processing.
1369 Default is 1. Allowed range is from 1 to 64.
1372 Curve the sharp knee around the threshold to enter gain reduction more softly.
1373 Default is 2.828427125. Allowed range is from 1 to 8.
1376 Choose if exact signal should be taken for detection or an RMS like one.
1377 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1380 Choose if the average level between all channels or the louder channel affects
1382 Default is @code{average}. Can be @code{average} or @code{maximum}.
1387 Apply an arbitrary Infinite Impulse Response filter.
1389 It accepts the following parameters:
1393 Set numerator/zeros coefficients.
1396 Set denominator/poles coefficients.
1408 Set coefficients format.
1414 Z-plane zeros/poles, cartesian (default)
1416 Z-plane zeros/poles, polar radians
1418 Z-plane zeros/poles, polar degrees
1422 Set kind of processing.
1423 Can be @code{d} - direct or @code{s} - serial cascading. Default is @code{s}.
1426 Set filtering precision.
1430 double-precision floating-point (default)
1432 single-precision floating-point
1440 How much to use filtered signal in output. Default is 1.
1441 Range is between 0 and 1.
1444 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1445 By default it is disabled.
1448 Set for which IR channel to display frequency response. By default is first channel
1449 displayed. This option is used only when @var{response} is enabled.
1452 Set video stream size. This option is used only when @var{response} is enabled.
1455 Coefficients in @code{tf} format are separated by spaces and are in ascending
1458 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1459 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1462 Different coefficients and gains can be provided for every channel, in such case
1463 use '|' to separate coefficients or gains. Last provided coefficients will be
1464 used for all remaining channels.
1466 @subsection Examples
1470 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1472 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
1476 Same as above but in @code{zp} format:
1478 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
1484 The limiter prevents an input signal from rising over a desired threshold.
1485 This limiter uses lookahead technology to prevent your signal from distorting.
1486 It means that there is a small delay after the signal is processed. Keep in mind
1487 that the delay it produces is the attack time you set.
1489 The filter accepts the following options:
1493 Set input gain. Default is 1.
1496 Set output gain. Default is 1.
1499 Don't let signals above this level pass the limiter. Default is 1.
1502 The limiter will reach its attenuation level in this amount of time in
1503 milliseconds. Default is 5 milliseconds.
1506 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1507 Default is 50 milliseconds.
1510 When gain reduction is always needed ASC takes care of releasing to an
1511 average reduction level rather than reaching a reduction of 0 in the release
1515 Select how much the release time is affected by ASC, 0 means nearly no changes
1516 in release time while 1 produces higher release times.
1519 Auto level output signal. Default is enabled.
1520 This normalizes audio back to 0dB if enabled.
1523 Depending on picked setting it is recommended to upsample input 2x or 4x times
1524 with @ref{aresample} before applying this filter.
1528 Apply a two-pole all-pass filter with central frequency (in Hz)
1529 @var{frequency}, and filter-width @var{width}.
1530 An all-pass filter changes the audio's frequency to phase relationship
1531 without changing its frequency to amplitude relationship.
1533 The filter accepts the following options:
1537 Set frequency in Hz.
1540 Set method to specify band-width of filter.
1555 Specify the band-width of a filter in width_type units.
1558 How much to use filtered signal in output. Default is 1.
1559 Range is between 0 and 1.
1562 Specify which channels to filter, by default all available are filtered.
1565 Normalize biquad coefficients, by default is disabled.
1566 Enabling it will normalize magnitude response at DC to 0dB.
1569 @subsection Commands
1571 This filter supports the following commands:
1574 Change allpass frequency.
1575 Syntax for the command is : "@var{frequency}"
1578 Change allpass width_type.
1579 Syntax for the command is : "@var{width_type}"
1582 Change allpass width.
1583 Syntax for the command is : "@var{width}"
1587 Syntax for the command is : "@var{mix}"
1594 The filter accepts the following options:
1598 Set the number of loops. Setting this value to -1 will result in infinite loops.
1602 Set maximal number of samples. Default is 0.
1605 Set first sample of loop. Default is 0.
1611 Merge two or more audio streams into a single multi-channel stream.
1613 The filter accepts the following options:
1618 Set the number of inputs. Default is 2.
1622 If the channel layouts of the inputs are disjoint, and therefore compatible,
1623 the channel layout of the output will be set accordingly and the channels
1624 will be reordered as necessary. If the channel layouts of the inputs are not
1625 disjoint, the output will have all the channels of the first input then all
1626 the channels of the second input, in that order, and the channel layout of
1627 the output will be the default value corresponding to the total number of
1630 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1631 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1632 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1633 first input, b1 is the first channel of the second input).
1635 On the other hand, if both input are in stereo, the output channels will be
1636 in the default order: a1, a2, b1, b2, and the channel layout will be
1637 arbitrarily set to 4.0, which may or may not be the expected value.
1639 All inputs must have the same sample rate, and format.
1641 If inputs do not have the same duration, the output will stop with the
1644 @subsection Examples
1648 Merge two mono files into a stereo stream:
1650 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1654 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1656 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
1662 Mixes multiple audio inputs into a single output.
1664 Note that this filter only supports float samples (the @var{amerge}
1665 and @var{pan} audio filters support many formats). If the @var{amix}
1666 input has integer samples then @ref{aresample} will be automatically
1667 inserted to perform the conversion to float samples.
1671 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1673 will mix 3 input audio streams to a single output with the same duration as the
1674 first input and a dropout transition time of 3 seconds.
1676 It accepts the following parameters:
1680 The number of inputs. If unspecified, it defaults to 2.
1683 How to determine the end-of-stream.
1687 The duration of the longest input. (default)
1690 The duration of the shortest input.
1693 The duration of the first input.
1697 @item dropout_transition
1698 The transition time, in seconds, for volume renormalization when an input
1699 stream ends. The default value is 2 seconds.
1702 Specify weight of each input audio stream as sequence.
1703 Each weight is separated by space. By default all inputs have same weight.
1708 Multiply first audio stream with second audio stream and store result
1709 in output audio stream. Multiplication is done by multiplying each
1710 sample from first stream with sample at same position from second stream.
1712 With this element-wise multiplication one can create amplitude fades and
1713 amplitude modulations.
1715 @section anequalizer
1717 High-order parametric multiband equalizer for each channel.
1719 It accepts the following parameters:
1723 This option string is in format:
1724 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1725 Each equalizer band is separated by '|'.
1729 Set channel number to which equalization will be applied.
1730 If input doesn't have that channel the entry is ignored.
1733 Set central frequency for band.
1734 If input doesn't have that frequency the entry is ignored.
1737 Set band width in hertz.
1740 Set band gain in dB.
1743 Set filter type for band, optional, can be:
1747 Butterworth, this is default.
1758 With this option activated frequency response of anequalizer is displayed
1762 Set video stream size. Only useful if curves option is activated.
1765 Set max gain that will be displayed. Only useful if curves option is activated.
1766 Setting this to a reasonable value makes it possible to display gain which is derived from
1767 neighbour bands which are too close to each other and thus produce higher gain
1768 when both are activated.
1771 Set frequency scale used to draw frequency response in video output.
1772 Can be linear or logarithmic. Default is logarithmic.
1775 Set color for each channel curve which is going to be displayed in video stream.
1776 This is list of color names separated by space or by '|'.
1777 Unrecognised or missing colors will be replaced by white color.
1780 @subsection Examples
1784 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1785 for first 2 channels using Chebyshev type 1 filter:
1787 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1791 @subsection Commands
1793 This filter supports the following commands:
1796 Alter existing filter parameters.
1797 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1799 @var{fN} is existing filter number, starting from 0, if no such filter is available
1801 @var{freq} set new frequency parameter.
1802 @var{width} set new width parameter in herz.
1803 @var{gain} set new gain parameter in dB.
1805 Full filter invocation with asendcmd may look like this:
1806 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1811 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1813 Each sample is adjusted by looking for other samples with similar contexts. This
1814 context similarity is defined by comparing their surrounding patches of size
1815 @option{p}. Patches are searched in an area of @option{r} around the sample.
1817 The filter accepts the following options:
1821 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1824 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1825 Default value is 2 milliseconds.
1828 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1829 Default value is 6 milliseconds.
1832 Set the output mode.
1834 It accepts the following values:
1837 Pass input unchanged.
1840 Pass noise filtered out.
1845 Default value is @var{o}.
1849 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
1852 @subsection Commands
1854 This filter supports the following commands:
1857 Change denoise strength. Argument is single float number.
1858 Syntax for the command is : "@var{s}"
1862 Syntax for the command is : "i", "o" or "n" string.
1866 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
1868 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
1869 relate to producing the least mean square of the error signal (difference between the desired,
1870 2nd input audio stream and the actual signal, the 1st input audio stream).
1872 A description of the accepted options follows.
1885 Set the filter leakage.
1888 It accepts the following values:
1897 Pass filtered samples.
1900 Pass difference between desired and filtered samples.
1902 Default value is @var{o}.
1906 @subsection Examples
1910 One of many usages of this filter is noise reduction, input audio is filtered
1911 with same samples that are delayed by fixed amount, one such example for stereo audio is:
1913 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
1917 @subsection Commands
1919 This filter supports the same commands as options, excluding option @code{order}.
1923 Pass the audio source unchanged to the output.
1927 Pad the end of an audio stream with silence.
1929 This can be used together with @command{ffmpeg} @option{-shortest} to
1930 extend audio streams to the same length as the video stream.
1932 A description of the accepted options follows.
1936 Set silence packet size. Default value is 4096.
1939 Set the number of samples of silence to add to the end. After the
1940 value is reached, the stream is terminated. This option is mutually
1941 exclusive with @option{whole_len}.
1944 Set the minimum total number of samples in the output audio stream. If
1945 the value is longer than the input audio length, silence is added to
1946 the end, until the value is reached. This option is mutually exclusive
1947 with @option{pad_len}.
1950 Specify the duration of samples of silence to add. See
1951 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1952 for the accepted syntax. Used only if set to non-zero value.
1955 Specify the minimum total duration in the output audio stream. See
1956 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1957 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
1958 the input audio length, silence is added to the end, until the value is reached.
1959 This option is mutually exclusive with @option{pad_dur}
1962 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
1963 nor @option{whole_dur} option is set, the filter will add silence to the end of
1964 the input stream indefinitely.
1966 @subsection Examples
1970 Add 1024 samples of silence to the end of the input:
1976 Make sure the audio output will contain at least 10000 samples, pad
1977 the input with silence if required:
1979 apad=whole_len=10000
1983 Use @command{ffmpeg} to pad the audio input with silence, so that the
1984 video stream will always result the shortest and will be converted
1985 until the end in the output file when using the @option{shortest}
1988 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
1993 Add a phasing effect to the input audio.
1995 A phaser filter creates series of peaks and troughs in the frequency spectrum.
1996 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
1998 A description of the accepted parameters follows.
2002 Set input gain. Default is 0.4.
2005 Set output gain. Default is 0.74
2008 Set delay in milliseconds. Default is 3.0.
2011 Set decay. Default is 0.4.
2014 Set modulation speed in Hz. Default is 0.5.
2017 Set modulation type. Default is triangular.
2019 It accepts the following values:
2028 Audio pulsator is something between an autopanner and a tremolo.
2029 But it can produce funny stereo effects as well. Pulsator changes the volume
2030 of the left and right channel based on a LFO (low frequency oscillator) with
2031 different waveforms and shifted phases.
2032 This filter have the ability to define an offset between left and right
2033 channel. An offset of 0 means that both LFO shapes match each other.
2034 The left and right channel are altered equally - a conventional tremolo.
2035 An offset of 50% means that the shape of the right channel is exactly shifted
2036 in phase (or moved backwards about half of the frequency) - pulsator acts as
2037 an autopanner. At 1 both curves match again. Every setting in between moves the
2038 phase shift gapless between all stages and produces some "bypassing" sounds with
2039 sine and triangle waveforms. The more you set the offset near 1 (starting from
2040 the 0.5) the faster the signal passes from the left to the right speaker.
2042 The filter accepts the following options:
2046 Set input gain. By default it is 1. Range is [0.015625 - 64].
2049 Set output gain. By default it is 1. Range is [0.015625 - 64].
2052 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2053 sawup or sawdown. Default is sine.
2056 Set modulation. Define how much of original signal is affected by the LFO.
2059 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2062 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2065 Set pulse width. Default is 1. Allowed range is [0 - 2].
2068 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2071 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2075 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2079 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2080 if timing is set to hz.
2086 Resample the input audio to the specified parameters, using the
2087 libswresample library. If none are specified then the filter will
2088 automatically convert between its input and output.
2090 This filter is also able to stretch/squeeze the audio data to make it match
2091 the timestamps or to inject silence / cut out audio to make it match the
2092 timestamps, do a combination of both or do neither.
2094 The filter accepts the syntax
2095 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2096 expresses a sample rate and @var{resampler_options} is a list of
2097 @var{key}=@var{value} pairs, separated by ":". See the
2098 @ref{Resampler Options,,"Resampler Options" section in the
2099 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2100 for the complete list of supported options.
2102 @subsection Examples
2106 Resample the input audio to 44100Hz:
2112 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2113 samples per second compensation:
2115 aresample=async=1000
2121 Reverse an audio clip.
2123 Warning: This filter requires memory to buffer the entire clip, so trimming
2126 @subsection Examples
2130 Take the first 5 seconds of a clip, and reverse it.
2132 atrim=end=5,areverse
2138 Reduce noise from speech using Recurrent Neural Networks.
2140 This filter accepts the following options:
2144 Set train model file to load. This option is always required.
2147 @section asetnsamples
2149 Set the number of samples per each output audio frame.
2151 The last output packet may contain a different number of samples, as
2152 the filter will flush all the remaining samples when the input audio
2155 The filter accepts the following options:
2159 @item nb_out_samples, n
2160 Set the number of frames per each output audio frame. The number is
2161 intended as the number of samples @emph{per each channel}.
2162 Default value is 1024.
2165 If set to 1, the filter will pad the last audio frame with zeroes, so
2166 that the last frame will contain the same number of samples as the
2167 previous ones. Default value is 1.
2170 For example, to set the number of per-frame samples to 1234 and
2171 disable padding for the last frame, use:
2173 asetnsamples=n=1234:p=0
2178 Set the sample rate without altering the PCM data.
2179 This will result in a change of speed and pitch.
2181 The filter accepts the following options:
2184 @item sample_rate, r
2185 Set the output sample rate. Default is 44100 Hz.
2190 Show a line containing various information for each input audio frame.
2191 The input audio is not modified.
2193 The shown line contains a sequence of key/value pairs of the form
2194 @var{key}:@var{value}.
2196 The following values are shown in the output:
2200 The (sequential) number of the input frame, starting from 0.
2203 The presentation timestamp of the input frame, in time base units; the time base
2204 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2207 The presentation timestamp of the input frame in seconds.
2210 position of the frame in the input stream, -1 if this information in
2211 unavailable and/or meaningless (for example in case of synthetic audio)
2220 The sample rate for the audio frame.
2223 The number of samples (per channel) in the frame.
2226 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2227 audio, the data is treated as if all the planes were concatenated.
2229 @item plane_checksums
2230 A list of Adler-32 checksums for each data plane.
2234 Apply audio soft clipping.
2236 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2237 along a smooth curve, rather than the abrupt shape of hard-clipping.
2239 This filter accepts the following options:
2243 Set type of soft-clipping.
2245 It accepts the following values:
2257 Set additional parameter which controls sigmoid function.
2260 @subsection Commands
2262 This filter supports the all above options as @ref{commands}.
2265 Automatic Speech Recognition
2267 This filter uses PocketSphinx for speech recognition. To enable
2268 compilation of this filter, you need to configure FFmpeg with
2269 @code{--enable-pocketsphinx}.
2271 It accepts the following options:
2275 Set sampling rate of input audio. Defaults is @code{16000}.
2276 This need to match speech models, otherwise one will get poor results.
2279 Set dictionary containing acoustic model files.
2282 Set pronunciation dictionary.
2285 Set language model file.
2288 Set language model set.
2291 Set which language model to use.
2294 Set output for log messages.
2297 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2302 Display time domain statistical information about the audio channels.
2303 Statistics are calculated and displayed for each audio channel and,
2304 where applicable, an overall figure is also given.
2306 It accepts the following option:
2309 Short window length in seconds, used for peak and trough RMS measurement.
2310 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2314 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2315 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2318 Available keys for each channel are:
2364 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2365 this @code{lavfi.astats.Overall.Peak_count}.
2367 For description what each key means read below.
2370 Set number of frame after which stats are going to be recalculated.
2371 Default is disabled.
2373 @item measure_perchannel
2374 Select the entries which need to be measured per channel. The metadata keys can
2375 be used as flags, default is @option{all} which measures everything.
2376 @option{none} disables all per channel measurement.
2378 @item measure_overall
2379 Select the entries which need to be measured overall. The metadata keys can
2380 be used as flags, default is @option{all} which measures everything.
2381 @option{none} disables all overall measurement.
2385 A description of each shown parameter follows:
2389 Mean amplitude displacement from zero.
2392 Minimal sample level.
2395 Maximal sample level.
2397 @item Min difference
2398 Minimal difference between two consecutive samples.
2400 @item Max difference
2401 Maximal difference between two consecutive samples.
2403 @item Mean difference
2404 Mean difference between two consecutive samples.
2405 The average of each difference between two consecutive samples.
2407 @item RMS difference
2408 Root Mean Square difference between two consecutive samples.
2412 Standard peak and RMS level measured in dBFS.
2416 Peak and trough values for RMS level measured over a short window.
2419 Standard ratio of peak to RMS level (note: not in dB).
2422 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2423 (i.e. either @var{Min level} or @var{Max level}).
2426 Number of occasions (not the number of samples) that the signal attained either
2427 @var{Min level} or @var{Max level}.
2429 @item Noise floor dB
2430 Minimum local peak measured in dBFS over a short window.
2432 @item Noise floor count
2433 Number of occasions (not the number of samples) that the signal attained
2437 Overall bit depth of audio. Number of bits used for each sample.
2440 Measured dynamic range of audio in dB.
2442 @item Zero crossings
2443 Number of points where the waveform crosses the zero level axis.
2445 @item Zero crossings rate
2446 Rate of Zero crossings and number of audio samples.
2453 The filter accepts exactly one parameter, the audio tempo. If not
2454 specified then the filter will assume nominal 1.0 tempo. Tempo must
2455 be in the [0.5, 100.0] range.
2457 Note that tempo greater than 2 will skip some samples rather than
2458 blend them in. If for any reason this is a concern it is always
2459 possible to daisy-chain several instances of atempo to achieve the
2460 desired product tempo.
2462 @subsection Examples
2466 Slow down audio to 80% tempo:
2472 To speed up audio to 300% tempo:
2478 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2480 atempo=sqrt(3),atempo=sqrt(3)
2484 @subsection Commands
2486 This filter supports the following commands:
2489 Change filter tempo scale factor.
2490 Syntax for the command is : "@var{tempo}"
2495 Trim the input so that the output contains one continuous subpart of the input.
2497 It accepts the following parameters:
2500 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2501 sample with the timestamp @var{start} will be the first sample in the output.
2504 Specify time of the first audio sample that will be dropped, i.e. the
2505 audio sample immediately preceding the one with the timestamp @var{end} will be
2506 the last sample in the output.
2509 Same as @var{start}, except this option sets the start timestamp in samples
2513 Same as @var{end}, except this option sets the end timestamp in samples instead
2517 The maximum duration of the output in seconds.
2520 The number of the first sample that should be output.
2523 The number of the first sample that should be dropped.
2526 @option{start}, @option{end}, and @option{duration} are expressed as time
2527 duration specifications; see
2528 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2530 Note that the first two sets of the start/end options and the @option{duration}
2531 option look at the frame timestamp, while the _sample options simply count the
2532 samples that pass through the filter. So start/end_pts and start/end_sample will
2533 give different results when the timestamps are wrong, inexact or do not start at
2534 zero. Also note that this filter does not modify the timestamps. If you wish
2535 to have the output timestamps start at zero, insert the asetpts filter after the
2538 If multiple start or end options are set, this filter tries to be greedy and
2539 keep all samples that match at least one of the specified constraints. To keep
2540 only the part that matches all the constraints at once, chain multiple atrim
2543 The defaults are such that all the input is kept. So it is possible to set e.g.
2544 just the end values to keep everything before the specified time.
2549 Drop everything except the second minute of input:
2551 ffmpeg -i INPUT -af atrim=60:120
2555 Keep only the first 1000 samples:
2557 ffmpeg -i INPUT -af atrim=end_sample=1000
2562 @section axcorrelate
2563 Calculate normalized cross-correlation between two input audio streams.
2565 Resulted samples are always between -1 and 1 inclusive.
2566 If result is 1 it means two input samples are highly correlated in that selected segment.
2567 Result 0 means they are not correlated at all.
2568 If result is -1 it means two input samples are out of phase, which means they cancel each
2571 The filter accepts the following options:
2575 Set size of segment over which cross-correlation is calculated.
2576 Default is 256. Allowed range is from 2 to 131072.
2579 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2580 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2581 are always zero and thus need much less calculations to make.
2582 This is generally not true, but is valid for typical audio streams.
2585 @subsection Examples
2589 Calculate correlation between channels in stereo audio stream:
2591 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2597 Apply a two-pole Butterworth band-pass filter with central
2598 frequency @var{frequency}, and (3dB-point) band-width width.
2599 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2600 instead of the default: constant 0dB peak gain.
2601 The filter roll off at 6dB per octave (20dB per decade).
2603 The filter accepts the following options:
2607 Set the filter's central frequency. Default is @code{3000}.
2610 Constant skirt gain if set to 1. Defaults to 0.
2613 Set method to specify band-width of filter.
2628 Specify the band-width of a filter in width_type units.
2631 How much to use filtered signal in output. Default is 1.
2632 Range is between 0 and 1.
2635 Specify which channels to filter, by default all available are filtered.
2638 Normalize biquad coefficients, by default is disabled.
2639 Enabling it will normalize magnitude response at DC to 0dB.
2642 @subsection Commands
2644 This filter supports the following commands:
2647 Change bandpass frequency.
2648 Syntax for the command is : "@var{frequency}"
2651 Change bandpass width_type.
2652 Syntax for the command is : "@var{width_type}"
2655 Change bandpass width.
2656 Syntax for the command is : "@var{width}"
2659 Change bandpass mix.
2660 Syntax for the command is : "@var{mix}"
2665 Apply a two-pole Butterworth band-reject filter with central
2666 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2667 The filter roll off at 6dB per octave (20dB per decade).
2669 The filter accepts the following options:
2673 Set the filter's central frequency. Default is @code{3000}.
2676 Set method to specify band-width of filter.
2691 Specify the band-width of a filter in width_type units.
2694 How much to use filtered signal in output. Default is 1.
2695 Range is between 0 and 1.
2698 Specify which channels to filter, by default all available are filtered.
2701 Normalize biquad coefficients, by default is disabled.
2702 Enabling it will normalize magnitude response at DC to 0dB.
2705 @subsection Commands
2707 This filter supports the following commands:
2710 Change bandreject frequency.
2711 Syntax for the command is : "@var{frequency}"
2714 Change bandreject width_type.
2715 Syntax for the command is : "@var{width_type}"
2718 Change bandreject width.
2719 Syntax for the command is : "@var{width}"
2722 Change bandreject mix.
2723 Syntax for the command is : "@var{mix}"
2726 @section bass, lowshelf
2728 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2729 shelving filter with a response similar to that of a standard
2730 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2732 The filter accepts the following options:
2736 Give the gain at 0 Hz. Its useful range is about -20
2737 (for a large cut) to +20 (for a large boost).
2738 Beware of clipping when using a positive gain.
2741 Set the filter's central frequency and so can be used
2742 to extend or reduce the frequency range to be boosted or cut.
2743 The default value is @code{100} Hz.
2746 Set method to specify band-width of filter.
2761 Determine how steep is the filter's shelf transition.
2764 How much to use filtered signal in output. Default is 1.
2765 Range is between 0 and 1.
2768 Specify which channels to filter, by default all available are filtered.
2771 Normalize biquad coefficients, by default is disabled.
2772 Enabling it will normalize magnitude response at DC to 0dB.
2775 @subsection Commands
2777 This filter supports the following commands:
2780 Change bass frequency.
2781 Syntax for the command is : "@var{frequency}"
2784 Change bass width_type.
2785 Syntax for the command is : "@var{width_type}"
2789 Syntax for the command is : "@var{width}"
2793 Syntax for the command is : "@var{gain}"
2797 Syntax for the command is : "@var{mix}"
2802 Apply a biquad IIR filter with the given coefficients.
2803 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2804 are the numerator and denominator coefficients respectively.
2805 and @var{channels}, @var{c} specify which channels to filter, by default all
2806 available are filtered.
2808 @subsection Commands
2810 This filter supports the following commands:
2818 Change biquad parameter.
2819 Syntax for the command is : "@var{value}"
2822 How much to use filtered signal in output. Default is 1.
2823 Range is between 0 and 1.
2826 Specify which channels to filter, by default all available are filtered.
2829 Normalize biquad coefficients, by default is disabled.
2830 Enabling it will normalize magnitude response at DC to 0dB.
2834 Bauer stereo to binaural transformation, which improves headphone listening of
2835 stereo audio records.
2837 To enable compilation of this filter you need to configure FFmpeg with
2838 @code{--enable-libbs2b}.
2840 It accepts the following parameters:
2844 Pre-defined crossfeed level.
2848 Default level (fcut=700, feed=50).
2851 Chu Moy circuit (fcut=700, feed=60).
2854 Jan Meier circuit (fcut=650, feed=95).
2859 Cut frequency (in Hz).
2868 Remap input channels to new locations.
2870 It accepts the following parameters:
2873 Map channels from input to output. The argument is a '|'-separated list of
2874 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2875 @var{in_channel} form. @var{in_channel} can be either the name of the input
2876 channel (e.g. FL for front left) or its index in the input channel layout.
2877 @var{out_channel} is the name of the output channel or its index in the output
2878 channel layout. If @var{out_channel} is not given then it is implicitly an
2879 index, starting with zero and increasing by one for each mapping.
2881 @item channel_layout
2882 The channel layout of the output stream.
2885 If no mapping is present, the filter will implicitly map input channels to
2886 output channels, preserving indices.
2888 @subsection Examples
2892 For example, assuming a 5.1+downmix input MOV file,
2894 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2896 will create an output WAV file tagged as stereo from the downmix channels of
2900 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2902 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2906 @section channelsplit
2908 Split each channel from an input audio stream into a separate output stream.
2910 It accepts the following parameters:
2912 @item channel_layout
2913 The channel layout of the input stream. The default is "stereo".
2915 A channel layout describing the channels to be extracted as separate output streams
2916 or "all" to extract each input channel as a separate stream. The default is "all".
2918 Choosing channels not present in channel layout in the input will result in an error.
2921 @subsection Examples
2925 For example, assuming a stereo input MP3 file,
2927 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2929 will create an output Matroska file with two audio streams, one containing only
2930 the left channel and the other the right channel.
2933 Split a 5.1 WAV file into per-channel files:
2935 ffmpeg -i in.wav -filter_complex
2936 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2937 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2938 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2943 Extract only LFE from a 5.1 WAV file:
2945 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
2946 -map '[LFE]' lfe.wav
2951 Add a chorus effect to the audio.
2953 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2955 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2956 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2957 The modulation depth defines the range the modulated delay is played before or after
2958 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2959 sound tuned around the original one, like in a chorus where some vocals are slightly
2962 It accepts the following parameters:
2965 Set input gain. Default is 0.4.
2968 Set output gain. Default is 0.4.
2971 Set delays. A typical delay is around 40ms to 60ms.
2983 @subsection Examples
2989 chorus=0.7:0.9:55:0.4:0.25:2
2995 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2999 Fuller sounding chorus with three delays:
3001 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
3006 Compress or expand the audio's dynamic range.
3008 It accepts the following parameters:
3014 A list of times in seconds for each channel over which the instantaneous level
3015 of the input signal is averaged to determine its volume. @var{attacks} refers to
3016 increase of volume and @var{decays} refers to decrease of volume. For most
3017 situations, the attack time (response to the audio getting louder) should be
3018 shorter than the decay time, because the human ear is more sensitive to sudden
3019 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3020 a typical value for decay is 0.8 seconds.
3021 If specified number of attacks & decays is lower than number of channels, the last
3022 set attack/decay will be used for all remaining channels.
3025 A list of points for the transfer function, specified in dB relative to the
3026 maximum possible signal amplitude. Each key points list must be defined using
3027 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3028 @code{x0/y0 x1/y1 x2/y2 ....}
3030 The input values must be in strictly increasing order but the transfer function
3031 does not have to be monotonically rising. The point @code{0/0} is assumed but
3032 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3033 function are @code{-70/-70|-60/-20|1/0}.
3036 Set the curve radius in dB for all joints. It defaults to 0.01.
3039 Set the additional gain in dB to be applied at all points on the transfer
3040 function. This allows for easy adjustment of the overall gain.
3044 Set an initial volume, in dB, to be assumed for each channel when filtering
3045 starts. This permits the user to supply a nominal level initially, so that, for
3046 example, a very large gain is not applied to initial signal levels before the
3047 companding has begun to operate. A typical value for audio which is initially
3048 quiet is -90 dB. It defaults to 0.
3051 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3052 delayed before being fed to the volume adjuster. Specifying a delay
3053 approximately equal to the attack/decay times allows the filter to effectively
3054 operate in predictive rather than reactive mode. It defaults to 0.
3058 @subsection Examples
3062 Make music with both quiet and loud passages suitable for listening to in a
3065 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3068 Another example for audio with whisper and explosion parts:
3070 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3074 A noise gate for when the noise is at a lower level than the signal:
3076 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3080 Here is another noise gate, this time for when the noise is at a higher level
3081 than the signal (making it, in some ways, similar to squelch):
3083 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3087 2:1 compression starting at -6dB:
3089 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3093 2:1 compression starting at -9dB:
3095 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3099 2:1 compression starting at -12dB:
3101 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3105 2:1 compression starting at -18dB:
3107 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3111 3:1 compression starting at -15dB:
3113 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3119 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3125 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
3129 Hard limiter at -6dB:
3131 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3135 Hard limiter at -12dB:
3137 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3141 Hard noise gate at -35 dB:
3143 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3149 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3153 @section compensationdelay
3155 Compensation Delay Line is a metric based delay to compensate differing
3156 positions of microphones or speakers.
3158 For example, you have recorded guitar with two microphones placed in
3159 different locations. Because the front of sound wave has fixed speed in
3160 normal conditions, the phasing of microphones can vary and depends on
3161 their location and interposition. The best sound mix can be achieved when
3162 these microphones are in phase (synchronized). Note that a distance of
3163 ~30 cm between microphones makes one microphone capture the signal in
3164 antiphase to the other microphone. That makes the final mix sound moody.
3165 This filter helps to solve phasing problems by adding different delays
3166 to each microphone track and make them synchronized.
3168 The best result can be reached when you take one track as base and
3169 synchronize other tracks one by one with it.
3170 Remember that synchronization/delay tolerance depends on sample rate, too.
3171 Higher sample rates will give more tolerance.
3173 The filter accepts the following parameters:
3177 Set millimeters distance. This is compensation distance for fine tuning.
3181 Set cm distance. This is compensation distance for tightening distance setup.
3185 Set meters distance. This is compensation distance for hard distance setup.
3189 Set dry amount. Amount of unprocessed (dry) signal.
3193 Set wet amount. Amount of processed (wet) signal.
3197 Set temperature in degrees Celsius. This is the temperature of the environment.
3202 Apply headphone crossfeed filter.
3204 Crossfeed is the process of blending the left and right channels of stereo
3206 It is mainly used to reduce extreme stereo separation of low frequencies.
3208 The intent is to produce more speaker like sound to the listener.
3210 The filter accepts the following options:
3214 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3215 This sets gain of low shelf filter for side part of stereo image.
3216 Default is -6dB. Max allowed is -30db when strength is set to 1.
3219 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3220 This sets cut off frequency of low shelf filter. Default is cut off near
3221 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3224 Set input gain. Default is 0.9.
3227 Set output gain. Default is 1.
3230 @section crystalizer
3231 Simple algorithm to expand audio dynamic range.
3233 The filter accepts the following options:
3237 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3238 (unchanged sound) to 10.0 (maximum effect).
3241 Enable clipping. By default is enabled.
3244 @subsection Commands
3246 This filter supports the all above options as @ref{commands}.
3249 Apply a DC shift to the audio.
3251 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3252 in the recording chain) from the audio. The effect of a DC offset is reduced
3253 headroom and hence volume. The @ref{astats} filter can be used to determine if
3254 a signal has a DC offset.
3258 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3262 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3263 used to prevent clipping.
3268 Apply de-essing to the audio samples.
3272 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3276 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3280 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3284 Set the output mode.
3286 It accepts the following values:
3289 Pass input unchanged.
3292 Pass ess filtered out.
3297 Default value is @var{o}.
3303 Measure audio dynamic range.
3305 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3306 is found in transition material. And anything less that 8 have very poor dynamics
3307 and is very compressed.
3309 The filter accepts the following options:
3313 Set window length in seconds used to split audio into segments of equal length.
3314 Default is 3 seconds.
3318 Dynamic Audio Normalizer.
3320 This filter applies a certain amount of gain to the input audio in order
3321 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3322 contrast to more "simple" normalization algorithms, the Dynamic Audio
3323 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3324 This allows for applying extra gain to the "quiet" sections of the audio
3325 while avoiding distortions or clipping the "loud" sections. In other words:
3326 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3327 sections, in the sense that the volume of each section is brought to the
3328 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3329 this goal *without* applying "dynamic range compressing". It will retain 100%
3330 of the dynamic range *within* each section of the audio file.
3334 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3335 Default is 500 milliseconds.
3336 The Dynamic Audio Normalizer processes the input audio in small chunks,
3337 referred to as frames. This is required, because a peak magnitude has no
3338 meaning for just a single sample value. Instead, we need to determine the
3339 peak magnitude for a contiguous sequence of sample values. While a "standard"
3340 normalizer would simply use the peak magnitude of the complete file, the
3341 Dynamic Audio Normalizer determines the peak magnitude individually for each
3342 frame. The length of a frame is specified in milliseconds. By default, the
3343 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3344 been found to give good results with most files.
3345 Note that the exact frame length, in number of samples, will be determined
3346 automatically, based on the sampling rate of the individual input audio file.
3349 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3350 number. Default is 31.
3351 Probably the most important parameter of the Dynamic Audio Normalizer is the
3352 @code{window size} of the Gaussian smoothing filter. The filter's window size
3353 is specified in frames, centered around the current frame. For the sake of
3354 simplicity, this must be an odd number. Consequently, the default value of 31
3355 takes into account the current frame, as well as the 15 preceding frames and
3356 the 15 subsequent frames. Using a larger window results in a stronger
3357 smoothing effect and thus in less gain variation, i.e. slower gain
3358 adaptation. Conversely, using a smaller window results in a weaker smoothing
3359 effect and thus in more gain variation, i.e. faster gain adaptation.
3360 In other words, the more you increase this value, the more the Dynamic Audio
3361 Normalizer will behave like a "traditional" normalization filter. On the
3362 contrary, the more you decrease this value, the more the Dynamic Audio
3363 Normalizer will behave like a dynamic range compressor.
3366 Set the target peak value. This specifies the highest permissible magnitude
3367 level for the normalized audio input. This filter will try to approach the
3368 target peak magnitude as closely as possible, but at the same time it also
3369 makes sure that the normalized signal will never exceed the peak magnitude.
3370 A frame's maximum local gain factor is imposed directly by the target peak
3371 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3372 It is not recommended to go above this value.
3375 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3376 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3377 factor for each input frame, i.e. the maximum gain factor that does not
3378 result in clipping or distortion. The maximum gain factor is determined by
3379 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3380 additionally bounds the frame's maximum gain factor by a predetermined
3381 (global) maximum gain factor. This is done in order to avoid excessive gain
3382 factors in "silent" or almost silent frames. By default, the maximum gain
3383 factor is 10.0, For most inputs the default value should be sufficient and
3384 it usually is not recommended to increase this value. Though, for input
3385 with an extremely low overall volume level, it may be necessary to allow even
3386 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3387 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3388 Instead, a "sigmoid" threshold function will be applied. This way, the
3389 gain factors will smoothly approach the threshold value, but never exceed that
3393 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3394 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3395 This means that the maximum local gain factor for each frame is defined
3396 (only) by the frame's highest magnitude sample. This way, the samples can
3397 be amplified as much as possible without exceeding the maximum signal
3398 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3399 Normalizer can also take into account the frame's root mean square,
3400 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3401 determine the power of a time-varying signal. It is therefore considered
3402 that the RMS is a better approximation of the "perceived loudness" than
3403 just looking at the signal's peak magnitude. Consequently, by adjusting all
3404 frames to a constant RMS value, a uniform "perceived loudness" can be
3405 established. If a target RMS value has been specified, a frame's local gain
3406 factor is defined as the factor that would result in exactly that RMS value.
3407 Note, however, that the maximum local gain factor is still restricted by the
3408 frame's highest magnitude sample, in order to prevent clipping.
3411 Enable channels coupling. By default is enabled.
3412 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3413 amount. This means the same gain factor will be applied to all channels, i.e.
3414 the maximum possible gain factor is determined by the "loudest" channel.
3415 However, in some recordings, it may happen that the volume of the different
3416 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3417 In this case, this option can be used to disable the channel coupling. This way,
3418 the gain factor will be determined independently for each channel, depending
3419 only on the individual channel's highest magnitude sample. This allows for
3420 harmonizing the volume of the different channels.
3423 Enable DC bias correction. By default is disabled.
3424 An audio signal (in the time domain) is a sequence of sample values.
3425 In the Dynamic Audio Normalizer these sample values are represented in the
3426 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3427 audio signal, or "waveform", should be centered around the zero point.
3428 That means if we calculate the mean value of all samples in a file, or in a
3429 single frame, then the result should be 0.0 or at least very close to that
3430 value. If, however, there is a significant deviation of the mean value from
3431 0.0, in either positive or negative direction, this is referred to as a
3432 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3433 Audio Normalizer provides optional DC bias correction.
3434 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3435 the mean value, or "DC correction" offset, of each input frame and subtract
3436 that value from all of the frame's sample values which ensures those samples
3437 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3438 boundaries, the DC correction offset values will be interpolated smoothly
3439 between neighbouring frames.
3441 @item altboundary, b
3442 Enable alternative boundary mode. By default is disabled.
3443 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3444 around each frame. This includes the preceding frames as well as the
3445 subsequent frames. However, for the "boundary" frames, located at the very
3446 beginning and at the very end of the audio file, not all neighbouring
3447 frames are available. In particular, for the first few frames in the audio
3448 file, the preceding frames are not known. And, similarly, for the last few
3449 frames in the audio file, the subsequent frames are not known. Thus, the
3450 question arises which gain factors should be assumed for the missing frames
3451 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3452 to deal with this situation. The default boundary mode assumes a gain factor
3453 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3454 "fade out" at the beginning and at the end of the input, respectively.
3457 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3458 By default, the Dynamic Audio Normalizer does not apply "traditional"
3459 compression. This means that signal peaks will not be pruned and thus the
3460 full dynamic range will be retained within each local neighbourhood. However,
3461 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3462 normalization algorithm with a more "traditional" compression.
3463 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3464 (thresholding) function. If (and only if) the compression feature is enabled,
3465 all input frames will be processed by a soft knee thresholding function prior
3466 to the actual normalization process. Put simply, the thresholding function is
3467 going to prune all samples whose magnitude exceeds a certain threshold value.
3468 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3469 value. Instead, the threshold value will be adjusted for each individual
3471 In general, smaller parameters result in stronger compression, and vice versa.
3472 Values below 3.0 are not recommended, because audible distortion may appear.
3475 Set the target threshold value. This specifies the lowest permissible
3476 magnitude level for the audio input which will be normalized.
3477 If input frame volume is above this value frame will be normalized.
3478 Otherwise frame may not be normalized at all. The default value is set
3479 to 0, which means all input frames will be normalized.
3480 This option is mostly useful if digital noise is not wanted to be amplified.
3483 @subsection Commands
3485 This filter supports the all above options as @ref{commands}.
3489 Make audio easier to listen to on headphones.
3491 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3492 so that when listened to on headphones the stereo image is moved from
3493 inside your head (standard for headphones) to outside and in front of
3494 the listener (standard for speakers).
3500 Apply a two-pole peaking equalisation (EQ) filter. With this
3501 filter, the signal-level at and around a selected frequency can
3502 be increased or decreased, whilst (unlike bandpass and bandreject
3503 filters) that at all other frequencies is unchanged.
3505 In order to produce complex equalisation curves, this filter can
3506 be given several times, each with a different central frequency.
3508 The filter accepts the following options:
3512 Set the filter's central frequency in Hz.
3515 Set method to specify band-width of filter.
3530 Specify the band-width of a filter in width_type units.
3533 Set the required gain or attenuation in dB.
3534 Beware of clipping when using a positive gain.
3537 How much to use filtered signal in output. Default is 1.
3538 Range is between 0 and 1.
3541 Specify which channels to filter, by default all available are filtered.
3544 Normalize biquad coefficients, by default is disabled.
3545 Enabling it will normalize magnitude response at DC to 0dB.
3548 @subsection Examples
3551 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3553 equalizer=f=1000:t=h:width=200:g=-10
3557 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3559 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3563 @subsection Commands
3565 This filter supports the following commands:
3568 Change equalizer frequency.
3569 Syntax for the command is : "@var{frequency}"
3572 Change equalizer width_type.
3573 Syntax for the command is : "@var{width_type}"
3576 Change equalizer width.
3577 Syntax for the command is : "@var{width}"
3580 Change equalizer gain.
3581 Syntax for the command is : "@var{gain}"
3584 Change equalizer mix.
3585 Syntax for the command is : "@var{mix}"
3588 @section extrastereo
3590 Linearly increases the difference between left and right channels which
3591 adds some sort of "live" effect to playback.
3593 The filter accepts the following options:
3597 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3598 (average of both channels), with 1.0 sound will be unchanged, with
3599 -1.0 left and right channels will be swapped.
3602 Enable clipping. By default is enabled.
3605 @subsection Commands
3607 This filter supports the all above options as @ref{commands}.
3609 @section firequalizer
3610 Apply FIR Equalization using arbitrary frequency response.
3612 The filter accepts the following option:
3616 Set gain curve equation (in dB). The expression can contain variables:
3619 the evaluated frequency
3623 channel number, set to 0 when multichannels evaluation is disabled
3625 channel id, see libavutil/channel_layout.h, set to the first channel id when
3626 multichannels evaluation is disabled
3630 channel_layout, see libavutil/channel_layout.h
3635 @item gain_interpolate(f)
3636 interpolate gain on frequency f based on gain_entry
3637 @item cubic_interpolate(f)
3638 same as gain_interpolate, but smoother
3640 This option is also available as command. Default is @code{gain_interpolate(f)}.
3643 Set gain entry for gain_interpolate function. The expression can
3647 store gain entry at frequency f with value g
3649 This option is also available as command.
3652 Set filter delay in seconds. Higher value means more accurate.
3653 Default is @code{0.01}.
3656 Set filter accuracy in Hz. Lower value means more accurate.
3657 Default is @code{5}.
3660 Set window function. Acceptable values are:
3663 rectangular window, useful when gain curve is already smooth
3665 hann window (default)
3671 3-terms continuous 1st derivative nuttall window
3673 minimum 3-terms discontinuous nuttall window
3675 4-terms continuous 1st derivative nuttall window
3677 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3679 blackman-harris window
3685 If enabled, use fixed number of audio samples. This improves speed when
3686 filtering with large delay. Default is disabled.
3689 Enable multichannels evaluation on gain. Default is disabled.
3692 Enable zero phase mode by subtracting timestamp to compensate delay.
3693 Default is disabled.
3696 Set scale used by gain. Acceptable values are:
3699 linear frequency, linear gain
3701 linear frequency, logarithmic (in dB) gain (default)
3703 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3705 logarithmic frequency, logarithmic gain
3709 Set file for dumping, suitable for gnuplot.
3712 Set scale for dumpfile. Acceptable values are same with scale option.
3716 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3717 Default is disabled.
3720 Enable minimum phase impulse response. Default is disabled.
3723 @subsection Examples
3728 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3731 lowpass at 1000 Hz with gain_entry:
3733 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3736 custom equalization:
3738 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3741 higher delay with zero phase to compensate delay:
3743 firequalizer=delay=0.1:fixed=on:zero_phase=on
3746 lowpass on left channel, highpass on right channel:
3748 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3749 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3754 Apply a flanging effect to the audio.
3756 The filter accepts the following options:
3760 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3763 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3766 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3770 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3771 Default value is 71.
3774 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3777 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3778 Default value is @var{sinusoidal}.
3781 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3782 Default value is 25.
3785 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3786 Default is @var{linear}.
3790 Apply Haas effect to audio.
3792 Note that this makes most sense to apply on mono signals.
3793 With this filter applied to mono signals it give some directionality and
3794 stretches its stereo image.
3796 The filter accepts the following options:
3800 Set input level. By default is @var{1}, or 0dB
3803 Set output level. By default is @var{1}, or 0dB.
3806 Set gain applied to side part of signal. By default is @var{1}.
3809 Set kind of middle source. Can be one of the following:
3819 Pick middle part signal of stereo image.
3822 Pick side part signal of stereo image.
3826 Change middle phase. By default is disabled.
3829 Set left channel delay. By default is @var{2.05} milliseconds.
3832 Set left channel balance. By default is @var{-1}.
3835 Set left channel gain. By default is @var{1}.
3838 Change left phase. By default is disabled.
3841 Set right channel delay. By defaults is @var{2.12} milliseconds.
3844 Set right channel balance. By default is @var{1}.
3847 Set right channel gain. By default is @var{1}.
3850 Change right phase. By default is enabled.
3855 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3856 embedded HDCD codes is expanded into a 20-bit PCM stream.
3858 The filter supports the Peak Extend and Low-level Gain Adjustment features
3859 of HDCD, and detects the Transient Filter flag.
3862 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3865 When using the filter with wav, note the default encoding for wav is 16-bit,
3866 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3867 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3869 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3870 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3873 The filter accepts the following options:
3876 @item disable_autoconvert
3877 Disable any automatic format conversion or resampling in the filter graph.
3879 @item process_stereo
3880 Process the stereo channels together. If target_gain does not match between
3881 channels, consider it invalid and use the last valid target_gain.
3884 Set the code detect timer period in ms.
3887 Always extend peaks above -3dBFS even if PE isn't signaled.
3890 Replace audio with a solid tone and adjust the amplitude to signal some
3891 specific aspect of the decoding process. The output file can be loaded in
3892 an audio editor alongside the original to aid analysis.
3894 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3901 Gain adjustment level at each sample
3903 Samples where peak extend occurs
3905 Samples where the code detect timer is active
3907 Samples where the target gain does not match between channels
3913 Apply head-related transfer functions (HRTFs) to create virtual
3914 loudspeakers around the user for binaural listening via headphones.
3915 The HRIRs are provided via additional streams, for each channel
3916 one stereo input stream is needed.
3918 The filter accepts the following options:
3922 Set mapping of input streams for convolution.
3923 The argument is a '|'-separated list of channel names in order as they
3924 are given as additional stream inputs for filter.
3925 This also specify number of input streams. Number of input streams
3926 must be not less than number of channels in first stream plus one.
3929 Set gain applied to audio. Value is in dB. Default is 0.
3932 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3933 processing audio in time domain which is slow.
3934 @var{freq} is processing audio in frequency domain which is fast.
3935 Default is @var{freq}.
3938 Set custom gain for LFE channels. Value is in dB. Default is 0.
3941 Set size of frame in number of samples which will be processed at once.
3942 Default value is @var{1024}. Allowed range is from 1024 to 96000.
3945 Set format of hrir stream.
3946 Default value is @var{stereo}. Alternative value is @var{multich}.
3947 If value is set to @var{stereo}, number of additional streams should
3948 be greater or equal to number of input channels in first input stream.
3949 Also each additional stream should have stereo number of channels.
3950 If value is set to @var{multich}, number of additional streams should
3951 be exactly one. Also number of input channels of additional stream
3952 should be equal or greater than twice number of channels of first input
3956 @subsection Examples
3960 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3961 each amovie filter use stereo file with IR coefficients as input.
3962 The files give coefficients for each position of virtual loudspeaker:
3965 -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"
3970 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3971 but now in @var{multich} @var{hrir} format.
3973 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"
3980 Apply a high-pass filter with 3dB point frequency.
3981 The filter can be either single-pole, or double-pole (the default).
3982 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3984 The filter accepts the following options:
3988 Set frequency in Hz. Default is 3000.
3991 Set number of poles. Default is 2.
3994 Set method to specify band-width of filter.
4009 Specify the band-width of a filter in width_type units.
4010 Applies only to double-pole filter.
4011 The default is 0.707q and gives a Butterworth response.
4014 How much to use filtered signal in output. Default is 1.
4015 Range is between 0 and 1.
4018 Specify which channels to filter, by default all available are filtered.
4021 Normalize biquad coefficients, by default is disabled.
4022 Enabling it will normalize magnitude response at DC to 0dB.
4025 @subsection Commands
4027 This filter supports the following commands:
4030 Change highpass frequency.
4031 Syntax for the command is : "@var{frequency}"
4034 Change highpass width_type.
4035 Syntax for the command is : "@var{width_type}"
4038 Change highpass width.
4039 Syntax for the command is : "@var{width}"
4042 Change highpass mix.
4043 Syntax for the command is : "@var{mix}"
4048 Join multiple input streams into one multi-channel stream.
4050 It accepts the following parameters:
4054 The number of input streams. It defaults to 2.
4056 @item channel_layout
4057 The desired output channel layout. It defaults to stereo.
4060 Map channels from inputs to output. The argument is a '|'-separated list of
4061 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4062 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4063 can be either the name of the input channel (e.g. FL for front left) or its
4064 index in the specified input stream. @var{out_channel} is the name of the output
4068 The filter will attempt to guess the mappings when they are not specified
4069 explicitly. It does so by first trying to find an unused matching input channel
4070 and if that fails it picks the first unused input channel.
4072 Join 3 inputs (with properly set channel layouts):
4074 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4077 Build a 5.1 output from 6 single-channel streams:
4079 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4080 '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'
4086 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4088 To enable compilation of this filter you need to configure FFmpeg with
4089 @code{--enable-ladspa}.
4093 Specifies the name of LADSPA plugin library to load. If the environment
4094 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4095 each one of the directories specified by the colon separated list in
4096 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4097 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4098 @file{/usr/lib/ladspa/}.
4101 Specifies the plugin within the library. Some libraries contain only
4102 one plugin, but others contain many of them. If this is not set filter
4103 will list all available plugins within the specified library.
4106 Set the '|' separated list of controls which are zero or more floating point
4107 values that determine the behavior of the loaded plugin (for example delay,
4109 Controls need to be defined using the following syntax:
4110 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4111 @var{valuei} is the value set on the @var{i}-th control.
4112 Alternatively they can be also defined using the following syntax:
4113 @var{value0}|@var{value1}|@var{value2}|..., where
4114 @var{valuei} is the value set on the @var{i}-th control.
4115 If @option{controls} is set to @code{help}, all available controls and
4116 their valid ranges are printed.
4118 @item sample_rate, s
4119 Specify the sample rate, default to 44100. Only used if plugin have
4123 Set the number of samples per channel per each output frame, default
4124 is 1024. Only used if plugin have zero inputs.
4127 Set the minimum duration of the sourced audio. See
4128 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4129 for the accepted syntax.
4130 Note that the resulting duration may be greater than the specified duration,
4131 as the generated audio is always cut at the end of a complete frame.
4132 If not specified, or the expressed duration is negative, the audio is
4133 supposed to be generated forever.
4134 Only used if plugin have zero inputs.
4138 @subsection Examples
4142 List all available plugins within amp (LADSPA example plugin) library:
4148 List all available controls and their valid ranges for @code{vcf_notch}
4149 plugin from @code{VCF} library:
4151 ladspa=f=vcf:p=vcf_notch:c=help
4155 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4158 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4162 Add reverberation to the audio using TAP-plugins
4163 (Tom's Audio Processing plugins):
4165 ladspa=file=tap_reverb:tap_reverb
4169 Generate white noise, with 0.2 amplitude:
4171 ladspa=file=cmt:noise_source_white:c=c0=.2
4175 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4176 @code{C* Audio Plugin Suite} (CAPS) library:
4178 ladspa=file=caps:Click:c=c1=20'
4182 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4184 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4188 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4189 @code{SWH Plugins} collection:
4191 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4195 Attenuate low frequencies using Multiband EQ from Steve Harris
4196 @code{SWH Plugins} collection:
4198 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4202 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4205 ladspa=caps:Narrower
4209 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4211 ladspa=caps:White:.2
4215 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4217 ladspa=caps:Fractal:c=c1=1
4221 Dynamic volume normalization using @code{VLevel} plugin:
4223 ladspa=vlevel-ladspa:vlevel_mono
4227 @subsection Commands
4229 This filter supports the following commands:
4232 Modify the @var{N}-th control value.
4234 If the specified value is not valid, it is ignored and prior one is kept.
4239 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4240 Support for both single pass (livestreams, files) and double pass (files) modes.
4241 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4242 detect true peaks, the audio stream will be upsampled to 192 kHz.
4243 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4245 The filter accepts the following options:
4249 Set integrated loudness target.
4250 Range is -70.0 - -5.0. Default value is -24.0.
4253 Set loudness range target.
4254 Range is 1.0 - 20.0. Default value is 7.0.
4257 Set maximum true peak.
4258 Range is -9.0 - +0.0. Default value is -2.0.
4260 @item measured_I, measured_i
4261 Measured IL of input file.
4262 Range is -99.0 - +0.0.
4264 @item measured_LRA, measured_lra
4265 Measured LRA of input file.
4266 Range is 0.0 - 99.0.
4268 @item measured_TP, measured_tp
4269 Measured true peak of input file.
4270 Range is -99.0 - +99.0.
4272 @item measured_thresh
4273 Measured threshold of input file.
4274 Range is -99.0 - +0.0.
4277 Set offset gain. Gain is applied before the true-peak limiter.
4278 Range is -99.0 - +99.0. Default is +0.0.
4281 Normalize by linearly scaling the source audio.
4282 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4283 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4284 be lower than source LRA and the change in integrated loudness shouldn't
4285 result in a true peak which exceeds the target TP. If any of these
4286 conditions aren't met, normalization mode will revert to @var{dynamic}.
4287 Options are @code{true} or @code{false}. Default is @code{true}.
4290 Treat mono input files as "dual-mono". If a mono file is intended for playback
4291 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4292 If set to @code{true}, this option will compensate for this effect.
4293 Multi-channel input files are not affected by this option.
4294 Options are true or false. Default is false.
4297 Set print format for stats. Options are summary, json, or none.
4298 Default value is none.
4303 Apply a low-pass filter with 3dB point frequency.
4304 The filter can be either single-pole or double-pole (the default).
4305 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4307 The filter accepts the following options:
4311 Set frequency in Hz. Default is 500.
4314 Set number of poles. Default is 2.
4317 Set method to specify band-width of filter.
4332 Specify the band-width of a filter in width_type units.
4333 Applies only to double-pole filter.
4334 The default is 0.707q and gives a Butterworth response.
4337 How much to use filtered signal in output. Default is 1.
4338 Range is between 0 and 1.
4341 Specify which channels to filter, by default all available are filtered.
4344 Normalize biquad coefficients, by default is disabled.
4345 Enabling it will normalize magnitude response at DC to 0dB.
4348 @subsection Examples
4351 Lowpass only LFE channel, it LFE is not present it does nothing:
4357 @subsection Commands
4359 This filter supports the following commands:
4362 Change lowpass frequency.
4363 Syntax for the command is : "@var{frequency}"
4366 Change lowpass width_type.
4367 Syntax for the command is : "@var{width_type}"
4370 Change lowpass width.
4371 Syntax for the command is : "@var{width}"
4375 Syntax for the command is : "@var{mix}"
4380 Load a LV2 (LADSPA Version 2) plugin.
4382 To enable compilation of this filter you need to configure FFmpeg with
4383 @code{--enable-lv2}.
4387 Specifies the plugin URI. You may need to escape ':'.
4390 Set the '|' separated list of controls which are zero or more floating point
4391 values that determine the behavior of the loaded plugin (for example delay,
4393 If @option{controls} is set to @code{help}, all available controls and
4394 their valid ranges are printed.
4396 @item sample_rate, s
4397 Specify the sample rate, default to 44100. Only used if plugin have
4401 Set the number of samples per channel per each output frame, default
4402 is 1024. Only used if plugin have zero inputs.
4405 Set the minimum duration of the sourced audio. See
4406 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4407 for the accepted syntax.
4408 Note that the resulting duration may be greater than the specified duration,
4409 as the generated audio is always cut at the end of a complete frame.
4410 If not specified, or the expressed duration is negative, the audio is
4411 supposed to be generated forever.
4412 Only used if plugin have zero inputs.
4415 @subsection Examples
4419 Apply bass enhancer plugin from Calf:
4421 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4425 Apply vinyl plugin from Calf:
4427 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4431 Apply bit crusher plugin from ArtyFX:
4433 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4438 Multiband Compress or expand the audio's dynamic range.
4440 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4441 This is akin to the crossover of a loudspeaker, and results in flat frequency
4442 response when absent compander action.
4444 It accepts the following parameters:
4448 This option syntax is:
4449 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4450 For explanation of each item refer to compand filter documentation.
4456 Mix channels with specific gain levels. The filter accepts the output
4457 channel layout followed by a set of channels definitions.
4459 This filter is also designed to efficiently remap the channels of an audio
4462 The filter accepts parameters of the form:
4463 "@var{l}|@var{outdef}|@var{outdef}|..."
4467 output channel layout or number of channels
4470 output channel specification, of the form:
4471 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4474 output channel to define, either a channel name (FL, FR, etc.) or a channel
4475 number (c0, c1, etc.)
4478 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4481 input channel to use, see out_name for details; it is not possible to mix
4482 named and numbered input channels
4485 If the `=' in a channel specification is replaced by `<', then the gains for
4486 that specification will be renormalized so that the total is 1, thus
4487 avoiding clipping noise.
4489 @subsection Mixing examples
4491 For example, if you want to down-mix from stereo to mono, but with a bigger
4492 factor for the left channel:
4494 pan=1c|c0=0.9*c0+0.1*c1
4497 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4498 7-channels surround:
4500 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4503 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4504 that should be preferred (see "-ac" option) unless you have very specific
4507 @subsection Remapping examples
4509 The channel remapping will be effective if, and only if:
4512 @item gain coefficients are zeroes or ones,
4513 @item only one input per channel output,
4516 If all these conditions are satisfied, the filter will notify the user ("Pure
4517 channel mapping detected"), and use an optimized and lossless method to do the
4520 For example, if you have a 5.1 source and want a stereo audio stream by
4521 dropping the extra channels:
4523 pan="stereo| c0=FL | c1=FR"
4526 Given the same source, you can also switch front left and front right channels
4527 and keep the input channel layout:
4529 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4532 If the input is a stereo audio stream, you can mute the front left channel (and
4533 still keep the stereo channel layout) with:
4538 Still with a stereo audio stream input, you can copy the right channel in both
4539 front left and right:
4541 pan="stereo| c0=FR | c1=FR"
4546 ReplayGain scanner filter. This filter takes an audio stream as an input and
4547 outputs it unchanged.
4548 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4552 Convert the audio sample format, sample rate and channel layout. It is
4553 not meant to be used directly.
4556 Apply time-stretching and pitch-shifting with librubberband.
4558 To enable compilation of this filter, you need to configure FFmpeg with
4559 @code{--enable-librubberband}.
4561 The filter accepts the following options:
4565 Set tempo scale factor.
4568 Set pitch scale factor.
4571 Set transients detector.
4572 Possible values are:
4581 Possible values are:
4590 Possible values are:
4597 Set processing window size.
4598 Possible values are:
4607 Possible values are:
4614 Enable formant preservation when shift pitching.
4615 Possible values are:
4623 Possible values are:
4632 Possible values are:
4639 @subsection Commands
4641 This filter supports the following commands:
4644 Change filter tempo scale factor.
4645 Syntax for the command is : "@var{tempo}"
4648 Change filter pitch scale factor.
4649 Syntax for the command is : "@var{pitch}"
4652 @section sidechaincompress
4654 This filter acts like normal compressor but has the ability to compress
4655 detected signal using second input signal.
4656 It needs two input streams and returns one output stream.
4657 First input stream will be processed depending on second stream signal.
4658 The filtered signal then can be filtered with other filters in later stages of
4659 processing. See @ref{pan} and @ref{amerge} filter.
4661 The filter accepts the following options:
4665 Set input gain. Default is 1. Range is between 0.015625 and 64.
4668 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4669 Default is @code{downward}.
4672 If a signal of second stream raises above this level it will affect the gain
4673 reduction of first stream.
4674 By default is 0.125. Range is between 0.00097563 and 1.
4677 Set a ratio about which the signal is reduced. 1:2 means that if the level
4678 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4679 Default is 2. Range is between 1 and 20.
4682 Amount of milliseconds the signal has to rise above the threshold before gain
4683 reduction starts. Default is 20. Range is between 0.01 and 2000.
4686 Amount of milliseconds the signal has to fall below the threshold before
4687 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4690 Set the amount by how much signal will be amplified after processing.
4691 Default is 1. Range is from 1 to 64.
4694 Curve the sharp knee around the threshold to enter gain reduction more softly.
4695 Default is 2.82843. Range is between 1 and 8.
4698 Choose if the @code{average} level between all channels of side-chain stream
4699 or the louder(@code{maximum}) channel of side-chain stream affects the
4700 reduction. Default is @code{average}.
4703 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4704 of @code{rms}. Default is @code{rms} which is mainly smoother.
4707 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4710 How much to use compressed signal in output. Default is 1.
4711 Range is between 0 and 1.
4714 @subsection Commands
4716 This filter supports the all above options as @ref{commands}.
4718 @subsection Examples
4722 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4723 depending on the signal of 2nd input and later compressed signal to be
4724 merged with 2nd input:
4726 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4730 @section sidechaingate
4732 A sidechain gate acts like a normal (wideband) gate but has the ability to
4733 filter the detected signal before sending it to the gain reduction stage.
4734 Normally a gate uses the full range signal to detect a level above the
4736 For example: If you cut all lower frequencies from your sidechain signal
4737 the gate will decrease the volume of your track only if not enough highs
4738 appear. With this technique you are able to reduce the resonation of a
4739 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4741 It needs two input streams and returns one output stream.
4742 First input stream will be processed depending on second stream signal.
4744 The filter accepts the following options:
4748 Set input level before filtering.
4749 Default is 1. Allowed range is from 0.015625 to 64.
4752 Set the mode of operation. Can be @code{upward} or @code{downward}.
4753 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
4754 will be amplified, expanding dynamic range in upward direction.
4755 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
4758 Set the level of gain reduction when the signal is below the threshold.
4759 Default is 0.06125. Allowed range is from 0 to 1.
4760 Setting this to 0 disables reduction and then filter behaves like expander.
4763 If a signal rises above this level the gain reduction is released.
4764 Default is 0.125. Allowed range is from 0 to 1.
4767 Set a ratio about which the signal is reduced.
4768 Default is 2. Allowed range is from 1 to 9000.
4771 Amount of milliseconds the signal has to rise above the threshold before gain
4773 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4776 Amount of milliseconds the signal has to fall below the threshold before the
4777 reduction is increased again. Default is 250 milliseconds.
4778 Allowed range is from 0.01 to 9000.
4781 Set amount of amplification of signal after processing.
4782 Default is 1. Allowed range is from 1 to 64.
4785 Curve the sharp knee around the threshold to enter gain reduction more softly.
4786 Default is 2.828427125. Allowed range is from 1 to 8.
4789 Choose if exact signal should be taken for detection or an RMS like one.
4790 Default is rms. Can be peak or rms.
4793 Choose if the average level between all channels or the louder channel affects
4795 Default is average. Can be average or maximum.
4798 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4801 @section silencedetect
4803 Detect silence in an audio stream.
4805 This filter logs a message when it detects that the input audio volume is less
4806 or equal to a noise tolerance value for a duration greater or equal to the
4807 minimum detected noise duration.
4809 The printed times and duration are expressed in seconds. The
4810 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
4811 is set on the first frame whose timestamp equals or exceeds the detection
4812 duration and it contains the timestamp of the first frame of the silence.
4814 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
4815 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
4816 keys are set on the first frame after the silence. If @option{mono} is
4817 enabled, and each channel is evaluated separately, the @code{.X}
4818 suffixed keys are used, and @code{X} corresponds to the channel number.
4820 The filter accepts the following options:
4824 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4825 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4828 Set silence duration until notification (default is 2 seconds). See
4829 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4830 for the accepted syntax.
4833 Process each channel separately, instead of combined. By default is disabled.
4836 @subsection Examples
4840 Detect 5 seconds of silence with -50dB noise tolerance:
4842 silencedetect=n=-50dB:d=5
4846 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4847 tolerance in @file{silence.mp3}:
4849 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4853 @section silenceremove
4855 Remove silence from the beginning, middle or end of the audio.
4857 The filter accepts the following options:
4861 This value is used to indicate if audio should be trimmed at beginning of
4862 the audio. A value of zero indicates no silence should be trimmed from the
4863 beginning. When specifying a non-zero value, it trims audio up until it
4864 finds non-silence. Normally, when trimming silence from beginning of audio
4865 the @var{start_periods} will be @code{1} but it can be increased to higher
4866 values to trim all audio up to specific count of non-silence periods.
4867 Default value is @code{0}.
4869 @item start_duration
4870 Specify the amount of time that non-silence must be detected before it stops
4871 trimming audio. By increasing the duration, bursts of noises can be treated
4872 as silence and trimmed off. Default value is @code{0}.
4874 @item start_threshold
4875 This indicates what sample value should be treated as silence. For digital
4876 audio, a value of @code{0} may be fine but for audio recorded from analog,
4877 you may wish to increase the value to account for background noise.
4878 Can be specified in dB (in case "dB" is appended to the specified value)
4879 or amplitude ratio. Default value is @code{0}.
4882 Specify max duration of silence at beginning that will be kept after
4883 trimming. Default is 0, which is equal to trimming all samples detected
4887 Specify mode of detection of silence end in start of multi-channel audio.
4888 Can be @var{any} or @var{all}. Default is @var{any}.
4889 With @var{any}, any sample that is detected as non-silence will cause
4890 stopped trimming of silence.
4891 With @var{all}, only if all channels are detected as non-silence will cause
4892 stopped trimming of silence.
4895 Set the count for trimming silence from the end of audio.
4896 To remove silence from the middle of a file, specify a @var{stop_periods}
4897 that is negative. This value is then treated as a positive value and is
4898 used to indicate the effect should restart processing as specified by
4899 @var{start_periods}, making it suitable for removing periods of silence
4900 in the middle of the audio.
4901 Default value is @code{0}.
4904 Specify a duration of silence that must exist before audio is not copied any
4905 more. By specifying a higher duration, silence that is wanted can be left in
4907 Default value is @code{0}.
4909 @item stop_threshold
4910 This is the same as @option{start_threshold} but for trimming silence from
4912 Can be specified in dB (in case "dB" is appended to the specified value)
4913 or amplitude ratio. Default value is @code{0}.
4916 Specify max duration of silence at end that will be kept after
4917 trimming. Default is 0, which is equal to trimming all samples detected
4921 Specify mode of detection of silence start in end of multi-channel audio.
4922 Can be @var{any} or @var{all}. Default is @var{any}.
4923 With @var{any}, any sample that is detected as non-silence will cause
4924 stopped trimming of silence.
4925 With @var{all}, only if all channels are detected as non-silence will cause
4926 stopped trimming of silence.
4929 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4930 and works better with digital silence which is exactly 0.
4931 Default value is @code{rms}.
4934 Set duration in number of seconds used to calculate size of window in number
4935 of samples for detecting silence.
4936 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4939 @subsection Examples
4943 The following example shows how this filter can be used to start a recording
4944 that does not contain the delay at the start which usually occurs between
4945 pressing the record button and the start of the performance:
4947 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
4951 Trim all silence encountered from beginning to end where there is more than 1
4952 second of silence in audio:
4954 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
4958 Trim all digital silence samples, using peak detection, from beginning to end
4959 where there is more than 0 samples of digital silence in audio and digital
4960 silence is detected in all channels at same positions in stream:
4962 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
4968 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4969 loudspeakers around the user for binaural listening via headphones (audio
4970 formats up to 9 channels supported).
4971 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4972 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4973 Austrian Academy of Sciences.
4975 To enable compilation of this filter you need to configure FFmpeg with
4976 @code{--enable-libmysofa}.
4978 The filter accepts the following options:
4982 Set the SOFA file used for rendering.
4985 Set gain applied to audio. Value is in dB. Default is 0.
4988 Set rotation of virtual loudspeakers in deg. Default is 0.
4991 Set elevation of virtual speakers in deg. Default is 0.
4994 Set distance in meters between loudspeakers and the listener with near-field
4995 HRTFs. Default is 1.
4998 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4999 processing audio in time domain which is slow.
5000 @var{freq} is processing audio in frequency domain which is fast.
5001 Default is @var{freq}.
5004 Set custom positions of virtual loudspeakers. Syntax for this option is:
5005 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5006 Each virtual loudspeaker is described with short channel name following with
5007 azimuth and elevation in degrees.
5008 Each virtual loudspeaker description is separated by '|'.
5009 For example to override front left and front right channel positions use:
5010 'speakers=FL 45 15|FR 345 15'.
5011 Descriptions with unrecognised channel names are ignored.
5014 Set custom gain for LFE channels. Value is in dB. Default is 0.
5017 Set custom frame size in number of samples. Default is 1024.
5018 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5019 is set to @var{freq}.
5022 Should all IRs be normalized upon importing SOFA file.
5023 By default is enabled.
5026 Should nearest IRs be interpolated with neighbor IRs if exact position
5027 does not match. By default is disabled.
5030 Minphase all IRs upon loading of SOFA file. By default is disabled.
5033 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5036 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5039 @subsection Examples
5043 Using ClubFritz6 sofa file:
5045 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5049 Using ClubFritz12 sofa file and bigger radius with small rotation:
5051 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5055 Similar as above but with custom speaker positions for front left, front right, back left and back right
5056 and also with custom gain:
5058 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5062 @section stereotools
5064 This filter has some handy utilities to manage stereo signals, for converting
5065 M/S stereo recordings to L/R signal while having control over the parameters
5066 or spreading the stereo image of master track.
5068 The filter accepts the following options:
5072 Set input level before filtering for both channels. Defaults is 1.
5073 Allowed range is from 0.015625 to 64.
5076 Set output level after filtering for both channels. Defaults is 1.
5077 Allowed range is from 0.015625 to 64.
5080 Set input balance between both channels. Default is 0.
5081 Allowed range is from -1 to 1.
5084 Set output balance between both channels. Default is 0.
5085 Allowed range is from -1 to 1.
5088 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5089 clipping. Disabled by default.
5092 Mute the left channel. Disabled by default.
5095 Mute the right channel. Disabled by default.
5098 Change the phase of the left channel. Disabled by default.
5101 Change the phase of the right channel. Disabled by default.
5104 Set stereo mode. Available values are:
5108 Left/Right to Left/Right, this is default.
5111 Left/Right to Mid/Side.
5114 Mid/Side to Left/Right.
5117 Left/Right to Left/Left.
5120 Left/Right to Right/Right.
5123 Left/Right to Left + Right.
5126 Left/Right to Right/Left.
5129 Mid/Side to Left/Left.
5132 Mid/Side to Right/Right.
5136 Set level of side signal. Default is 1.
5137 Allowed range is from 0.015625 to 64.
5140 Set balance of side signal. Default is 0.
5141 Allowed range is from -1 to 1.
5144 Set level of the middle signal. Default is 1.
5145 Allowed range is from 0.015625 to 64.
5148 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5151 Set stereo base between mono and inversed channels. Default is 0.
5152 Allowed range is from -1 to 1.
5155 Set delay in milliseconds how much to delay left from right channel and
5156 vice versa. Default is 0. Allowed range is from -20 to 20.
5159 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5162 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5164 @item bmode_in, bmode_out
5165 Set balance mode for balance_in/balance_out option.
5167 Can be one of the following:
5171 Classic balance mode. Attenuate one channel at time.
5172 Gain is raised up to 1.
5175 Similar as classic mode above but gain is raised up to 2.
5178 Equal power distribution, from -6dB to +6dB range.
5182 @subsection Examples
5186 Apply karaoke like effect:
5188 stereotools=mlev=0.015625
5192 Convert M/S signal to L/R:
5194 "stereotools=mode=ms>lr"
5198 @section stereowiden
5200 This filter enhance the stereo effect by suppressing signal common to both
5201 channels and by delaying the signal of left into right and vice versa,
5202 thereby widening the stereo effect.
5204 The filter accepts the following options:
5208 Time in milliseconds of the delay of left signal into right and vice versa.
5209 Default is 20 milliseconds.
5212 Amount of gain in delayed signal into right and vice versa. Gives a delay
5213 effect of left signal in right output and vice versa which gives widening
5214 effect. Default is 0.3.
5217 Cross feed of left into right with inverted phase. This helps in suppressing
5218 the mono. If the value is 1 it will cancel all the signal common to both
5219 channels. Default is 0.3.
5222 Set level of input signal of original channel. Default is 0.8.
5225 @subsection Commands
5227 This filter supports the all above options except @code{delay} as @ref{commands}.
5229 @section superequalizer
5230 Apply 18 band equalizer.
5232 The filter accepts the following options:
5239 Set 131Hz band gain.
5241 Set 185Hz band gain.
5243 Set 262Hz band gain.
5245 Set 370Hz band gain.
5247 Set 523Hz band gain.
5249 Set 740Hz band gain.
5251 Set 1047Hz band gain.
5253 Set 1480Hz band gain.
5255 Set 2093Hz band gain.
5257 Set 2960Hz band gain.
5259 Set 4186Hz band gain.
5261 Set 5920Hz band gain.
5263 Set 8372Hz band gain.
5265 Set 11840Hz band gain.
5267 Set 16744Hz band gain.
5269 Set 20000Hz band gain.
5273 Apply audio surround upmix filter.
5275 This filter allows to produce multichannel output from audio stream.
5277 The filter accepts the following options:
5281 Set output channel layout. By default, this is @var{5.1}.
5283 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5284 for the required syntax.
5287 Set input channel layout. By default, this is @var{stereo}.
5289 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5290 for the required syntax.
5293 Set input volume level. By default, this is @var{1}.
5296 Set output volume level. By default, this is @var{1}.
5299 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5302 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5305 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5308 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5309 In @var{add} mode, LFE channel is created from input audio and added to output.
5310 In @var{sub} mode, LFE channel is created from input audio and added to output but
5311 also all non-LFE output channels are subtracted with output LFE channel.
5314 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5315 Default is @var{90}.
5318 Set front center input volume. By default, this is @var{1}.
5321 Set front center output volume. By default, this is @var{1}.
5324 Set front left input volume. By default, this is @var{1}.
5327 Set front left output volume. By default, this is @var{1}.
5330 Set front right input volume. By default, this is @var{1}.
5333 Set front right output volume. By default, this is @var{1}.
5336 Set side left input volume. By default, this is @var{1}.
5339 Set side left output volume. By default, this is @var{1}.
5342 Set side right input volume. By default, this is @var{1}.
5345 Set side right output volume. By default, this is @var{1}.
5348 Set back left input volume. By default, this is @var{1}.
5351 Set back left output volume. By default, this is @var{1}.
5354 Set back right input volume. By default, this is @var{1}.
5357 Set back right output volume. By default, this is @var{1}.
5360 Set back center input volume. By default, this is @var{1}.
5363 Set back center output volume. By default, this is @var{1}.
5366 Set LFE input volume. By default, this is @var{1}.
5369 Set LFE output volume. By default, this is @var{1}.
5372 Set spread usage of stereo image across X axis for all channels.
5375 Set spread usage of stereo image across Y axis for all channels.
5377 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5378 Set spread usage of stereo image across X axis for each channel.
5380 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5381 Set spread usage of stereo image across Y axis for each channel.
5384 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5387 Set window function.
5389 It accepts the following values:
5412 Default is @code{hann}.
5415 Set window overlap. If set to 1, the recommended overlap for selected
5416 window function will be picked. Default is @code{0.5}.
5419 @section treble, highshelf
5421 Boost or cut treble (upper) frequencies of the audio using a two-pole
5422 shelving filter with a response similar to that of a standard
5423 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5425 The filter accepts the following options:
5429 Give the gain at whichever is the lower of ~22 kHz and the
5430 Nyquist frequency. Its useful range is about -20 (for a large cut)
5431 to +20 (for a large boost). Beware of clipping when using a positive gain.
5434 Set the filter's central frequency and so can be used
5435 to extend or reduce the frequency range to be boosted or cut.
5436 The default value is @code{3000} Hz.
5439 Set method to specify band-width of filter.
5454 Determine how steep is the filter's shelf transition.
5457 How much to use filtered signal in output. Default is 1.
5458 Range is between 0 and 1.
5461 Specify which channels to filter, by default all available are filtered.
5464 Normalize biquad coefficients, by default is disabled.
5465 Enabling it will normalize magnitude response at DC to 0dB.
5468 @subsection Commands
5470 This filter supports the following commands:
5473 Change treble frequency.
5474 Syntax for the command is : "@var{frequency}"
5477 Change treble width_type.
5478 Syntax for the command is : "@var{width_type}"
5481 Change treble width.
5482 Syntax for the command is : "@var{width}"
5486 Syntax for the command is : "@var{gain}"
5490 Syntax for the command is : "@var{mix}"
5495 Sinusoidal amplitude modulation.
5497 The filter accepts the following options:
5501 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5502 (20 Hz or lower) will result in a tremolo effect.
5503 This filter may also be used as a ring modulator by specifying
5504 a modulation frequency higher than 20 Hz.
5505 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5508 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5509 Default value is 0.5.
5514 Sinusoidal phase modulation.
5516 The filter accepts the following options:
5520 Modulation frequency in Hertz.
5521 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5524 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5525 Default value is 0.5.
5530 Adjust the input audio volume.
5532 It accepts the following parameters:
5536 Set audio volume expression.
5538 Output values are clipped to the maximum value.
5540 The output audio volume is given by the relation:
5542 @var{output_volume} = @var{volume} * @var{input_volume}
5545 The default value for @var{volume} is "1.0".
5548 This parameter represents the mathematical precision.
5550 It determines which input sample formats will be allowed, which affects the
5551 precision of the volume scaling.
5555 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5557 32-bit floating-point; this limits input sample format to FLT. (default)
5559 64-bit floating-point; this limits input sample format to DBL.
5563 Choose the behaviour on encountering ReplayGain side data in input frames.
5567 Remove ReplayGain side data, ignoring its contents (the default).
5570 Ignore ReplayGain side data, but leave it in the frame.
5573 Prefer the track gain, if present.
5576 Prefer the album gain, if present.
5579 @item replaygain_preamp
5580 Pre-amplification gain in dB to apply to the selected replaygain gain.
5582 Default value for @var{replaygain_preamp} is 0.0.
5584 @item replaygain_noclip
5585 Prevent clipping by limiting the gain applied.
5587 Default value for @var{replaygain_noclip} is 1.
5590 Set when the volume expression is evaluated.
5592 It accepts the following values:
5595 only evaluate expression once during the filter initialization, or
5596 when the @samp{volume} command is sent
5599 evaluate expression for each incoming frame
5602 Default value is @samp{once}.
5605 The volume expression can contain the following parameters.
5609 frame number (starting at zero)
5612 @item nb_consumed_samples
5613 number of samples consumed by the filter
5615 number of samples in the current frame
5617 original frame position in the file
5623 PTS at start of stream
5625 time at start of stream
5631 last set volume value
5634 Note that when @option{eval} is set to @samp{once} only the
5635 @var{sample_rate} and @var{tb} variables are available, all other
5636 variables will evaluate to NAN.
5638 @subsection Commands
5640 This filter supports the following commands:
5643 Modify the volume expression.
5644 The command accepts the same syntax of the corresponding option.
5646 If the specified expression is not valid, it is kept at its current
5650 @subsection Examples
5654 Halve the input audio volume:
5658 volume=volume=-6.0206dB
5661 In all the above example the named key for @option{volume} can be
5662 omitted, for example like in:
5668 Increase input audio power by 6 decibels using fixed-point precision:
5670 volume=volume=6dB:precision=fixed
5674 Fade volume after time 10 with an annihilation period of 5 seconds:
5676 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5680 @section volumedetect
5682 Detect the volume of the input video.
5684 The filter has no parameters. The input is not modified. Statistics about
5685 the volume will be printed in the log when the input stream end is reached.
5687 In particular it will show the mean volume (root mean square), maximum
5688 volume (on a per-sample basis), and the beginning of a histogram of the
5689 registered volume values (from the maximum value to a cumulated 1/1000 of
5692 All volumes are in decibels relative to the maximum PCM value.
5694 @subsection Examples
5696 Here is an excerpt of the output:
5698 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5699 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5700 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5701 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5702 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5703 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5704 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5705 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5706 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5712 The mean square energy is approximately -27 dB, or 10^-2.7.
5714 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5716 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5719 In other words, raising the volume by +4 dB does not cause any clipping,
5720 raising it by +5 dB causes clipping for 6 samples, etc.
5722 @c man end AUDIO FILTERS
5724 @chapter Audio Sources
5725 @c man begin AUDIO SOURCES
5727 Below is a description of the currently available audio sources.
5731 Buffer audio frames, and make them available to the filter chain.
5733 This source is mainly intended for a programmatic use, in particular
5734 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5736 It accepts the following parameters:
5740 The timebase which will be used for timestamps of submitted frames. It must be
5741 either a floating-point number or in @var{numerator}/@var{denominator} form.
5744 The sample rate of the incoming audio buffers.
5747 The sample format of the incoming audio buffers.
5748 Either a sample format name or its corresponding integer representation from
5749 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5751 @item channel_layout
5752 The channel layout of the incoming audio buffers.
5753 Either a channel layout name from channel_layout_map in
5754 @file{libavutil/channel_layout.c} or its corresponding integer representation
5755 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5758 The number of channels of the incoming audio buffers.
5759 If both @var{channels} and @var{channel_layout} are specified, then they
5764 @subsection Examples
5767 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5770 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5771 Since the sample format with name "s16p" corresponds to the number
5772 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5775 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5780 Generate an audio signal specified by an expression.
5782 This source accepts in input one or more expressions (one for each
5783 channel), which are evaluated and used to generate a corresponding
5786 This source accepts the following options:
5790 Set the '|'-separated expressions list for each separate channel. In case the
5791 @option{channel_layout} option is not specified, the selected channel layout
5792 depends on the number of provided expressions. Otherwise the last
5793 specified expression is applied to the remaining output channels.
5795 @item channel_layout, c
5796 Set the channel layout. The number of channels in the specified layout
5797 must be equal to the number of specified expressions.
5800 Set the minimum duration of the sourced audio. See
5801 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5802 for the accepted syntax.
5803 Note that the resulting duration may be greater than the specified
5804 duration, as the generated audio is always cut at the end of a
5807 If not specified, or the expressed duration is negative, the audio is
5808 supposed to be generated forever.
5811 Set the number of samples per channel per each output frame,
5814 @item sample_rate, s
5815 Specify the sample rate, default to 44100.
5818 Each expression in @var{exprs} can contain the following constants:
5822 number of the evaluated sample, starting from 0
5825 time of the evaluated sample expressed in seconds, starting from 0
5832 @subsection Examples
5842 Generate a sin signal with frequency of 440 Hz, set sample rate to
5845 aevalsrc="sin(440*2*PI*t):s=8000"
5849 Generate a two channels signal, specify the channel layout (Front
5850 Center + Back Center) explicitly:
5852 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5856 Generate white noise:
5858 aevalsrc="-2+random(0)"
5862 Generate an amplitude modulated signal:
5864 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5868 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5870 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5877 Generate a FIR coefficients using frequency sampling method.
5879 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
5881 The filter accepts the following options:
5885 Set number of filter coefficents in output audio stream.
5886 Default value is 1025.
5889 Set frequency points from where magnitude and phase are set.
5890 This must be in non decreasing order, and first element must be 0, while last element
5891 must be 1. Elements are separated by white spaces.
5894 Set magnitude value for every frequency point set by @option{frequency}.
5895 Number of values must be same as number of frequency points.
5896 Values are separated by white spaces.
5899 Set phase value for every frequency point set by @option{frequency}.
5900 Number of values must be same as number of frequency points.
5901 Values are separated by white spaces.
5903 @item sample_rate, r
5904 Set sample rate, default is 44100.
5907 Set number of samples per each frame. Default is 1024.
5910 Set window function. Default is blackman.
5915 The null audio source, return unprocessed audio frames. It is mainly useful
5916 as a template and to be employed in analysis / debugging tools, or as
5917 the source for filters which ignore the input data (for example the sox
5920 This source accepts the following options:
5924 @item channel_layout, cl
5926 Specifies the channel layout, and can be either an integer or a string
5927 representing a channel layout. The default value of @var{channel_layout}
5930 Check the channel_layout_map definition in
5931 @file{libavutil/channel_layout.c} for the mapping between strings and
5932 channel layout values.
5934 @item sample_rate, r
5935 Specifies the sample rate, and defaults to 44100.
5938 Set the number of samples per requested frames.
5942 @subsection Examples
5946 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5948 anullsrc=r=48000:cl=4
5952 Do the same operation with a more obvious syntax:
5954 anullsrc=r=48000:cl=mono
5958 All the parameters need to be explicitly defined.
5962 Synthesize a voice utterance using the libflite library.
5964 To enable compilation of this filter you need to configure FFmpeg with
5965 @code{--enable-libflite}.
5967 Note that versions of the flite library prior to 2.0 are not thread-safe.
5969 The filter accepts the following options:
5974 If set to 1, list the names of the available voices and exit
5975 immediately. Default value is 0.
5978 Set the maximum number of samples per frame. Default value is 512.
5981 Set the filename containing the text to speak.
5984 Set the text to speak.
5987 Set the voice to use for the speech synthesis. Default value is
5988 @code{kal}. See also the @var{list_voices} option.
5991 @subsection Examples
5995 Read from file @file{speech.txt}, and synthesize the text using the
5996 standard flite voice:
5998 flite=textfile=speech.txt
6002 Read the specified text selecting the @code{slt} voice:
6004 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6008 Input text to ffmpeg:
6010 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6014 Make @file{ffplay} speak the specified text, using @code{flite} and
6015 the @code{lavfi} device:
6017 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6021 For more information about libflite, check:
6022 @url{http://www.festvox.org/flite/}
6026 Generate a noise audio signal.
6028 The filter accepts the following options:
6031 @item sample_rate, r
6032 Specify the sample rate. Default value is 48000 Hz.
6035 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6039 Specify the duration of the generated audio stream. Not specifying this option
6040 results in noise with an infinite length.
6042 @item color, colour, c
6043 Specify the color of noise. Available noise colors are white, pink, brown,
6044 blue, violet and velvet. Default color is white.
6047 Specify a value used to seed the PRNG.
6050 Set the number of samples per each output frame, default is 1024.
6053 @subsection Examples
6058 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6060 anoisesrc=d=60:c=pink:r=44100:a=0.5
6066 Generate odd-tap Hilbert transform FIR coefficients.
6068 The resulting stream can be used with @ref{afir} filter for phase-shifting
6069 the signal by 90 degrees.
6071 This is used in many matrix coding schemes and for analytic signal generation.
6072 The process is often written as a multiplication by i (or j), the imaginary unit.
6074 The filter accepts the following options:
6078 @item sample_rate, s
6079 Set sample rate, default is 44100.
6082 Set length of FIR filter, default is 22051.
6085 Set number of samples per each frame.
6088 Set window function to be used when generating FIR coefficients.
6093 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6095 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6097 The filter accepts the following options:
6100 @item sample_rate, r
6101 Set sample rate, default is 44100.
6104 Set number of samples per each frame. Default is 1024.
6107 Set high-pass frequency. Default is 0.
6110 Set low-pass frequency. Default is 0.
6111 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6112 is higher than 0 then filter will create band-pass filter coefficients,
6113 otherwise band-reject filter coefficients.
6116 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6119 Set Kaiser window beta.
6122 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6125 Enable rounding, by default is disabled.
6128 Set number of taps for high-pass filter.
6131 Set number of taps for low-pass filter.
6136 Generate an audio signal made of a sine wave with amplitude 1/8.
6138 The audio signal is bit-exact.
6140 The filter accepts the following options:
6145 Set the carrier frequency. Default is 440 Hz.
6147 @item beep_factor, b
6148 Enable a periodic beep every second with frequency @var{beep_factor} times
6149 the carrier frequency. Default is 0, meaning the beep is disabled.
6151 @item sample_rate, r
6152 Specify the sample rate, default is 44100.
6155 Specify the duration of the generated audio stream.
6157 @item samples_per_frame
6158 Set the number of samples per output frame.
6160 The expression can contain the following constants:
6164 The (sequential) number of the output audio frame, starting from 0.
6167 The PTS (Presentation TimeStamp) of the output audio frame,
6168 expressed in @var{TB} units.
6171 The PTS of the output audio frame, expressed in seconds.
6174 The timebase of the output audio frames.
6177 Default is @code{1024}.
6180 @subsection Examples
6185 Generate a simple 440 Hz sine wave:
6191 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6195 sine=frequency=220:beep_factor=4:duration=5
6199 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6202 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6206 @c man end AUDIO SOURCES
6208 @chapter Audio Sinks
6209 @c man begin AUDIO SINKS
6211 Below is a description of the currently available audio sinks.
6213 @section abuffersink
6215 Buffer audio frames, and make them available to the end of filter chain.
6217 This sink is mainly intended for programmatic use, in particular
6218 through the interface defined in @file{libavfilter/buffersink.h}
6219 or the options system.
6221 It accepts a pointer to an AVABufferSinkContext structure, which
6222 defines the incoming buffers' formats, to be passed as the opaque
6223 parameter to @code{avfilter_init_filter} for initialization.
6226 Null audio sink; do absolutely nothing with the input audio. It is
6227 mainly useful as a template and for use in analysis / debugging
6230 @c man end AUDIO SINKS
6232 @chapter Video Filters
6233 @c man begin VIDEO FILTERS
6235 When you configure your FFmpeg build, you can disable any of the
6236 existing filters using @code{--disable-filters}.
6237 The configure output will show the video filters included in your
6240 Below is a description of the currently available video filters.
6244 Mark a region of interest in a video frame.
6246 The frame data is passed through unchanged, but metadata is attached
6247 to the frame indicating regions of interest which can affect the
6248 behaviour of later encoding. Multiple regions can be marked by
6249 applying the filter multiple times.
6253 Region distance in pixels from the left edge of the frame.
6255 Region distance in pixels from the top edge of the frame.
6257 Region width in pixels.
6259 Region height in pixels.
6261 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6262 and may contain the following variables:
6265 Width of the input frame.
6267 Height of the input frame.
6271 Quantisation offset to apply within the region.
6273 This must be a real value in the range -1 to +1. A value of zero
6274 indicates no quality change. A negative value asks for better quality
6275 (less quantisation), while a positive value asks for worse quality
6276 (greater quantisation).
6278 The range is calibrated so that the extreme values indicate the
6279 largest possible offset - if the rest of the frame is encoded with the
6280 worst possible quality, an offset of -1 indicates that this region
6281 should be encoded with the best possible quality anyway. Intermediate
6282 values are then interpolated in some codec-dependent way.
6284 For example, in 10-bit H.264 the quantisation parameter varies between
6285 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6286 this region should be encoded with a QP around one-tenth of the full
6287 range better than the rest of the frame. So, if most of the frame
6288 were to be encoded with a QP of around 30, this region would get a QP
6289 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6290 An extreme value of -1 would indicate that this region should be
6291 encoded with the best possible quality regardless of the treatment of
6292 the rest of the frame - that is, should be encoded at a QP of -12.
6294 If set to true, remove any existing regions of interest marked on the
6295 frame before adding the new one.
6298 @subsection Examples
6302 Mark the centre quarter of the frame as interesting.
6304 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6307 Mark the 100-pixel-wide region on the left edge of the frame as very
6308 uninteresting (to be encoded at much lower quality than the rest of
6311 addroi=0:0:100:ih:+1/5
6315 @section alphaextract
6317 Extract the alpha component from the input as a grayscale video. This
6318 is especially useful with the @var{alphamerge} filter.
6322 Add or replace the alpha component of the primary input with the
6323 grayscale value of a second input. This is intended for use with
6324 @var{alphaextract} to allow the transmission or storage of frame
6325 sequences that have alpha in a format that doesn't support an alpha
6328 For example, to reconstruct full frames from a normal YUV-encoded video
6329 and a separate video created with @var{alphaextract}, you might use:
6331 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6334 Since this filter is designed for reconstruction, it operates on frame
6335 sequences without considering timestamps, and terminates when either
6336 input reaches end of stream. This will cause problems if your encoding
6337 pipeline drops frames. If you're trying to apply an image as an
6338 overlay to a video stream, consider the @var{overlay} filter instead.
6342 Amplify differences between current pixel and pixels of adjacent frames in
6343 same pixel location.
6345 This filter accepts the following options:
6349 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6350 For example radius of 3 will instruct filter to calculate average of 7 frames.
6353 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6356 Set threshold for difference amplification. Any difference greater or equal to
6357 this value will not alter source pixel. Default is 10.
6358 Allowed range is from 0 to 65535.
6361 Set tolerance for difference amplification. Any difference lower to
6362 this value will not alter source pixel. Default is 0.
6363 Allowed range is from 0 to 65535.
6366 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6367 This option controls maximum possible value that will decrease source pixel value.
6370 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6371 This option controls maximum possible value that will increase source pixel value.
6374 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6377 @subsection Commands
6379 This filter supports the following @ref{commands} that corresponds to option of same name:
6391 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6392 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6393 Substation Alpha) subtitles files.
6395 This filter accepts the following option in addition to the common options from
6396 the @ref{subtitles} filter:
6400 Set the shaping engine
6402 Available values are:
6405 The default libass shaping engine, which is the best available.
6407 Fast, font-agnostic shaper that can do only substitutions
6409 Slower shaper using OpenType for substitutions and positioning
6412 The default is @code{auto}.
6416 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6418 The filter accepts the following options:
6422 Set threshold A for 1st plane. Default is 0.02.
6423 Valid range is 0 to 0.3.
6426 Set threshold B for 1st plane. Default is 0.04.
6427 Valid range is 0 to 5.
6430 Set threshold A for 2nd plane. Default is 0.02.
6431 Valid range is 0 to 0.3.
6434 Set threshold B for 2nd plane. Default is 0.04.
6435 Valid range is 0 to 5.
6438 Set threshold A for 3rd plane. Default is 0.02.
6439 Valid range is 0 to 0.3.
6442 Set threshold B for 3rd plane. Default is 0.04.
6443 Valid range is 0 to 5.
6445 Threshold A is designed to react on abrupt changes in the input signal and
6446 threshold B is designed to react on continuous changes in the input signal.
6449 Set number of frames filter will use for averaging. Default is 9. Must be odd
6450 number in range [5, 129].
6453 Set what planes of frame filter will use for averaging. Default is all.
6456 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
6457 Alternatively can be set to @code{s} serial.
6459 Parallel can be faster then serial, while other way around is never true.
6460 Parallel will abort early on first change being greater then thresholds, while serial
6461 will continue processing other side of frames if they are equal or bellow thresholds.
6464 @subsection Commands
6465 This filter supports same @ref{commands} as options except option @code{s}.
6466 The command accepts the same syntax of the corresponding option.
6470 Apply average blur filter.
6472 The filter accepts the following options:
6476 Set horizontal radius size.
6479 Set which planes to filter. By default all planes are filtered.
6482 Set vertical radius size, if zero it will be same as @code{sizeX}.
6483 Default is @code{0}.
6486 @subsection Commands
6487 This filter supports same commands as options.
6488 The command accepts the same syntax of the corresponding option.
6490 If the specified expression is not valid, it is kept at its current
6495 Compute the bounding box for the non-black pixels in the input frame
6498 This filter computes the bounding box containing all the pixels with a
6499 luminance value greater than the minimum allowed value.
6500 The parameters describing the bounding box are printed on the filter
6503 The filter accepts the following option:
6507 Set the minimal luminance value. Default is @code{16}.
6511 Apply bilateral filter, spatial smoothing while preserving edges.
6513 The filter accepts the following options:
6516 Set sigma of gaussian function to calculate spatial weight.
6517 Allowed range is 0 to 10. Default is 0.1.
6520 Set sigma of gaussian function to calculate range weight.
6521 Allowed range is 0 to 1. Default is 0.1.
6524 Set planes to filter. Default is first only.
6527 @section bitplanenoise
6529 Show and measure bit plane noise.
6531 The filter accepts the following options:
6535 Set which plane to analyze. Default is @code{1}.
6538 Filter out noisy pixels from @code{bitplane} set above.
6539 Default is disabled.
6542 @section blackdetect
6544 Detect video intervals that are (almost) completely black. Can be
6545 useful to detect chapter transitions, commercials, or invalid
6548 The filter outputs its detection analysis to both the log as well as
6549 frame metadata. If a black segment of at least the specified minimum
6550 duration is found, a line with the start and end timestamps as well
6551 as duration is printed to the log with level @code{info}. In addition,
6552 a log line with level @code{debug} is printed per frame showing the
6553 black amount detected for that frame.
6555 The filter also attaches metadata to the first frame of a black
6556 segment with key @code{lavfi.black_start} and to the first frame
6557 after the black segment ends with key @code{lavfi.black_end}. The
6558 value is the frame's timestamp. This metadata is added regardless
6559 of the minimum duration specified.
6561 The filter accepts the following options:
6564 @item black_min_duration, d
6565 Set the minimum detected black duration expressed in seconds. It must
6566 be a non-negative floating point number.
6568 Default value is 2.0.
6570 @item picture_black_ratio_th, pic_th
6571 Set the threshold for considering a picture "black".
6572 Express the minimum value for the ratio:
6574 @var{nb_black_pixels} / @var{nb_pixels}
6577 for which a picture is considered black.
6578 Default value is 0.98.
6580 @item pixel_black_th, pix_th
6581 Set the threshold for considering a pixel "black".
6583 The threshold expresses the maximum pixel luminance value for which a
6584 pixel is considered "black". The provided value is scaled according to
6585 the following equation:
6587 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6590 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6591 the input video format, the range is [0-255] for YUV full-range
6592 formats and [16-235] for YUV non full-range formats.
6594 Default value is 0.10.
6597 The following example sets the maximum pixel threshold to the minimum
6598 value, and detects only black intervals of 2 or more seconds:
6600 blackdetect=d=2:pix_th=0.00
6605 Detect frames that are (almost) completely black. Can be useful to
6606 detect chapter transitions or commercials. Output lines consist of
6607 the frame number of the detected frame, the percentage of blackness,
6608 the position in the file if known or -1 and the timestamp in seconds.
6610 In order to display the output lines, you need to set the loglevel at
6611 least to the AV_LOG_INFO value.
6613 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6614 The value represents the percentage of pixels in the picture that
6615 are below the threshold value.
6617 It accepts the following parameters:
6622 The percentage of the pixels that have to be below the threshold; it defaults to
6625 @item threshold, thresh
6626 The threshold below which a pixel value is considered black; it defaults to
6634 Blend two video frames into each other.
6636 The @code{blend} filter takes two input streams and outputs one
6637 stream, the first input is the "top" layer and second input is
6638 "bottom" layer. By default, the output terminates when the longest input terminates.
6640 The @code{tblend} (time blend) filter takes two consecutive frames
6641 from one single stream, and outputs the result obtained by blending
6642 the new frame on top of the old frame.
6644 A description of the accepted options follows.
6652 Set blend mode for specific pixel component or all pixel components in case
6653 of @var{all_mode}. Default value is @code{normal}.
6655 Available values for component modes are:
6697 Set blend opacity for specific pixel component or all pixel components in case
6698 of @var{all_opacity}. Only used in combination with pixel component blend modes.
6705 Set blend expression for specific pixel component or all pixel components in case
6706 of @var{all_expr}. Note that related mode options will be ignored if those are set.
6708 The expressions can use the following variables:
6712 The sequential number of the filtered frame, starting from @code{0}.
6716 the coordinates of the current sample
6720 the width and height of currently filtered plane
6724 Width and height scale for the plane being filtered. It is the
6725 ratio between the dimensions of the current plane to the luma plane,
6726 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
6727 the luma plane and @code{0.5,0.5} for the chroma planes.
6730 Time of the current frame, expressed in seconds.
6733 Value of pixel component at current location for first video frame (top layer).
6736 Value of pixel component at current location for second video frame (bottom layer).
6740 The @code{blend} filter also supports the @ref{framesync} options.
6742 @subsection Examples
6746 Apply transition from bottom layer to top layer in first 10 seconds:
6748 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
6752 Apply linear horizontal transition from top layer to bottom layer:
6754 blend=all_expr='A*(X/W)+B*(1-X/W)'
6758 Apply 1x1 checkerboard effect:
6760 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
6764 Apply uncover left effect:
6766 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6770 Apply uncover down effect:
6772 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6776 Apply uncover up-left effect:
6778 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6782 Split diagonally video and shows top and bottom layer on each side:
6784 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6788 Display differences between the current and the previous frame:
6790 tblend=all_mode=grainextract
6796 Denoise frames using Block-Matching 3D algorithm.
6798 The filter accepts the following options.
6802 Set denoising strength. Default value is 1.
6803 Allowed range is from 0 to 999.9.
6804 The denoising algorithm is very sensitive to sigma, so adjust it
6805 according to the source.
6808 Set local patch size. This sets dimensions in 2D.
6811 Set sliding step for processing blocks. Default value is 4.
6812 Allowed range is from 1 to 64.
6813 Smaller values allows processing more reference blocks and is slower.
6816 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6817 When set to 1, no block matching is done. Larger values allows more blocks
6819 Allowed range is from 1 to 256.
6822 Set radius for search block matching. Default is 9.
6823 Allowed range is from 1 to INT32_MAX.
6826 Set step between two search locations for block matching. Default is 1.
6827 Allowed range is from 1 to 64. Smaller is slower.
6830 Set threshold of mean square error for block matching. Valid range is 0 to
6834 Set thresholding parameter for hard thresholding in 3D transformed domain.
6835 Larger values results in stronger hard-thresholding filtering in frequency
6839 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6840 Default is @code{basic}.
6843 If enabled, filter will use 2nd stream for block matching.
6844 Default is disabled for @code{basic} value of @var{estim} option,
6845 and always enabled if value of @var{estim} is @code{final}.
6848 Set planes to filter. Default is all available except alpha.
6851 @subsection Examples
6855 Basic filtering with bm3d:
6857 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
6861 Same as above, but filtering only luma:
6863 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
6867 Same as above, but with both estimation modes:
6869 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
6873 Same as above, but prefilter with @ref{nlmeans} filter instead:
6875 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
6881 Apply a boxblur algorithm to the input video.
6883 It accepts the following parameters:
6887 @item luma_radius, lr
6888 @item luma_power, lp
6889 @item chroma_radius, cr
6890 @item chroma_power, cp
6891 @item alpha_radius, ar
6892 @item alpha_power, ap
6896 A description of the accepted options follows.
6899 @item luma_radius, lr
6900 @item chroma_radius, cr
6901 @item alpha_radius, ar
6902 Set an expression for the box radius in pixels used for blurring the
6903 corresponding input plane.
6905 The radius value must be a non-negative number, and must not be
6906 greater than the value of the expression @code{min(w,h)/2} for the
6907 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
6910 Default value for @option{luma_radius} is "2". If not specified,
6911 @option{chroma_radius} and @option{alpha_radius} default to the
6912 corresponding value set for @option{luma_radius}.
6914 The expressions can contain the following constants:
6918 The input width and height in pixels.
6922 The input chroma image width and height in pixels.
6926 The horizontal and vertical chroma subsample values. For example, for the
6927 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6930 @item luma_power, lp
6931 @item chroma_power, cp
6932 @item alpha_power, ap
6933 Specify how many times the boxblur filter is applied to the
6934 corresponding plane.
6936 Default value for @option{luma_power} is 2. If not specified,
6937 @option{chroma_power} and @option{alpha_power} default to the
6938 corresponding value set for @option{luma_power}.
6940 A value of 0 will disable the effect.
6943 @subsection Examples
6947 Apply a boxblur filter with the luma, chroma, and alpha radii
6950 boxblur=luma_radius=2:luma_power=1
6955 Set the luma radius to 2, and alpha and chroma radius to 0:
6957 boxblur=2:1:cr=0:ar=0
6961 Set the luma and chroma radii to a fraction of the video dimension:
6963 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6969 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6970 Deinterlacing Filter").
6972 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6973 interpolation algorithms.
6974 It accepts the following parameters:
6978 The interlacing mode to adopt. It accepts one of the following values:
6982 Output one frame for each frame.
6984 Output one frame for each field.
6987 The default value is @code{send_field}.
6990 The picture field parity assumed for the input interlaced video. It accepts one
6991 of the following values:
6995 Assume the top field is first.
6997 Assume the bottom field is first.
6999 Enable automatic detection of field parity.
7002 The default value is @code{auto}.
7003 If the interlacing is unknown or the decoder does not export this information,
7004 top field first will be assumed.
7007 Specify which frames to deinterlace. Accepts one of the following
7012 Deinterlace all frames.
7014 Only deinterlace frames marked as interlaced.
7017 The default value is @code{all}.
7022 Apply Contrast Adaptive Sharpen filter to video stream.
7024 The filter accepts the following options:
7028 Set the sharpening strength. Default value is 0.
7031 Set planes to filter. Default value is to filter all
7032 planes except alpha plane.
7036 Remove all color information for all colors except for certain one.
7038 The filter accepts the following options:
7042 The color which will not be replaced with neutral chroma.
7045 Similarity percentage with the above color.
7046 0.01 matches only the exact key color, while 1.0 matches everything.
7050 0.0 makes pixels either fully gray, or not gray at all.
7051 Higher values result in more preserved color.
7054 Signals that the color passed is already in YUV instead of RGB.
7056 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7057 This can be used to pass exact YUV values as hexadecimal numbers.
7060 @subsection Commands
7061 This filter supports same @ref{commands} as options.
7062 The command accepts the same syntax of the corresponding option.
7064 If the specified expression is not valid, it is kept at its current
7068 YUV colorspace color/chroma keying.
7070 The filter accepts the following options:
7074 The color which will be replaced with transparency.
7077 Similarity percentage with the key color.
7079 0.01 matches only the exact key color, while 1.0 matches everything.
7084 0.0 makes pixels either fully transparent, or not transparent at all.
7086 Higher values result in semi-transparent pixels, with a higher transparency
7087 the more similar the pixels color is to the key color.
7090 Signals that the color passed is already in YUV instead of RGB.
7092 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7093 This can be used to pass exact YUV values as hexadecimal numbers.
7096 @subsection Commands
7097 This filter supports same @ref{commands} as options.
7098 The command accepts the same syntax of the corresponding option.
7100 If the specified expression is not valid, it is kept at its current
7103 @subsection Examples
7107 Make every green pixel in the input image transparent:
7109 ffmpeg -i input.png -vf chromakey=green out.png
7113 Overlay a greenscreen-video on top of a static black background.
7115 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
7119 @section chromashift
7120 Shift chroma pixels horizontally and/or vertically.
7122 The filter accepts the following options:
7125 Set amount to shift chroma-blue horizontally.
7127 Set amount to shift chroma-blue vertically.
7129 Set amount to shift chroma-red horizontally.
7131 Set amount to shift chroma-red vertically.
7133 Set edge mode, can be @var{smear}, default, or @var{warp}.
7136 @subsection Commands
7138 This filter supports the all above options as @ref{commands}.
7142 Display CIE color diagram with pixels overlaid onto it.
7144 The filter accepts the following options:
7159 @item uhdtv, rec2020
7173 Set what gamuts to draw.
7175 See @code{system} option for available values.
7178 Set ciescope size, by default set to 512.
7181 Set intensity used to map input pixel values to CIE diagram.
7184 Set contrast used to draw tongue colors that are out of active color system gamut.
7187 Correct gamma displayed on scope, by default enabled.
7190 Show white point on CIE diagram, by default disabled.
7193 Set input gamma. Used only with XYZ input color space.
7198 Visualize information exported by some codecs.
7200 Some codecs can export information through frames using side-data or other
7201 means. For example, some MPEG based codecs export motion vectors through the
7202 @var{export_mvs} flag in the codec @option{flags2} option.
7204 The filter accepts the following option:
7208 Set motion vectors to visualize.
7210 Available flags for @var{mv} are:
7214 forward predicted MVs of P-frames
7216 forward predicted MVs of B-frames
7218 backward predicted MVs of B-frames
7222 Display quantization parameters using the chroma planes.
7225 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7227 Available flags for @var{mv_type} are:
7231 forward predicted MVs
7233 backward predicted MVs
7236 @item frame_type, ft
7237 Set frame type to visualize motion vectors of.
7239 Available flags for @var{frame_type} are:
7243 intra-coded frames (I-frames)
7245 predicted frames (P-frames)
7247 bi-directionally predicted frames (B-frames)
7251 @subsection Examples
7255 Visualize forward predicted MVs of all frames using @command{ffplay}:
7257 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7261 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7263 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7267 @section colorbalance
7268 Modify intensity of primary colors (red, green and blue) of input frames.
7270 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7271 regions for the red-cyan, green-magenta or blue-yellow balance.
7273 A positive adjustment value shifts the balance towards the primary color, a negative
7274 value towards the complementary color.
7276 The filter accepts the following options:
7282 Adjust red, green and blue shadows (darkest pixels).
7287 Adjust red, green and blue midtones (medium pixels).
7292 Adjust red, green and blue highlights (brightest pixels).
7294 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7297 Preserve lightness when changing color balance. Default is disabled.
7300 @subsection Examples
7304 Add red color cast to shadows:
7310 @subsection Commands
7312 This filter supports the all above options as @ref{commands}.
7314 @section colorchannelmixer
7316 Adjust video input frames by re-mixing color channels.
7318 This filter modifies a color channel by adding the values associated to
7319 the other channels of the same pixels. For example if the value to
7320 modify is red, the output value will be:
7322 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7325 The filter accepts the following options:
7332 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7333 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7339 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7340 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7346 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7347 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7353 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7354 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7356 Allowed ranges for options are @code{[-2.0, 2.0]}.
7359 @subsection Examples
7363 Convert source to grayscale:
7365 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7368 Simulate sepia tones:
7370 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7374 @subsection Commands
7376 This filter supports the all above options as @ref{commands}.
7379 RGB colorspace color keying.
7381 The filter accepts the following options:
7385 The color which will be replaced with transparency.
7388 Similarity percentage with the key color.
7390 0.01 matches only the exact key color, while 1.0 matches everything.
7395 0.0 makes pixels either fully transparent, or not transparent at all.
7397 Higher values result in semi-transparent pixels, with a higher transparency
7398 the more similar the pixels color is to the key color.
7401 @subsection Examples
7405 Make every green pixel in the input image transparent:
7407 ffmpeg -i input.png -vf colorkey=green out.png
7411 Overlay a greenscreen-video on top of a static background image.
7413 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
7417 @subsection Commands
7418 This filter supports same @ref{commands} as options.
7419 The command accepts the same syntax of the corresponding option.
7421 If the specified expression is not valid, it is kept at its current
7425 Remove all color information for all RGB colors except for certain one.
7427 The filter accepts the following options:
7431 The color which will not be replaced with neutral gray.
7434 Similarity percentage with the above color.
7435 0.01 matches only the exact key color, while 1.0 matches everything.
7438 Blend percentage. 0.0 makes pixels fully gray.
7439 Higher values result in more preserved color.
7442 @subsection Commands
7443 This filter supports same @ref{commands} as options.
7444 The command accepts the same syntax of the corresponding option.
7446 If the specified expression is not valid, it is kept at its current
7449 @section colorlevels
7451 Adjust video input frames using levels.
7453 The filter accepts the following options:
7460 Adjust red, green, blue and alpha input black point.
7461 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7467 Adjust red, green, blue and alpha input white point.
7468 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7470 Input levels are used to lighten highlights (bright tones), darken shadows
7471 (dark tones), change the balance of bright and dark tones.
7477 Adjust red, green, blue and alpha output black point.
7478 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7484 Adjust red, green, blue and alpha output white point.
7485 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7487 Output levels allows manual selection of a constrained output level range.
7490 @subsection Examples
7494 Make video output darker:
7496 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7502 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7506 Make video output lighter:
7508 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7512 Increase brightness:
7514 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7518 @subsection Commands
7520 This filter supports the all above options as @ref{commands}.
7522 @section colormatrix
7524 Convert color matrix.
7526 The filter accepts the following options:
7531 Specify the source and destination color matrix. Both values must be
7534 The accepted values are:
7562 For example to convert from BT.601 to SMPTE-240M, use the command:
7564 colormatrix=bt601:smpte240m
7569 Convert colorspace, transfer characteristics or color primaries.
7570 Input video needs to have an even size.
7572 The filter accepts the following options:
7577 Specify all color properties at once.
7579 The accepted values are:
7609 Specify output colorspace.
7611 The accepted values are:
7620 BT.470BG or BT.601-6 625
7623 SMPTE-170M or BT.601-6 525
7632 BT.2020 with non-constant luminance
7638 Specify output transfer characteristics.
7640 The accepted values are:
7652 Constant gamma of 2.2
7655 Constant gamma of 2.8
7658 SMPTE-170M, BT.601-6 625 or BT.601-6 525
7676 BT.2020 for 10-bits content
7679 BT.2020 for 12-bits content
7685 Specify output color primaries.
7687 The accepted values are:
7696 BT.470BG or BT.601-6 625
7699 SMPTE-170M or BT.601-6 525
7723 Specify output color range.
7725 The accepted values are:
7728 TV (restricted) range
7731 MPEG (restricted) range
7742 Specify output color format.
7744 The accepted values are:
7747 YUV 4:2:0 planar 8-bits
7750 YUV 4:2:0 planar 10-bits
7753 YUV 4:2:0 planar 12-bits
7756 YUV 4:2:2 planar 8-bits
7759 YUV 4:2:2 planar 10-bits
7762 YUV 4:2:2 planar 12-bits
7765 YUV 4:4:4 planar 8-bits
7768 YUV 4:4:4 planar 10-bits
7771 YUV 4:4:4 planar 12-bits
7776 Do a fast conversion, which skips gamma/primary correction. This will take
7777 significantly less CPU, but will be mathematically incorrect. To get output
7778 compatible with that produced by the colormatrix filter, use fast=1.
7781 Specify dithering mode.
7783 The accepted values are:
7789 Floyd-Steinberg dithering
7793 Whitepoint adaptation mode.
7795 The accepted values are:
7798 Bradford whitepoint adaptation
7801 von Kries whitepoint adaptation
7804 identity whitepoint adaptation (i.e. no whitepoint adaptation)
7808 Override all input properties at once. Same accepted values as @ref{all}.
7811 Override input colorspace. Same accepted values as @ref{space}.
7814 Override input color primaries. Same accepted values as @ref{primaries}.
7817 Override input transfer characteristics. Same accepted values as @ref{trc}.
7820 Override input color range. Same accepted values as @ref{range}.
7824 The filter converts the transfer characteristics, color space and color
7825 primaries to the specified user values. The output value, if not specified,
7826 is set to a default value based on the "all" property. If that property is
7827 also not specified, the filter will log an error. The output color range and
7828 format default to the same value as the input color range and format. The
7829 input transfer characteristics, color space, color primaries and color range
7830 should be set on the input data. If any of these are missing, the filter will
7831 log an error and no conversion will take place.
7833 For example to convert the input to SMPTE-240M, use the command:
7835 colorspace=smpte240m
7838 @section convolution
7840 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
7842 The filter accepts the following options:
7849 Set matrix for each plane.
7850 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
7851 and from 1 to 49 odd number of signed integers in @var{row} mode.
7857 Set multiplier for calculated value for each plane.
7858 If unset or 0, it will be sum of all matrix elements.
7864 Set bias for each plane. This value is added to the result of the multiplication.
7865 Useful for making the overall image brighter or darker. Default is 0.0.
7871 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
7872 Default is @var{square}.
7875 @subsection Examples
7881 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"
7887 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"
7893 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"
7899 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"
7903 Apply laplacian edge detector which includes diagonals:
7905 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"
7911 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"
7917 Apply 2D convolution of video stream in frequency domain using second stream
7920 The filter accepts the following options:
7924 Set which planes to process.
7927 Set which impulse video frames will be processed, can be @var{first}
7928 or @var{all}. Default is @var{all}.
7931 The @code{convolve} filter also supports the @ref{framesync} options.
7935 Copy the input video source unchanged to the output. This is mainly useful for
7940 Video filtering on GPU using Apple's CoreImage API on OSX.
7942 Hardware acceleration is based on an OpenGL context. Usually, this means it is
7943 processed by video hardware. However, software-based OpenGL implementations
7944 exist which means there is no guarantee for hardware processing. It depends on
7947 There are many filters and image generators provided by Apple that come with a
7948 large variety of options. The filter has to be referenced by its name along
7951 The coreimage filter accepts the following options:
7954 List all available filters and generators along with all their respective
7955 options as well as possible minimum and maximum values along with the default
7962 Specify all filters by their respective name and options.
7963 Use @var{list_filters} to determine all valid filter names and options.
7964 Numerical options are specified by a float value and are automatically clamped
7965 to their respective value range. Vector and color options have to be specified
7966 by a list of space separated float values. Character escaping has to be done.
7967 A special option name @code{default} is available to use default options for a
7970 It is required to specify either @code{default} or at least one of the filter options.
7971 All omitted options are used with their default values.
7972 The syntax of the filter string is as follows:
7974 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
7978 Specify a rectangle where the output of the filter chain is copied into the
7979 input image. It is given by a list of space separated float values:
7981 output_rect=x\ y\ width\ height
7983 If not given, the output rectangle equals the dimensions of the input image.
7984 The output rectangle is automatically cropped at the borders of the input
7985 image. Negative values are valid for each component.
7987 output_rect=25\ 25\ 100\ 100
7991 Several filters can be chained for successive processing without GPU-HOST
7992 transfers allowing for fast processing of complex filter chains.
7993 Currently, only filters with zero (generators) or exactly one (filters) input
7994 image and one output image are supported. Also, transition filters are not yet
7997 Some filters generate output images with additional padding depending on the
7998 respective filter kernel. The padding is automatically removed to ensure the
7999 filter output has the same size as the input image.
8001 For image generators, the size of the output image is determined by the
8002 previous output image of the filter chain or the input image of the whole
8003 filterchain, respectively. The generators do not use the pixel information of
8004 this image to generate their output. However, the generated output is
8005 blended onto this image, resulting in partial or complete coverage of the
8008 The @ref{coreimagesrc} video source can be used for generating input images
8009 which are directly fed into the filter chain. By using it, providing input
8010 images by another video source or an input video is not required.
8012 @subsection Examples
8017 List all filters available:
8019 coreimage=list_filters=true
8023 Use the CIBoxBlur filter with default options to blur an image:
8025 coreimage=filter=CIBoxBlur@@default
8029 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8030 its center at 100x100 and a radius of 50 pixels:
8032 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8036 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8037 given as complete and escaped command-line for Apple's standard bash shell:
8039 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8045 Cover a rectangular object
8047 It accepts the following options:
8051 Filepath of the optional cover image, needs to be in yuv420.
8056 It accepts the following values:
8059 cover it by the supplied image
8061 cover it by interpolating the surrounding pixels
8064 Default value is @var{blur}.
8067 @subsection Examples
8071 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8073 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8079 Crop the input video to given dimensions.
8081 It accepts the following parameters:
8085 The width of the output video. It defaults to @code{iw}.
8086 This expression is evaluated only once during the filter
8087 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8090 The height of the output video. It defaults to @code{ih}.
8091 This expression is evaluated only once during the filter
8092 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8095 The horizontal position, in the input video, of the left edge of the output
8096 video. It defaults to @code{(in_w-out_w)/2}.
8097 This expression is evaluated per-frame.
8100 The vertical position, in the input video, of the top edge of the output video.
8101 It defaults to @code{(in_h-out_h)/2}.
8102 This expression is evaluated per-frame.
8105 If set to 1 will force the output display aspect ratio
8106 to be the same of the input, by changing the output sample aspect
8107 ratio. It defaults to 0.
8110 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8111 width/height/x/y as specified and will not be rounded to nearest smaller value.
8115 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8116 expressions containing the following constants:
8121 The computed values for @var{x} and @var{y}. They are evaluated for
8126 The input width and height.
8130 These are the same as @var{in_w} and @var{in_h}.
8134 The output (cropped) width and height.
8138 These are the same as @var{out_w} and @var{out_h}.
8141 same as @var{iw} / @var{ih}
8144 input sample aspect ratio
8147 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8151 horizontal and vertical chroma subsample values. For example for the
8152 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8155 The number of the input frame, starting from 0.
8158 the position in the file of the input frame, NAN if unknown
8161 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8165 The expression for @var{out_w} may depend on the value of @var{out_h},
8166 and the expression for @var{out_h} may depend on @var{out_w}, but they
8167 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8168 evaluated after @var{out_w} and @var{out_h}.
8170 The @var{x} and @var{y} parameters specify the expressions for the
8171 position of the top-left corner of the output (non-cropped) area. They
8172 are evaluated for each frame. If the evaluated value is not valid, it
8173 is approximated to the nearest valid value.
8175 The expression for @var{x} may depend on @var{y}, and the expression
8176 for @var{y} may depend on @var{x}.
8178 @subsection Examples
8182 Crop area with size 100x100 at position (12,34).
8187 Using named options, the example above becomes:
8189 crop=w=100:h=100:x=12:y=34
8193 Crop the central input area with size 100x100:
8199 Crop the central input area with size 2/3 of the input video:
8201 crop=2/3*in_w:2/3*in_h
8205 Crop the input video central square:
8212 Delimit the rectangle with the top-left corner placed at position
8213 100:100 and the right-bottom corner corresponding to the right-bottom
8214 corner of the input image.
8216 crop=in_w-100:in_h-100:100:100
8220 Crop 10 pixels from the left and right borders, and 20 pixels from
8221 the top and bottom borders
8223 crop=in_w-2*10:in_h-2*20
8227 Keep only the bottom right quarter of the input image:
8229 crop=in_w/2:in_h/2:in_w/2:in_h/2
8233 Crop height for getting Greek harmony:
8235 crop=in_w:1/PHI*in_w
8239 Apply trembling effect:
8241 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)
8245 Apply erratic camera effect depending on timestamp:
8247 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)"
8251 Set x depending on the value of y:
8253 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8257 @subsection Commands
8259 This filter supports the following commands:
8265 Set width/height of the output video and the horizontal/vertical position
8267 The command accepts the same syntax of the corresponding option.
8269 If the specified expression is not valid, it is kept at its current
8275 Auto-detect the crop size.
8277 It calculates the necessary cropping parameters and prints the
8278 recommended parameters via the logging system. The detected dimensions
8279 correspond to the non-black area of the input video.
8281 It accepts the following parameters:
8286 Set higher black value threshold, which can be optionally specified
8287 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8288 value greater to the set value is considered non-black. It defaults to 24.
8289 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8290 on the bitdepth of the pixel format.
8293 The value which the width/height should be divisible by. It defaults to
8294 16. The offset is automatically adjusted to center the video. Use 2 to
8295 get only even dimensions (needed for 4:2:2 video). 16 is best when
8296 encoding to most video codecs.
8298 @item reset_count, reset
8299 Set the counter that determines after how many frames cropdetect will
8300 reset the previously detected largest video area and start over to
8301 detect the current optimal crop area. Default value is 0.
8303 This can be useful when channel logos distort the video area. 0
8304 indicates 'never reset', and returns the largest area encountered during
8311 Delay video filtering until a given wallclock timestamp. The filter first
8312 passes on @option{preroll} amount of frames, then it buffers at most
8313 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8314 it forwards the buffered frames and also any subsequent frames coming in its
8317 The filter can be used synchronize the output of multiple ffmpeg processes for
8318 realtime output devices like decklink. By putting the delay in the filtering
8319 chain and pre-buffering frames the process can pass on data to output almost
8320 immediately after the target wallclock timestamp is reached.
8322 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8328 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8331 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8334 The maximum duration of content to buffer before waiting for the cue expressed
8335 in seconds. Default is 0.
8342 Apply color adjustments using curves.
8344 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8345 component (red, green and blue) has its values defined by @var{N} key points
8346 tied from each other using a smooth curve. The x-axis represents the pixel
8347 values from the input frame, and the y-axis the new pixel values to be set for
8350 By default, a component curve is defined by the two points @var{(0;0)} and
8351 @var{(1;1)}. This creates a straight line where each original pixel value is
8352 "adjusted" to its own value, which means no change to the image.
8354 The filter allows you to redefine these two points and add some more. A new
8355 curve (using a natural cubic spline interpolation) will be define to pass
8356 smoothly through all these new coordinates. The new defined points needs to be
8357 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8358 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8359 the vector spaces, the values will be clipped accordingly.
8361 The filter accepts the following options:
8365 Select one of the available color presets. This option can be used in addition
8366 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8367 options takes priority on the preset values.
8368 Available presets are:
8371 @item color_negative
8374 @item increase_contrast
8376 @item linear_contrast
8377 @item medium_contrast
8379 @item strong_contrast
8382 Default is @code{none}.
8384 Set the master key points. These points will define a second pass mapping. It
8385 is sometimes called a "luminance" or "value" mapping. It can be used with
8386 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8387 post-processing LUT.
8389 Set the key points for the red component.
8391 Set the key points for the green component.
8393 Set the key points for the blue component.
8395 Set the key points for all components (not including master).
8396 Can be used in addition to the other key points component
8397 options. In this case, the unset component(s) will fallback on this
8398 @option{all} setting.
8400 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
8402 Save Gnuplot script of the curves in specified file.
8405 To avoid some filtergraph syntax conflicts, each key points list need to be
8406 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
8408 @subsection Examples
8412 Increase slightly the middle level of blue:
8414 curves=blue='0/0 0.5/0.58 1/1'
8420 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'
8422 Here we obtain the following coordinates for each components:
8425 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8427 @code{(0;0) (0.50;0.48) (1;1)}
8429 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8433 The previous example can also be achieved with the associated built-in preset:
8435 curves=preset=vintage
8445 Use a Photoshop preset and redefine the points of the green component:
8447 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8451 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8452 and @command{gnuplot}:
8454 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8455 gnuplot -p /tmp/curves.plt
8461 Video data analysis filter.
8463 This filter shows hexadecimal pixel values of part of video.
8465 The filter accepts the following options:
8469 Set output video size.
8472 Set x offset from where to pick pixels.
8475 Set y offset from where to pick pixels.
8478 Set scope mode, can be one of the following:
8481 Draw hexadecimal pixel values with white color on black background.
8484 Draw hexadecimal pixel values with input video pixel color on black
8488 Draw hexadecimal pixel values on color background picked from input video,
8489 the text color is picked in such way so its always visible.
8493 Draw rows and columns numbers on left and top of video.
8496 Set background opacity.
8499 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
8504 Denoise frames using 2D DCT (frequency domain filtering).
8506 This filter is not designed for real time.
8508 The filter accepts the following options:
8512 Set the noise sigma constant.
8514 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8515 coefficient (absolute value) below this threshold with be dropped.
8517 If you need a more advanced filtering, see @option{expr}.
8519 Default is @code{0}.
8522 Set number overlapping pixels for each block. Since the filter can be slow, you
8523 may want to reduce this value, at the cost of a less effective filter and the
8524 risk of various artefacts.
8526 If the overlapping value doesn't permit processing the whole input width or
8527 height, a warning will be displayed and according borders won't be denoised.
8529 Default value is @var{blocksize}-1, which is the best possible setting.
8532 Set the coefficient factor expression.
8534 For each coefficient of a DCT block, this expression will be evaluated as a
8535 multiplier value for the coefficient.
8537 If this is option is set, the @option{sigma} option will be ignored.
8539 The absolute value of the coefficient can be accessed through the @var{c}
8543 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8544 @var{blocksize}, which is the width and height of the processed blocks.
8546 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8547 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8548 on the speed processing. Also, a larger block size does not necessarily means a
8552 @subsection Examples
8554 Apply a denoise with a @option{sigma} of @code{4.5}:
8559 The same operation can be achieved using the expression system:
8561 dctdnoiz=e='gte(c, 4.5*3)'
8564 Violent denoise using a block size of @code{16x16}:
8571 Remove banding artifacts from input video.
8572 It works by replacing banded pixels with average value of referenced pixels.
8574 The filter accepts the following options:
8581 Set banding detection threshold for each plane. Default is 0.02.
8582 Valid range is 0.00003 to 0.5.
8583 If difference between current pixel and reference pixel is less than threshold,
8584 it will be considered as banded.
8587 Banding detection range in pixels. Default is 16. If positive, random number
8588 in range 0 to set value will be used. If negative, exact absolute value
8590 The range defines square of four pixels around current pixel.
8593 Set direction in radians from which four pixel will be compared. If positive,
8594 random direction from 0 to set direction will be picked. If negative, exact of
8595 absolute value will be picked. For example direction 0, -PI or -2*PI radians
8596 will pick only pixels on same row and -PI/2 will pick only pixels on same
8600 If enabled, current pixel is compared with average value of all four
8601 surrounding pixels. The default is enabled. If disabled current pixel is
8602 compared with all four surrounding pixels. The pixel is considered banded
8603 if only all four differences with surrounding pixels are less than threshold.
8606 If enabled, current pixel is changed if and only if all pixel components are banded,
8607 e.g. banding detection threshold is triggered for all color components.
8608 The default is disabled.
8613 Remove blocking artifacts from input video.
8615 The filter accepts the following options:
8619 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
8620 This controls what kind of deblocking is applied.
8623 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
8629 Set blocking detection thresholds. Allowed range is 0 to 1.
8630 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
8631 Using higher threshold gives more deblocking strength.
8632 Setting @var{alpha} controls threshold detection at exact edge of block.
8633 Remaining options controls threshold detection near the edge. Each one for
8634 below/above or left/right. Setting any of those to @var{0} disables
8638 Set planes to filter. Default is to filter all available planes.
8641 @subsection Examples
8645 Deblock using weak filter and block size of 4 pixels.
8647 deblock=filter=weak:block=4
8651 Deblock using strong filter, block size of 4 pixels and custom thresholds for
8652 deblocking more edges.
8654 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
8658 Similar as above, but filter only first plane.
8660 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
8664 Similar as above, but filter only second and third plane.
8666 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
8673 Drop duplicated frames at regular intervals.
8675 The filter accepts the following options:
8679 Set the number of frames from which one will be dropped. Setting this to
8680 @var{N} means one frame in every batch of @var{N} frames will be dropped.
8681 Default is @code{5}.
8684 Set the threshold for duplicate detection. If the difference metric for a frame
8685 is less than or equal to this value, then it is declared as duplicate. Default
8689 Set scene change threshold. Default is @code{15}.
8693 Set the size of the x and y-axis blocks used during metric calculations.
8694 Larger blocks give better noise suppression, but also give worse detection of
8695 small movements. Must be a power of two. Default is @code{32}.
8698 Mark main input as a pre-processed input and activate clean source input
8699 stream. This allows the input to be pre-processed with various filters to help
8700 the metrics calculation while keeping the frame selection lossless. When set to
8701 @code{1}, the first stream is for the pre-processed input, and the second
8702 stream is the clean source from where the kept frames are chosen. Default is
8706 Set whether or not chroma is considered in the metric calculations. Default is
8712 Apply 2D deconvolution of video stream in frequency domain using second stream
8715 The filter accepts the following options:
8719 Set which planes to process.
8722 Set which impulse video frames will be processed, can be @var{first}
8723 or @var{all}. Default is @var{all}.
8726 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
8727 and height are not same and not power of 2 or if stream prior to convolving
8731 The @code{deconvolve} filter also supports the @ref{framesync} options.
8735 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
8737 It accepts the following options:
8741 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
8742 @var{rainbows} for cross-color reduction.
8745 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
8748 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
8751 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
8754 Set temporal chroma threshold. Lower values increases reduction of cross-color.
8759 Apply deflate effect to the video.
8761 This filter replaces the pixel by the local(3x3) average by taking into account
8762 only values lower than the pixel.
8764 It accepts the following options:
8771 Limit the maximum change for each plane, default is 65535.
8772 If 0, plane will remain unchanged.
8775 @subsection Commands
8777 This filter supports the all above options as @ref{commands}.
8781 Remove temporal frame luminance variations.
8783 It accepts the following options:
8787 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
8790 Set averaging mode to smooth temporal luminance variations.
8792 Available values are:
8817 Do not actually modify frame. Useful when one only wants metadata.
8822 Remove judder produced by partially interlaced telecined content.
8824 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
8825 source was partially telecined content then the output of @code{pullup,dejudder}
8826 will have a variable frame rate. May change the recorded frame rate of the
8827 container. Aside from that change, this filter will not affect constant frame
8830 The option available in this filter is:
8834 Specify the length of the window over which the judder repeats.
8836 Accepts any integer greater than 1. Useful values are:
8840 If the original was telecined from 24 to 30 fps (Film to NTSC).
8843 If the original was telecined from 25 to 30 fps (PAL to NTSC).
8846 If a mixture of the two.
8849 The default is @samp{4}.
8854 Suppress a TV station logo by a simple interpolation of the surrounding
8855 pixels. Just set a rectangle covering the logo and watch it disappear
8856 (and sometimes something even uglier appear - your mileage may vary).
8858 It accepts the following parameters:
8863 Specify the top left corner coordinates of the logo. They must be
8868 Specify the width and height of the logo to clear. They must be
8872 Specify the thickness of the fuzzy edge of the rectangle (added to
8873 @var{w} and @var{h}). The default value is 1. This option is
8874 deprecated, setting higher values should no longer be necessary and
8878 When set to 1, a green rectangle is drawn on the screen to simplify
8879 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
8880 The default value is 0.
8882 The rectangle is drawn on the outermost pixels which will be (partly)
8883 replaced with interpolated values. The values of the next pixels
8884 immediately outside this rectangle in each direction will be used to
8885 compute the interpolated pixel values inside the rectangle.
8889 @subsection Examples
8893 Set a rectangle covering the area with top left corner coordinates 0,0
8894 and size 100x77, and a band of size 10:
8896 delogo=x=0:y=0:w=100:h=77:band=10
8904 Remove the rain in the input image/video by applying the derain methods based on
8905 convolutional neural networks. Supported models:
8909 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
8910 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
8913 Training as well as model generation scripts are provided in
8914 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
8916 Native model files (.model) can be generated from TensorFlow model
8917 files (.pb) by using tools/python/convert.py
8919 The filter accepts the following options:
8923 Specify which filter to use. This option accepts the following values:
8927 Derain filter. To conduct derain filter, you need to use a derain model.
8930 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
8932 Default value is @samp{derain}.
8935 Specify which DNN backend to use for model loading and execution. This option accepts
8936 the following values:
8940 Native implementation of DNN loading and execution.
8943 TensorFlow backend. To enable this backend you
8944 need to install the TensorFlow for C library (see
8945 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
8946 @code{--enable-libtensorflow}
8948 Default value is @samp{native}.
8951 Set path to model file specifying network architecture and its parameters.
8952 Note that different backends use different file formats. TensorFlow and native
8953 backend can load files for only its format.
8956 It can also be finished with @ref{dnn_processing} filter.
8960 Attempt to fix small changes in horizontal and/or vertical shift. This
8961 filter helps remove camera shake from hand-holding a camera, bumping a
8962 tripod, moving on a vehicle, etc.
8964 The filter accepts the following options:
8972 Specify a rectangular area where to limit the search for motion
8974 If desired the search for motion vectors can be limited to a
8975 rectangular area of the frame defined by its top left corner, width
8976 and height. These parameters have the same meaning as the drawbox
8977 filter which can be used to visualise the position of the bounding
8980 This is useful when simultaneous movement of subjects within the frame
8981 might be confused for camera motion by the motion vector search.
8983 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
8984 then the full frame is used. This allows later options to be set
8985 without specifying the bounding box for the motion vector search.
8987 Default - search the whole frame.
8991 Specify the maximum extent of movement in x and y directions in the
8992 range 0-64 pixels. Default 16.
8995 Specify how to generate pixels to fill blanks at the edge of the
8996 frame. Available values are:
8999 Fill zeroes at blank locations
9001 Original image at blank locations
9003 Extruded edge value at blank locations
9005 Mirrored edge at blank locations
9007 Default value is @samp{mirror}.
9010 Specify the blocksize to use for motion search. Range 4-128 pixels,
9014 Specify the contrast threshold for blocks. Only blocks with more than
9015 the specified contrast (difference between darkest and lightest
9016 pixels) will be considered. Range 1-255, default 125.
9019 Specify the search strategy. Available values are:
9022 Set exhaustive search
9024 Set less exhaustive search.
9026 Default value is @samp{exhaustive}.
9029 If set then a detailed log of the motion search is written to the
9036 Remove unwanted contamination of foreground colors, caused by reflected color of
9037 greenscreen or bluescreen.
9039 This filter accepts the following options:
9043 Set what type of despill to use.
9046 Set how spillmap will be generated.
9049 Set how much to get rid of still remaining spill.
9052 Controls amount of red in spill area.
9055 Controls amount of green in spill area.
9056 Should be -1 for greenscreen.
9059 Controls amount of blue in spill area.
9060 Should be -1 for bluescreen.
9063 Controls brightness of spill area, preserving colors.
9066 Modify alpha from generated spillmap.
9071 Apply an exact inverse of the telecine operation. It requires a predefined
9072 pattern specified using the pattern option which must be the same as that passed
9073 to the telecine filter.
9075 This filter accepts the following options:
9084 The default value is @code{top}.
9088 A string of numbers representing the pulldown pattern you wish to apply.
9089 The default value is @code{23}.
9092 A number representing position of the first frame with respect to the telecine
9093 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9098 Apply dilation effect to the video.
9100 This filter replaces the pixel by the local(3x3) maximum.
9102 It accepts the following options:
9109 Limit the maximum change for each plane, default is 65535.
9110 If 0, plane will remain unchanged.
9113 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9116 Flags to local 3x3 coordinates maps like this:
9123 @subsection Commands
9125 This filter supports the all above options as @ref{commands}.
9129 Displace pixels as indicated by second and third input stream.
9131 It takes three input streams and outputs one stream, the first input is the
9132 source, and second and third input are displacement maps.
9134 The second input specifies how much to displace pixels along the
9135 x-axis, while the third input specifies how much to displace pixels
9137 If one of displacement map streams terminates, last frame from that
9138 displacement map will be used.
9140 Note that once generated, displacements maps can be reused over and over again.
9142 A description of the accepted options follows.
9146 Set displace behavior for pixels that are out of range.
9148 Available values are:
9151 Missing pixels are replaced by black pixels.
9154 Adjacent pixels will spread out to replace missing pixels.
9157 Out of range pixels are wrapped so they point to pixels of other side.
9160 Out of range pixels will be replaced with mirrored pixels.
9162 Default is @samp{smear}.
9166 @subsection Examples
9170 Add ripple effect to rgb input of video size hd720:
9172 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
9176 Add wave effect to rgb input of video size hd720:
9178 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
9182 @anchor{dnn_processing}
9183 @section dnn_processing
9185 Do image processing with deep neural networks. It works together with another filter
9186 which converts the pixel format of the Frame to what the dnn network requires.
9188 The filter accepts the following options:
9192 Specify which DNN backend to use for model loading and execution. This option accepts
9193 the following values:
9197 Native implementation of DNN loading and execution.
9200 TensorFlow backend. To enable this backend you
9201 need to install the TensorFlow for C library (see
9202 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9203 @code{--enable-libtensorflow}
9206 Default value is @samp{native}.
9209 Set path to model file specifying network architecture and its parameters.
9210 Note that different backends use different file formats. TensorFlow and native
9211 backend can load files for only its format.
9213 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9216 Set the input name of the dnn network.
9219 Set the output name of the dnn network.
9223 @subsection Examples
9227 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9229 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9233 Halve the pixel value of the frame with format gray32f:
9235 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
9239 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9241 ./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
9245 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9247 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9254 Draw a colored box on the input image.
9256 It accepts the following parameters:
9261 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9265 The expressions which specify the width and height of the box; if 0 they are interpreted as
9266 the input width and height. It defaults to 0.
9269 Specify the color of the box to write. For the general syntax of this option,
9270 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9271 value @code{invert} is used, the box edge color is the same as the
9272 video with inverted luma.
9275 The expression which sets the thickness of the box edge.
9276 A value of @code{fill} will create a filled box. Default value is @code{3}.
9278 See below for the list of accepted constants.
9281 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
9282 will overwrite the video's color and alpha pixels.
9283 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
9286 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9287 following constants:
9291 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9295 horizontal and vertical chroma subsample values. For example for the
9296 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9300 The input width and height.
9303 The input sample aspect ratio.
9307 The x and y offset coordinates where the box is drawn.
9311 The width and height of the drawn box.
9314 The thickness of the drawn box.
9316 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9317 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9321 @subsection Examples
9325 Draw a black box around the edge of the input image:
9331 Draw a box with color red and an opacity of 50%:
9333 drawbox=10:20:200:60:red@@0.5
9336 The previous example can be specified as:
9338 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
9342 Fill the box with pink color:
9344 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
9348 Draw a 2-pixel red 2.40:1 mask:
9350 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
9354 @subsection Commands
9355 This filter supports same commands as options.
9356 The command accepts the same syntax of the corresponding option.
9358 If the specified expression is not valid, it is kept at its current
9363 Draw a graph using input video metadata.
9365 It accepts the following parameters:
9369 Set 1st frame metadata key from which metadata values will be used to draw a graph.
9372 Set 1st foreground color expression.
9375 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
9378 Set 2nd foreground color expression.
9381 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
9384 Set 3rd foreground color expression.
9387 Set 4th frame metadata key from which metadata values will be used to draw a graph.
9390 Set 4th foreground color expression.
9393 Set minimal value of metadata value.
9396 Set maximal value of metadata value.
9399 Set graph background color. Default is white.
9404 Available values for mode is:
9411 Default is @code{line}.
9416 Available values for slide is:
9419 Draw new frame when right border is reached.
9422 Replace old columns with new ones.
9425 Scroll from right to left.
9428 Scroll from left to right.
9431 Draw single picture.
9434 Default is @code{frame}.
9437 Set size of graph video. For the syntax of this option, check the
9438 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
9439 The default value is @code{900x256}.
9442 Set the output frame rate. Default value is @code{25}.
9444 The foreground color expressions can use the following variables:
9447 Minimal value of metadata value.
9450 Maximal value of metadata value.
9453 Current metadata key value.
9456 The color is defined as 0xAABBGGRR.
9459 Example using metadata from @ref{signalstats} filter:
9461 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
9464 Example using metadata from @ref{ebur128} filter:
9466 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
9471 Draw a grid on the input image.
9473 It accepts the following parameters:
9478 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
9482 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
9483 input width and height, respectively, minus @code{thickness}, so image gets
9484 framed. Default to 0.
9487 Specify the color of the grid. For the general syntax of this option,
9488 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9489 value @code{invert} is used, the grid color is the same as the
9490 video with inverted luma.
9493 The expression which sets the thickness of the grid line. Default value is @code{1}.
9495 See below for the list of accepted constants.
9498 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
9499 will overwrite the video's color and alpha pixels.
9500 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
9503 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9504 following constants:
9508 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9512 horizontal and vertical chroma subsample values. For example for the
9513 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9517 The input grid cell width and height.
9520 The input sample aspect ratio.
9524 The x and y coordinates of some point of grid intersection (meant to configure offset).
9528 The width and height of the drawn cell.
9531 The thickness of the drawn cell.
9533 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9534 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9538 @subsection Examples
9542 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
9544 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
9548 Draw a white 3x3 grid with an opacity of 50%:
9550 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
9554 @subsection Commands
9555 This filter supports same commands as options.
9556 The command accepts the same syntax of the corresponding option.
9558 If the specified expression is not valid, it is kept at its current
9564 Draw a text string or text from a specified file on top of a video, using the
9565 libfreetype library.
9567 To enable compilation of this filter, you need to configure FFmpeg with
9568 @code{--enable-libfreetype}.
9569 To enable default font fallback and the @var{font} option you need to
9570 configure FFmpeg with @code{--enable-libfontconfig}.
9571 To enable the @var{text_shaping} option, you need to configure FFmpeg with
9572 @code{--enable-libfribidi}.
9576 It accepts the following parameters:
9581 Used to draw a box around text using the background color.
9582 The value must be either 1 (enable) or 0 (disable).
9583 The default value of @var{box} is 0.
9586 Set the width of the border to be drawn around the box using @var{boxcolor}.
9587 The default value of @var{boxborderw} is 0.
9590 The color to be used for drawing box around text. For the syntax of this
9591 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9593 The default value of @var{boxcolor} is "white".
9596 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
9597 The default value of @var{line_spacing} is 0.
9600 Set the width of the border to be drawn around the text using @var{bordercolor}.
9601 The default value of @var{borderw} is 0.
9604 Set the color to be used for drawing border around text. For the syntax of this
9605 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9607 The default value of @var{bordercolor} is "black".
9610 Select how the @var{text} is expanded. Can be either @code{none},
9611 @code{strftime} (deprecated) or
9612 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
9616 Set a start time for the count. Value is in microseconds. Only applied
9617 in the deprecated strftime expansion mode. To emulate in normal expansion
9618 mode use the @code{pts} function, supplying the start time (in seconds)
9619 as the second argument.
9622 If true, check and fix text coords to avoid clipping.
9625 The color to be used for drawing fonts. For the syntax of this option, check
9626 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9628 The default value of @var{fontcolor} is "black".
9630 @item fontcolor_expr
9631 String which is expanded the same way as @var{text} to obtain dynamic
9632 @var{fontcolor} value. By default this option has empty value and is not
9633 processed. When this option is set, it overrides @var{fontcolor} option.
9636 The font family to be used for drawing text. By default Sans.
9639 The font file to be used for drawing text. The path must be included.
9640 This parameter is mandatory if the fontconfig support is disabled.
9643 Draw the text applying alpha blending. The value can
9644 be a number between 0.0 and 1.0.
9645 The expression accepts the same variables @var{x, y} as well.
9646 The default value is 1.
9647 Please see @var{fontcolor_expr}.
9650 The font size to be used for drawing text.
9651 The default value of @var{fontsize} is 16.
9654 If set to 1, attempt to shape the text (for example, reverse the order of
9655 right-to-left text and join Arabic characters) before drawing it.
9656 Otherwise, just draw the text exactly as given.
9657 By default 1 (if supported).
9660 The flags to be used for loading the fonts.
9662 The flags map the corresponding flags supported by libfreetype, and are
9663 a combination of the following values:
9670 @item vertical_layout
9671 @item force_autohint
9674 @item ignore_global_advance_width
9676 @item ignore_transform
9682 Default value is "default".
9684 For more information consult the documentation for the FT_LOAD_*
9688 The color to be used for drawing a shadow behind the drawn text. For the
9689 syntax of this option, check the @ref{color syntax,,"Color" section in the
9690 ffmpeg-utils manual,ffmpeg-utils}.
9692 The default value of @var{shadowcolor} is "black".
9696 The x and y offsets for the text shadow position with respect to the
9697 position of the text. They can be either positive or negative
9698 values. The default value for both is "0".
9701 The starting frame number for the n/frame_num variable. The default value
9705 The size in number of spaces to use for rendering the tab.
9709 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
9710 format. It can be used with or without text parameter. @var{timecode_rate}
9711 option must be specified.
9713 @item timecode_rate, rate, r
9714 Set the timecode frame rate (timecode only). Value will be rounded to nearest
9715 integer. Minimum value is "1".
9716 Drop-frame timecode is supported for frame rates 30 & 60.
9719 If set to 1, the output of the timecode option will wrap around at 24 hours.
9720 Default is 0 (disabled).
9723 The text string to be drawn. The text must be a sequence of UTF-8
9725 This parameter is mandatory if no file is specified with the parameter
9729 A text file containing text to be drawn. The text must be a sequence
9730 of UTF-8 encoded characters.
9732 This parameter is mandatory if no text string is specified with the
9733 parameter @var{text}.
9735 If both @var{text} and @var{textfile} are specified, an error is thrown.
9738 If set to 1, the @var{textfile} will be reloaded before each frame.
9739 Be sure to update it atomically, or it may be read partially, or even fail.
9743 The expressions which specify the offsets where text will be drawn
9744 within the video frame. They are relative to the top/left border of the
9747 The default value of @var{x} and @var{y} is "0".
9749 See below for the list of accepted constants and functions.
9752 The parameters for @var{x} and @var{y} are expressions containing the
9753 following constants and functions:
9757 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
9761 horizontal and vertical chroma subsample values. For example for the
9762 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9765 the height of each text line
9773 @item max_glyph_a, ascent
9774 the maximum distance from the baseline to the highest/upper grid
9775 coordinate used to place a glyph outline point, for all the rendered
9777 It is a positive value, due to the grid's orientation with the Y axis
9780 @item max_glyph_d, descent
9781 the maximum distance from the baseline to the lowest grid coordinate
9782 used to place a glyph outline point, for all the rendered glyphs.
9783 This is a negative value, due to the grid's orientation, with the Y axis
9787 maximum glyph height, that is the maximum height for all the glyphs
9788 contained in the rendered text, it is equivalent to @var{ascent} -
9792 maximum glyph width, that is the maximum width for all the glyphs
9793 contained in the rendered text
9796 the number of input frame, starting from 0
9798 @item rand(min, max)
9799 return a random number included between @var{min} and @var{max}
9802 The input sample aspect ratio.
9805 timestamp expressed in seconds, NAN if the input timestamp is unknown
9808 the height of the rendered text
9811 the width of the rendered text
9815 the x and y offset coordinates where the text is drawn.
9817 These parameters allow the @var{x} and @var{y} expressions to refer
9818 to each other, so you can for example specify @code{y=x/dar}.
9821 A one character description of the current frame's picture type.
9824 The current packet's position in the input file or stream
9825 (in bytes, from the start of the input). A value of -1 indicates
9826 this info is not available.
9829 The current packet's duration, in seconds.
9832 The current packet's size (in bytes).
9835 @anchor{drawtext_expansion}
9836 @subsection Text expansion
9838 If @option{expansion} is set to @code{strftime},
9839 the filter recognizes strftime() sequences in the provided text and
9840 expands them accordingly. Check the documentation of strftime(). This
9841 feature is deprecated.
9843 If @option{expansion} is set to @code{none}, the text is printed verbatim.
9845 If @option{expansion} is set to @code{normal} (which is the default),
9846 the following expansion mechanism is used.
9848 The backslash character @samp{\}, followed by any character, always expands to
9849 the second character.
9851 Sequences of the form @code{%@{...@}} are expanded. The text between the
9852 braces is a function name, possibly followed by arguments separated by ':'.
9853 If the arguments contain special characters or delimiters (':' or '@}'),
9854 they should be escaped.
9856 Note that they probably must also be escaped as the value for the
9857 @option{text} option in the filter argument string and as the filter
9858 argument in the filtergraph description, and possibly also for the shell,
9859 that makes up to four levels of escaping; using a text file avoids these
9862 The following functions are available:
9867 The expression evaluation result.
9869 It must take one argument specifying the expression to be evaluated,
9870 which accepts the same constants and functions as the @var{x} and
9871 @var{y} values. Note that not all constants should be used, for
9872 example the text size is not known when evaluating the expression, so
9873 the constants @var{text_w} and @var{text_h} will have an undefined
9876 @item expr_int_format, eif
9877 Evaluate the expression's value and output as formatted integer.
9879 The first argument is the expression to be evaluated, just as for the @var{expr} function.
9880 The second argument specifies the output format. Allowed values are @samp{x},
9881 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
9882 @code{printf} function.
9883 The third parameter is optional and sets the number of positions taken by the output.
9884 It can be used to add padding with zeros from the left.
9887 The time at which the filter is running, expressed in UTC.
9888 It can accept an argument: a strftime() format string.
9891 The time at which the filter is running, expressed in the local time zone.
9892 It can accept an argument: a strftime() format string.
9895 Frame metadata. Takes one or two arguments.
9897 The first argument is mandatory and specifies the metadata key.
9899 The second argument is optional and specifies a default value, used when the
9900 metadata key is not found or empty.
9902 Available metadata can be identified by inspecting entries
9903 starting with TAG included within each frame section
9904 printed by running @code{ffprobe -show_frames}.
9906 String metadata generated in filters leading to
9907 the drawtext filter are also available.
9910 The frame number, starting from 0.
9913 A one character description of the current picture type.
9916 The timestamp of the current frame.
9917 It can take up to three arguments.
9919 The first argument is the format of the timestamp; it defaults to @code{flt}
9920 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
9921 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
9922 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
9923 @code{localtime} stands for the timestamp of the frame formatted as
9924 local time zone time.
9926 The second argument is an offset added to the timestamp.
9928 If the format is set to @code{hms}, a third argument @code{24HH} may be
9929 supplied to present the hour part of the formatted timestamp in 24h format
9932 If the format is set to @code{localtime} or @code{gmtime},
9933 a third argument may be supplied: a strftime() format string.
9934 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
9937 @subsection Commands
9939 This filter supports altering parameters via commands:
9942 Alter existing filter parameters.
9944 Syntax for the argument is the same as for filter invocation, e.g.
9947 fontsize=56:fontcolor=green:text='Hello World'
9950 Full filter invocation with sendcmd would look like this:
9953 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
9957 If the entire argument can't be parsed or applied as valid values then the filter will
9958 continue with its existing parameters.
9960 @subsection Examples
9964 Draw "Test Text" with font FreeSerif, using the default values for the
9965 optional parameters.
9968 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
9972 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
9973 and y=50 (counting from the top-left corner of the screen), text is
9974 yellow with a red box around it. Both the text and the box have an
9978 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
9979 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
9982 Note that the double quotes are not necessary if spaces are not used
9983 within the parameter list.
9986 Show the text at the center of the video frame:
9988 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
9992 Show the text at a random position, switching to a new position every 30 seconds:
9994 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)"
9998 Show a text line sliding from right to left in the last row of the video
9999 frame. The file @file{LONG_LINE} is assumed to contain a single line
10002 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10006 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10008 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10012 Draw a single green letter "g", at the center of the input video.
10013 The glyph baseline is placed at half screen height.
10015 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10019 Show text for 1 second every 3 seconds:
10021 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10025 Use fontconfig to set the font. Note that the colons need to be escaped.
10027 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10031 Print the date of a real-time encoding (see strftime(3)):
10033 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10037 Show text fading in and out (appearing/disappearing):
10040 DS=1.0 # display start
10041 DE=10.0 # display end
10042 FID=1.5 # fade in duration
10043 FOD=5 # fade out duration
10044 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 @}"
10048 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10049 and the @option{fontsize} value are included in the @option{y} offset.
10051 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10052 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10056 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10057 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10058 must have option @option{-export_path_metadata 1} for the special metadata fields
10059 to be available for filters.
10061 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10066 For more information about libfreetype, check:
10067 @url{http://www.freetype.org/}.
10069 For more information about fontconfig, check:
10070 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10072 For more information about libfribidi, check:
10073 @url{http://fribidi.org/}.
10075 @section edgedetect
10077 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10079 The filter accepts the following options:
10084 Set low and high threshold values used by the Canny thresholding
10087 The high threshold selects the "strong" edge pixels, which are then
10088 connected through 8-connectivity with the "weak" edge pixels selected
10089 by the low threshold.
10091 @var{low} and @var{high} threshold values must be chosen in the range
10092 [0,1], and @var{low} should be lesser or equal to @var{high}.
10094 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10098 Define the drawing mode.
10102 Draw white/gray wires on black background.
10105 Mix the colors to create a paint/cartoon effect.
10108 Apply Canny edge detector on all selected planes.
10110 Default value is @var{wires}.
10113 Select planes for filtering. By default all available planes are filtered.
10116 @subsection Examples
10120 Standard edge detection with custom values for the hysteresis thresholding:
10122 edgedetect=low=0.1:high=0.4
10126 Painting effect without thresholding:
10128 edgedetect=mode=colormix:high=0
10134 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10136 For each input image, the filter will compute the optimal mapping from
10137 the input to the output given the codebook length, that is the number
10138 of distinct output colors.
10140 This filter accepts the following options.
10143 @item codebook_length, l
10144 Set codebook length. The value must be a positive integer, and
10145 represents the number of distinct output colors. Default value is 256.
10148 Set the maximum number of iterations to apply for computing the optimal
10149 mapping. The higher the value the better the result and the higher the
10150 computation time. Default value is 1.
10153 Set a random seed, must be an integer included between 0 and
10154 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10155 will try to use a good random seed on a best effort basis.
10158 Set pal8 output pixel format. This option does not work with codebook
10159 length greater than 256.
10164 Measure graylevel entropy in histogram of color channels of video frames.
10166 It accepts the following parameters:
10170 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10172 @var{diff} mode measures entropy of histogram delta values, absolute differences
10173 between neighbour histogram values.
10177 Set brightness, contrast, saturation and approximate gamma adjustment.
10179 The filter accepts the following options:
10183 Set the contrast expression. The value must be a float value in range
10184 @code{-1000.0} to @code{1000.0}. The default value is "1".
10187 Set the brightness expression. The value must be a float value in
10188 range @code{-1.0} to @code{1.0}. The default value is "0".
10191 Set the saturation expression. The value must be a float in
10192 range @code{0.0} to @code{3.0}. The default value is "1".
10195 Set the gamma expression. The value must be a float in range
10196 @code{0.1} to @code{10.0}. The default value is "1".
10199 Set the gamma expression for red. The value must be a float in
10200 range @code{0.1} to @code{10.0}. The default value is "1".
10203 Set the gamma expression for green. The value must be a float in range
10204 @code{0.1} to @code{10.0}. The default value is "1".
10207 Set the gamma expression for blue. The value must be a float in range
10208 @code{0.1} to @code{10.0}. The default value is "1".
10211 Set the gamma weight expression. It can be used to reduce the effect
10212 of a high gamma value on bright image areas, e.g. keep them from
10213 getting overamplified and just plain white. The value must be a float
10214 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10215 gamma correction all the way down while @code{1.0} leaves it at its
10216 full strength. Default is "1".
10219 Set when the expressions for brightness, contrast, saturation and
10220 gamma expressions are evaluated.
10222 It accepts the following values:
10225 only evaluate expressions once during the filter initialization or
10226 when a command is processed
10229 evaluate expressions for each incoming frame
10232 Default value is @samp{init}.
10235 The expressions accept the following parameters:
10238 frame count of the input frame starting from 0
10241 byte position of the corresponding packet in the input file, NAN if
10245 frame rate of the input video, NAN if the input frame rate is unknown
10248 timestamp expressed in seconds, NAN if the input timestamp is unknown
10251 @subsection Commands
10252 The filter supports the following commands:
10256 Set the contrast expression.
10259 Set the brightness expression.
10262 Set the saturation expression.
10265 Set the gamma expression.
10268 Set the gamma_r expression.
10271 Set gamma_g expression.
10274 Set gamma_b expression.
10277 Set gamma_weight expression.
10279 The command accepts the same syntax of the corresponding option.
10281 If the specified expression is not valid, it is kept at its current
10288 Apply erosion effect to the video.
10290 This filter replaces the pixel by the local(3x3) minimum.
10292 It accepts the following options:
10299 Limit the maximum change for each plane, default is 65535.
10300 If 0, plane will remain unchanged.
10303 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
10306 Flags to local 3x3 coordinates maps like this:
10313 @subsection Commands
10315 This filter supports the all above options as @ref{commands}.
10317 @section extractplanes
10319 Extract color channel components from input video stream into
10320 separate grayscale video streams.
10322 The filter accepts the following option:
10326 Set plane(s) to extract.
10328 Available values for planes are:
10339 Choosing planes not available in the input will result in an error.
10340 That means you cannot select @code{r}, @code{g}, @code{b} planes
10341 with @code{y}, @code{u}, @code{v} planes at same time.
10344 @subsection Examples
10348 Extract luma, u and v color channel component from input video frame
10349 into 3 grayscale outputs:
10351 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
10357 Apply a fade-in/out effect to the input video.
10359 It accepts the following parameters:
10363 The effect type can be either "in" for a fade-in, or "out" for a fade-out
10365 Default is @code{in}.
10367 @item start_frame, s
10368 Specify the number of the frame to start applying the fade
10369 effect at. Default is 0.
10372 The number of frames that the fade effect lasts. At the end of the
10373 fade-in effect, the output video will have the same intensity as the input video.
10374 At the end of the fade-out transition, the output video will be filled with the
10375 selected @option{color}.
10379 If set to 1, fade only alpha channel, if one exists on the input.
10380 Default value is 0.
10382 @item start_time, st
10383 Specify the timestamp (in seconds) of the frame to start to apply the fade
10384 effect. If both start_frame and start_time are specified, the fade will start at
10385 whichever comes last. Default is 0.
10388 The number of seconds for which the fade effect has to last. At the end of the
10389 fade-in effect the output video will have the same intensity as the input video,
10390 at the end of the fade-out transition the output video will be filled with the
10391 selected @option{color}.
10392 If both duration and nb_frames are specified, duration is used. Default is 0
10393 (nb_frames is used by default).
10396 Specify the color of the fade. Default is "black".
10399 @subsection Examples
10403 Fade in the first 30 frames of video:
10408 The command above is equivalent to:
10414 Fade out the last 45 frames of a 200-frame video:
10417 fade=type=out:start_frame=155:nb_frames=45
10421 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
10423 fade=in:0:25, fade=out:975:25
10427 Make the first 5 frames yellow, then fade in from frame 5-24:
10429 fade=in:5:20:color=yellow
10433 Fade in alpha over first 25 frames of video:
10435 fade=in:0:25:alpha=1
10439 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
10441 fade=t=in:st=5.5:d=0.5
10447 Denoise frames using 3D FFT (frequency domain filtering).
10449 The filter accepts the following options:
10453 Set the noise sigma constant. This sets denoising strength.
10454 Default value is 1. Allowed range is from 0 to 30.
10455 Using very high sigma with low overlap may give blocking artifacts.
10458 Set amount of denoising. By default all detected noise is reduced.
10459 Default value is 1. Allowed range is from 0 to 1.
10462 Set size of block, Default is 4, can be 3, 4, 5 or 6.
10463 Actual size of block in pixels is 2 to power of @var{block}, so by default
10464 block size in pixels is 2^4 which is 16.
10467 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
10470 Set number of previous frames to use for denoising. By default is set to 0.
10473 Set number of next frames to to use for denoising. By default is set to 0.
10476 Set planes which will be filtered, by default are all available filtered
10481 Apply arbitrary expressions to samples in frequency domain
10485 Adjust the dc value (gain) of the luma plane of the image. The filter
10486 accepts an integer value in range @code{0} to @code{1000}. The default
10487 value is set to @code{0}.
10490 Adjust the dc value (gain) of the 1st chroma plane of the image. The
10491 filter accepts an integer value in range @code{0} to @code{1000}. The
10492 default value is set to @code{0}.
10495 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
10496 filter accepts an integer value in range @code{0} to @code{1000}. The
10497 default value is set to @code{0}.
10500 Set the frequency domain weight expression for the luma plane.
10503 Set the frequency domain weight expression for the 1st chroma plane.
10506 Set the frequency domain weight expression for the 2nd chroma plane.
10509 Set when the expressions are evaluated.
10511 It accepts the following values:
10514 Only evaluate expressions once during the filter initialization.
10517 Evaluate expressions for each incoming frame.
10520 Default value is @samp{init}.
10522 The filter accepts the following variables:
10525 The coordinates of the current sample.
10529 The width and height of the image.
10532 The number of input frame, starting from 0.
10535 @subsection Examples
10541 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
10547 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
10553 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
10559 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
10566 Extract a single field from an interlaced image using stride
10567 arithmetic to avoid wasting CPU time. The output frames are marked as
10570 The filter accepts the following options:
10574 Specify whether to extract the top (if the value is @code{0} or
10575 @code{top}) or the bottom field (if the value is @code{1} or
10581 Create new frames by copying the top and bottom fields from surrounding frames
10582 supplied as numbers by the hint file.
10586 Set file containing hints: absolute/relative frame numbers.
10588 There must be one line for each frame in a clip. Each line must contain two
10589 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
10590 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
10591 is current frame number for @code{absolute} mode or out of [-1, 1] range
10592 for @code{relative} mode. First number tells from which frame to pick up top
10593 field and second number tells from which frame to pick up bottom field.
10595 If optionally followed by @code{+} output frame will be marked as interlaced,
10596 else if followed by @code{-} output frame will be marked as progressive, else
10597 it will be marked same as input frame.
10598 If optionally followed by @code{t} output frame will use only top field, or in
10599 case of @code{b} it will use only bottom field.
10600 If line starts with @code{#} or @code{;} that line is skipped.
10603 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
10606 Example of first several lines of @code{hint} file for @code{relative} mode:
10608 0,0 - # first frame
10609 1,0 - # second frame, use third's frame top field and second's frame bottom field
10610 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
10625 @section fieldmatch
10627 Field matching filter for inverse telecine. It is meant to reconstruct the
10628 progressive frames from a telecined stream. The filter does not drop duplicated
10629 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
10630 followed by a decimation filter such as @ref{decimate} in the filtergraph.
10632 The separation of the field matching and the decimation is notably motivated by
10633 the possibility of inserting a de-interlacing filter fallback between the two.
10634 If the source has mixed telecined and real interlaced content,
10635 @code{fieldmatch} will not be able to match fields for the interlaced parts.
10636 But these remaining combed frames will be marked as interlaced, and thus can be
10637 de-interlaced by a later filter such as @ref{yadif} before decimation.
10639 In addition to the various configuration options, @code{fieldmatch} can take an
10640 optional second stream, activated through the @option{ppsrc} option. If
10641 enabled, the frames reconstruction will be based on the fields and frames from
10642 this second stream. This allows the first input to be pre-processed in order to
10643 help the various algorithms of the filter, while keeping the output lossless
10644 (assuming the fields are matched properly). Typically, a field-aware denoiser,
10645 or brightness/contrast adjustments can help.
10647 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
10648 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
10649 which @code{fieldmatch} is based on. While the semantic and usage are very
10650 close, some behaviour and options names can differ.
10652 The @ref{decimate} filter currently only works for constant frame rate input.
10653 If your input has mixed telecined (30fps) and progressive content with a lower
10654 framerate like 24fps use the following filterchain to produce the necessary cfr
10655 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
10657 The filter accepts the following options:
10661 Specify the assumed field order of the input stream. Available values are:
10665 Auto detect parity (use FFmpeg's internal parity value).
10667 Assume bottom field first.
10669 Assume top field first.
10672 Note that it is sometimes recommended not to trust the parity announced by the
10675 Default value is @var{auto}.
10678 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
10679 sense that it won't risk creating jerkiness due to duplicate frames when
10680 possible, but if there are bad edits or blended fields it will end up
10681 outputting combed frames when a good match might actually exist. On the other
10682 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
10683 but will almost always find a good frame if there is one. The other values are
10684 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
10685 jerkiness and creating duplicate frames versus finding good matches in sections
10686 with bad edits, orphaned fields, blended fields, etc.
10688 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
10690 Available values are:
10694 2-way matching (p/c)
10696 2-way matching, and trying 3rd match if still combed (p/c + n)
10698 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
10700 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
10701 still combed (p/c + n + u/b)
10703 3-way matching (p/c/n)
10705 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
10706 detected as combed (p/c/n + u/b)
10709 The parenthesis at the end indicate the matches that would be used for that
10710 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
10713 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
10716 Default value is @var{pc_n}.
10719 Mark the main input stream as a pre-processed input, and enable the secondary
10720 input stream as the clean source to pick the fields from. See the filter
10721 introduction for more details. It is similar to the @option{clip2} feature from
10724 Default value is @code{0} (disabled).
10727 Set the field to match from. It is recommended to set this to the same value as
10728 @option{order} unless you experience matching failures with that setting. In
10729 certain circumstances changing the field that is used to match from can have a
10730 large impact on matching performance. Available values are:
10734 Automatic (same value as @option{order}).
10736 Match from the bottom field.
10738 Match from the top field.
10741 Default value is @var{auto}.
10744 Set whether or not chroma is included during the match comparisons. In most
10745 cases it is recommended to leave this enabled. You should set this to @code{0}
10746 only if your clip has bad chroma problems such as heavy rainbowing or other
10747 artifacts. Setting this to @code{0} could also be used to speed things up at
10748 the cost of some accuracy.
10750 Default value is @code{1}.
10754 These define an exclusion band which excludes the lines between @option{y0} and
10755 @option{y1} from being included in the field matching decision. An exclusion
10756 band can be used to ignore subtitles, a logo, or other things that may
10757 interfere with the matching. @option{y0} sets the starting scan line and
10758 @option{y1} sets the ending line; all lines in between @option{y0} and
10759 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
10760 @option{y0} and @option{y1} to the same value will disable the feature.
10761 @option{y0} and @option{y1} defaults to @code{0}.
10764 Set the scene change detection threshold as a percentage of maximum change on
10765 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
10766 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
10767 @option{scthresh} is @code{[0.0, 100.0]}.
10769 Default value is @code{12.0}.
10772 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
10773 account the combed scores of matches when deciding what match to use as the
10774 final match. Available values are:
10778 No final matching based on combed scores.
10780 Combed scores are only used when a scene change is detected.
10782 Use combed scores all the time.
10785 Default is @var{sc}.
10788 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
10789 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
10790 Available values are:
10794 No forced calculation.
10796 Force p/c/n calculations.
10798 Force p/c/n/u/b calculations.
10801 Default value is @var{none}.
10804 This is the area combing threshold used for combed frame detection. This
10805 essentially controls how "strong" or "visible" combing must be to be detected.
10806 Larger values mean combing must be more visible and smaller values mean combing
10807 can be less visible or strong and still be detected. Valid settings are from
10808 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
10809 be detected as combed). This is basically a pixel difference value. A good
10810 range is @code{[8, 12]}.
10812 Default value is @code{9}.
10815 Sets whether or not chroma is considered in the combed frame decision. Only
10816 disable this if your source has chroma problems (rainbowing, etc.) that are
10817 causing problems for the combed frame detection with chroma enabled. Actually,
10818 using @option{chroma}=@var{0} is usually more reliable, except for the case
10819 where there is chroma only combing in the source.
10821 Default value is @code{0}.
10825 Respectively set the x-axis and y-axis size of the window used during combed
10826 frame detection. This has to do with the size of the area in which
10827 @option{combpel} pixels are required to be detected as combed for a frame to be
10828 declared combed. See the @option{combpel} parameter description for more info.
10829 Possible values are any number that is a power of 2 starting at 4 and going up
10832 Default value is @code{16}.
10835 The number of combed pixels inside any of the @option{blocky} by
10836 @option{blockx} size blocks on the frame for the frame to be detected as
10837 combed. While @option{cthresh} controls how "visible" the combing must be, this
10838 setting controls "how much" combing there must be in any localized area (a
10839 window defined by the @option{blockx} and @option{blocky} settings) on the
10840 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
10841 which point no frames will ever be detected as combed). This setting is known
10842 as @option{MI} in TFM/VFM vocabulary.
10844 Default value is @code{80}.
10847 @anchor{p/c/n/u/b meaning}
10848 @subsection p/c/n/u/b meaning
10850 @subsubsection p/c/n
10852 We assume the following telecined stream:
10855 Top fields: 1 2 2 3 4
10856 Bottom fields: 1 2 3 4 4
10859 The numbers correspond to the progressive frame the fields relate to. Here, the
10860 first two frames are progressive, the 3rd and 4th are combed, and so on.
10862 When @code{fieldmatch} is configured to run a matching from bottom
10863 (@option{field}=@var{bottom}) this is how this input stream get transformed:
10868 B 1 2 3 4 4 <-- matching reference
10877 As a result of the field matching, we can see that some frames get duplicated.
10878 To perform a complete inverse telecine, you need to rely on a decimation filter
10879 after this operation. See for instance the @ref{decimate} filter.
10881 The same operation now matching from top fields (@option{field}=@var{top})
10886 T 1 2 2 3 4 <-- matching reference
10896 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
10897 basically, they refer to the frame and field of the opposite parity:
10900 @item @var{p} matches the field of the opposite parity in the previous frame
10901 @item @var{c} matches the field of the opposite parity in the current frame
10902 @item @var{n} matches the field of the opposite parity in the next frame
10907 The @var{u} and @var{b} matching are a bit special in the sense that they match
10908 from the opposite parity flag. In the following examples, we assume that we are
10909 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
10910 'x' is placed above and below each matched fields.
10912 With bottom matching (@option{field}=@var{bottom}):
10917 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10918 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10926 With top matching (@option{field}=@var{top}):
10931 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10932 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10940 @subsection Examples
10942 Simple IVTC of a top field first telecined stream:
10944 fieldmatch=order=tff:combmatch=none, decimate
10947 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
10949 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
10952 @section fieldorder
10954 Transform the field order of the input video.
10956 It accepts the following parameters:
10961 The output field order. Valid values are @var{tff} for top field first or @var{bff}
10962 for bottom field first.
10965 The default value is @samp{tff}.
10967 The transformation is done by shifting the picture content up or down
10968 by one line, and filling the remaining line with appropriate picture content.
10969 This method is consistent with most broadcast field order converters.
10971 If the input video is not flagged as being interlaced, or it is already
10972 flagged as being of the required output field order, then this filter does
10973 not alter the incoming video.
10975 It is very useful when converting to or from PAL DV material,
10976 which is bottom field first.
10980 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
10983 @section fifo, afifo
10985 Buffer input images and send them when they are requested.
10987 It is mainly useful when auto-inserted by the libavfilter
10990 It does not take parameters.
10992 @section fillborders
10994 Fill borders of the input video, without changing video stream dimensions.
10995 Sometimes video can have garbage at the four edges and you may not want to
10996 crop video input to keep size multiple of some number.
10998 This filter accepts the following options:
11002 Number of pixels to fill from left border.
11005 Number of pixels to fill from right border.
11008 Number of pixels to fill from top border.
11011 Number of pixels to fill from bottom border.
11016 It accepts the following values:
11019 fill pixels using outermost pixels
11022 fill pixels using mirroring
11025 fill pixels with constant value
11028 Default is @var{smear}.
11031 Set color for pixels in fixed mode. Default is @var{black}.
11034 @subsection Commands
11035 This filter supports same @ref{commands} as options.
11036 The command accepts the same syntax of the corresponding option.
11038 If the specified expression is not valid, it is kept at its current
11043 Find a rectangular object
11045 It accepts the following options:
11049 Filepath of the object image, needs to be in gray8.
11052 Detection threshold, default is 0.5.
11055 Number of mipmaps, default is 3.
11057 @item xmin, ymin, xmax, ymax
11058 Specifies the rectangle in which to search.
11061 @subsection Examples
11065 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11067 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11073 Flood area with values of same pixel components with another values.
11075 It accepts the following options:
11078 Set pixel x coordinate.
11081 Set pixel y coordinate.
11084 Set source #0 component value.
11087 Set source #1 component value.
11090 Set source #2 component value.
11093 Set source #3 component value.
11096 Set destination #0 component value.
11099 Set destination #1 component value.
11102 Set destination #2 component value.
11105 Set destination #3 component value.
11111 Convert the input video to one of the specified pixel formats.
11112 Libavfilter will try to pick one that is suitable as input to
11115 It accepts the following parameters:
11119 A '|'-separated list of pixel format names, such as
11120 "pix_fmts=yuv420p|monow|rgb24".
11124 @subsection Examples
11128 Convert the input video to the @var{yuv420p} format
11130 format=pix_fmts=yuv420p
11133 Convert the input video to any of the formats in the list
11135 format=pix_fmts=yuv420p|yuv444p|yuv410p
11142 Convert the video to specified constant frame rate by duplicating or dropping
11143 frames as necessary.
11145 It accepts the following parameters:
11149 The desired output frame rate. The default is @code{25}.
11152 Assume the first PTS should be the given value, in seconds. This allows for
11153 padding/trimming at the start of stream. By default, no assumption is made
11154 about the first frame's expected PTS, so no padding or trimming is done.
11155 For example, this could be set to 0 to pad the beginning with duplicates of
11156 the first frame if a video stream starts after the audio stream or to trim any
11157 frames with a negative PTS.
11160 Timestamp (PTS) rounding method.
11162 Possible values are:
11169 round towards -infinity
11171 round towards +infinity
11175 The default is @code{near}.
11178 Action performed when reading the last frame.
11180 Possible values are:
11183 Use same timestamp rounding method as used for other frames.
11185 Pass through last frame if input duration has not been reached yet.
11187 The default is @code{round}.
11191 Alternatively, the options can be specified as a flat string:
11192 @var{fps}[:@var{start_time}[:@var{round}]].
11194 See also the @ref{setpts} filter.
11196 @subsection Examples
11200 A typical usage in order to set the fps to 25:
11206 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11208 fps=fps=film:round=near
11214 Pack two different video streams into a stereoscopic video, setting proper
11215 metadata on supported codecs. The two views should have the same size and
11216 framerate and processing will stop when the shorter video ends. Please note
11217 that you may conveniently adjust view properties with the @ref{scale} and
11220 It accepts the following parameters:
11224 The desired packing format. Supported values are:
11229 The views are next to each other (default).
11232 The views are on top of each other.
11235 The views are packed by line.
11238 The views are packed by column.
11241 The views are temporally interleaved.
11250 # Convert left and right views into a frame-sequential video
11251 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11253 # Convert views into a side-by-side video with the same output resolution as the input
11254 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
11259 Change the frame rate by interpolating new video output frames from the source
11262 This filter is not designed to function correctly with interlaced media. If
11263 you wish to change the frame rate of interlaced media then you are required
11264 to deinterlace before this filter and re-interlace after this filter.
11266 A description of the accepted options follows.
11270 Specify the output frames per second. This option can also be specified
11271 as a value alone. The default is @code{50}.
11274 Specify the start of a range where the output frame will be created as a
11275 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11276 the default is @code{15}.
11279 Specify the end of a range where the output frame will be created as a
11280 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11281 the default is @code{240}.
11284 Specify the level at which a scene change is detected as a value between
11285 0 and 100 to indicate a new scene; a low value reflects a low
11286 probability for the current frame to introduce a new scene, while a higher
11287 value means the current frame is more likely to be one.
11288 The default is @code{8.2}.
11291 Specify flags influencing the filter process.
11293 Available value for @var{flags} is:
11296 @item scene_change_detect, scd
11297 Enable scene change detection using the value of the option @var{scene}.
11298 This flag is enabled by default.
11304 Select one frame every N-th frame.
11306 This filter accepts the following option:
11309 Select frame after every @code{step} frames.
11310 Allowed values are positive integers higher than 0. Default value is @code{1}.
11313 @section freezedetect
11315 Detect frozen video.
11317 This filter logs a message and sets frame metadata when it detects that the
11318 input video has no significant change in content during a specified duration.
11319 Video freeze detection calculates the mean average absolute difference of all
11320 the components of video frames and compares it to a noise floor.
11322 The printed times and duration are expressed in seconds. The
11323 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
11324 whose timestamp equals or exceeds the detection duration and it contains the
11325 timestamp of the first frame of the freeze. The
11326 @code{lavfi.freezedetect.freeze_duration} and
11327 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
11330 The filter accepts the following options:
11334 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
11335 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
11339 Set freeze duration until notification (default is 2 seconds).
11342 @section freezeframes
11344 Freeze video frames.
11346 This filter freezes video frames using frame from 2nd input.
11348 The filter accepts the following options:
11352 Set number of first frame from which to start freeze.
11355 Set number of last frame from which to end freeze.
11358 Set number of frame from 2nd input which will be used instead of replaced frames.
11364 Apply a frei0r effect to the input video.
11366 To enable the compilation of this filter, you need to install the frei0r
11367 header and configure FFmpeg with @code{--enable-frei0r}.
11369 It accepts the following parameters:
11374 The name of the frei0r effect to load. If the environment variable
11375 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
11376 directories specified by the colon-separated list in @env{FREI0R_PATH}.
11377 Otherwise, the standard frei0r paths are searched, in this order:
11378 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
11379 @file{/usr/lib/frei0r-1/}.
11381 @item filter_params
11382 A '|'-separated list of parameters to pass to the frei0r effect.
11386 A frei0r effect parameter can be a boolean (its value is either
11387 "y" or "n"), a double, a color (specified as
11388 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
11389 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
11390 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
11391 a position (specified as @var{X}/@var{Y}, where
11392 @var{X} and @var{Y} are floating point numbers) and/or a string.
11394 The number and types of parameters depend on the loaded effect. If an
11395 effect parameter is not specified, the default value is set.
11397 @subsection Examples
11401 Apply the distort0r effect, setting the first two double parameters:
11403 frei0r=filter_name=distort0r:filter_params=0.5|0.01
11407 Apply the colordistance effect, taking a color as the first parameter:
11409 frei0r=colordistance:0.2/0.3/0.4
11410 frei0r=colordistance:violet
11411 frei0r=colordistance:0x112233
11415 Apply the perspective effect, specifying the top left and top right image
11418 frei0r=perspective:0.2/0.2|0.8/0.2
11422 For more information, see
11423 @url{http://frei0r.dyne.org}
11427 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
11429 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
11430 processing filter, one of them is performed once per block, not per pixel.
11431 This allows for much higher speed.
11433 The filter accepts the following options:
11437 Set quality. This option defines the number of levels for averaging. It accepts
11438 an integer in the range 4-5. Default value is @code{4}.
11441 Force a constant quantization parameter. It accepts an integer in range 0-63.
11442 If not set, the filter will use the QP from the video stream (if available).
11445 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
11446 more details but also more artifacts, while higher values make the image smoother
11447 but also blurrier. Default value is @code{0} − PSNR optimal.
11449 @item use_bframe_qp
11450 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
11451 option may cause flicker since the B-Frames have often larger QP. Default is
11452 @code{0} (not enabled).
11458 Apply Gaussian blur filter.
11460 The filter accepts the following options:
11464 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
11467 Set number of steps for Gaussian approximation. Default is @code{1}.
11470 Set which planes to filter. By default all planes are filtered.
11473 Set vertical sigma, if negative it will be same as @code{sigma}.
11474 Default is @code{-1}.
11477 @subsection Commands
11478 This filter supports same commands as options.
11479 The command accepts the same syntax of the corresponding option.
11481 If the specified expression is not valid, it is kept at its current
11486 Apply generic equation to each pixel.
11488 The filter accepts the following options:
11491 @item lum_expr, lum
11492 Set the luminance expression.
11494 Set the chrominance blue expression.
11496 Set the chrominance red expression.
11497 @item alpha_expr, a
11498 Set the alpha expression.
11500 Set the red expression.
11501 @item green_expr, g
11502 Set the green expression.
11504 Set the blue expression.
11507 The colorspace is selected according to the specified options. If one
11508 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
11509 options is specified, the filter will automatically select a YCbCr
11510 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
11511 @option{blue_expr} options is specified, it will select an RGB
11514 If one of the chrominance expression is not defined, it falls back on the other
11515 one. If no alpha expression is specified it will evaluate to opaque value.
11516 If none of chrominance expressions are specified, they will evaluate
11517 to the luminance expression.
11519 The expressions can use the following variables and functions:
11523 The sequential number of the filtered frame, starting from @code{0}.
11527 The coordinates of the current sample.
11531 The width and height of the image.
11535 Width and height scale depending on the currently filtered plane. It is the
11536 ratio between the corresponding luma plane number of pixels and the current
11537 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
11538 @code{0.5,0.5} for chroma planes.
11541 Time of the current frame, expressed in seconds.
11544 Return the value of the pixel at location (@var{x},@var{y}) of the current
11548 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
11552 Return the value of the pixel at location (@var{x},@var{y}) of the
11553 blue-difference chroma plane. Return 0 if there is no such plane.
11556 Return the value of the pixel at location (@var{x},@var{y}) of the
11557 red-difference chroma plane. Return 0 if there is no such plane.
11562 Return the value of the pixel at location (@var{x},@var{y}) of the
11563 red/green/blue component. Return 0 if there is no such component.
11566 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
11567 plane. Return 0 if there is no such plane.
11569 @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)
11570 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
11571 sums of samples within a rectangle. See the functions without the sum postfix.
11573 @item interpolation
11574 Set one of interpolation methods:
11579 Default is bilinear.
11582 For functions, if @var{x} and @var{y} are outside the area, the value will be
11583 automatically clipped to the closer edge.
11585 Please note that this filter can use multiple threads in which case each slice
11586 will have its own expression state. If you want to use only a single expression
11587 state because your expressions depend on previous state then you should limit
11588 the number of filter threads to 1.
11590 @subsection Examples
11594 Flip the image horizontally:
11600 Generate a bidimensional sine wave, with angle @code{PI/3} and a
11601 wavelength of 100 pixels:
11603 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
11607 Generate a fancy enigmatic moving light:
11609 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
11613 Generate a quick emboss effect:
11615 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
11619 Modify RGB components depending on pixel position:
11621 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
11625 Create a radial gradient that is the same size as the input (also see
11626 the @ref{vignette} filter):
11628 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
11634 Fix the banding artifacts that are sometimes introduced into nearly flat
11635 regions by truncation to 8-bit color depth.
11636 Interpolate the gradients that should go where the bands are, and
11639 It is designed for playback only. Do not use it prior to
11640 lossy compression, because compression tends to lose the dither and
11641 bring back the bands.
11643 It accepts the following parameters:
11648 The maximum amount by which the filter will change any one pixel. This is also
11649 the threshold for detecting nearly flat regions. Acceptable values range from
11650 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
11654 The neighborhood to fit the gradient to. A larger radius makes for smoother
11655 gradients, but also prevents the filter from modifying the pixels near detailed
11656 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
11657 values will be clipped to the valid range.
11661 Alternatively, the options can be specified as a flat string:
11662 @var{strength}[:@var{radius}]
11664 @subsection Examples
11668 Apply the filter with a @code{3.5} strength and radius of @code{8}:
11674 Specify radius, omitting the strength (which will fall-back to the default
11682 @anchor{graphmonitor}
11683 @section graphmonitor
11684 Show various filtergraph stats.
11686 With this filter one can debug complete filtergraph.
11687 Especially issues with links filling with queued frames.
11689 The filter accepts the following options:
11693 Set video output size. Default is @var{hd720}.
11696 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
11699 Set output mode, can be @var{fulll} or @var{compact}.
11700 In @var{compact} mode only filters with some queued frames have displayed stats.
11703 Set flags which enable which stats are shown in video.
11705 Available values for flags are:
11708 Display number of queued frames in each link.
11710 @item frame_count_in
11711 Display number of frames taken from filter.
11713 @item frame_count_out
11714 Display number of frames given out from filter.
11717 Display current filtered frame pts.
11720 Display current filtered frame time.
11723 Display time base for filter link.
11726 Display used format for filter link.
11729 Display video size or number of audio channels in case of audio used by filter link.
11732 Display video frame rate or sample rate in case of audio used by filter link.
11736 Set upper limit for video rate of output stream, Default value is @var{25}.
11737 This guarantee that output video frame rate will not be higher than this value.
11741 A color constancy variation filter which estimates scene illumination via grey edge algorithm
11742 and corrects the scene colors accordingly.
11744 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
11746 The filter accepts the following options:
11750 The order of differentiation to be applied on the scene. Must be chosen in the range
11751 [0,2] and default value is 1.
11754 The Minkowski parameter to be used for calculating the Minkowski distance. Must
11755 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
11756 max value instead of calculating Minkowski distance.
11759 The standard deviation of Gaussian blur to be applied on the scene. Must be
11760 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
11761 can't be equal to 0 if @var{difford} is greater than 0.
11764 @subsection Examples
11770 greyedge=difford=1:minknorm=5:sigma=2
11776 greyedge=difford=1:minknorm=0:sigma=2
11784 Apply a Hald CLUT to a video stream.
11786 First input is the video stream to process, and second one is the Hald CLUT.
11787 The Hald CLUT input can be a simple picture or a complete video stream.
11789 The filter accepts the following options:
11793 Force termination when the shortest input terminates. Default is @code{0}.
11795 Continue applying the last CLUT after the end of the stream. A value of
11796 @code{0} disable the filter after the last frame of the CLUT is reached.
11797 Default is @code{1}.
11800 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
11801 filters share the same internals).
11803 This filter also supports the @ref{framesync} options.
11805 More information about the Hald CLUT can be found on Eskil Steenberg's website
11806 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
11808 @subsection Workflow examples
11810 @subsubsection Hald CLUT video stream
11812 Generate an identity Hald CLUT stream altered with various effects:
11814 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
11817 Note: make sure you use a lossless codec.
11819 Then use it with @code{haldclut} to apply it on some random stream:
11821 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
11824 The Hald CLUT will be applied to the 10 first seconds (duration of
11825 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
11826 to the remaining frames of the @code{mandelbrot} stream.
11828 @subsubsection Hald CLUT with preview
11830 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
11831 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
11832 biggest possible square starting at the top left of the picture. The remaining
11833 padding pixels (bottom or right) will be ignored. This area can be used to add
11834 a preview of the Hald CLUT.
11836 Typically, the following generated Hald CLUT will be supported by the
11837 @code{haldclut} filter:
11840 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
11841 pad=iw+320 [padded_clut];
11842 smptebars=s=320x256, split [a][b];
11843 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
11844 [main][b] overlay=W-320" -frames:v 1 clut.png
11847 It contains the original and a preview of the effect of the CLUT: SMPTE color
11848 bars are displayed on the right-top, and below the same color bars processed by
11851 Then, the effect of this Hald CLUT can be visualized with:
11853 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
11858 Flip the input video horizontally.
11860 For example, to horizontally flip the input video with @command{ffmpeg}:
11862 ffmpeg -i in.avi -vf "hflip" out.avi
11866 This filter applies a global color histogram equalization on a
11869 It can be used to correct video that has a compressed range of pixel
11870 intensities. The filter redistributes the pixel intensities to
11871 equalize their distribution across the intensity range. It may be
11872 viewed as an "automatically adjusting contrast filter". This filter is
11873 useful only for correcting degraded or poorly captured source
11876 The filter accepts the following options:
11880 Determine the amount of equalization to be applied. As the strength
11881 is reduced, the distribution of pixel intensities more-and-more
11882 approaches that of the input frame. The value must be a float number
11883 in the range [0,1] and defaults to 0.200.
11886 Set the maximum intensity that can generated and scale the output
11887 values appropriately. The strength should be set as desired and then
11888 the intensity can be limited if needed to avoid washing-out. The value
11889 must be a float number in the range [0,1] and defaults to 0.210.
11892 Set the antibanding level. If enabled the filter will randomly vary
11893 the luminance of output pixels by a small amount to avoid banding of
11894 the histogram. Possible values are @code{none}, @code{weak} or
11895 @code{strong}. It defaults to @code{none}.
11901 Compute and draw a color distribution histogram for the input video.
11903 The computed histogram is a representation of the color component
11904 distribution in an image.
11906 Standard histogram displays the color components distribution in an image.
11907 Displays color graph for each color component. Shows distribution of
11908 the Y, U, V, A or R, G, B components, depending on input format, in the
11909 current frame. Below each graph a color component scale meter is shown.
11911 The filter accepts the following options:
11915 Set height of level. Default value is @code{200}.
11916 Allowed range is [50, 2048].
11919 Set height of color scale. Default value is @code{12}.
11920 Allowed range is [0, 40].
11924 It accepts the following values:
11927 Per color component graphs are placed below each other.
11930 Per color component graphs are placed side by side.
11933 Presents information identical to that in the @code{parade}, except
11934 that the graphs representing color components are superimposed directly
11937 Default is @code{stack}.
11940 Set mode. Can be either @code{linear}, or @code{logarithmic}.
11941 Default is @code{linear}.
11944 Set what color components to display.
11945 Default is @code{7}.
11948 Set foreground opacity. Default is @code{0.7}.
11951 Set background opacity. Default is @code{0.5}.
11954 @subsection Examples
11959 Calculate and draw histogram:
11961 ffplay -i input -vf histogram
11969 This is a high precision/quality 3d denoise filter. It aims to reduce
11970 image noise, producing smooth images and making still images really
11971 still. It should enhance compressibility.
11973 It accepts the following optional parameters:
11977 A non-negative floating point number which specifies spatial luma strength.
11978 It defaults to 4.0.
11980 @item chroma_spatial
11981 A non-negative floating point number which specifies spatial chroma strength.
11982 It defaults to 3.0*@var{luma_spatial}/4.0.
11985 A floating point number which specifies luma temporal strength. It defaults to
11986 6.0*@var{luma_spatial}/4.0.
11989 A floating point number which specifies chroma temporal strength. It defaults to
11990 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
11993 @subsection Commands
11994 This filter supports same @ref{commands} as options.
11995 The command accepts the same syntax of the corresponding option.
11997 If the specified expression is not valid, it is kept at its current
12000 @anchor{hwdownload}
12001 @section hwdownload
12003 Download hardware frames to system memory.
12005 The input must be in hardware frames, and the output a non-hardware format.
12006 Not all formats will be supported on the output - it may be necessary to insert
12007 an additional @option{format} filter immediately following in the graph to get
12008 the output in a supported format.
12012 Map hardware frames to system memory or to another device.
12014 This filter has several different modes of operation; which one is used depends
12015 on the input and output formats:
12018 Hardware frame input, normal frame output
12020 Map the input frames to system memory and pass them to the output. If the
12021 original hardware frame is later required (for example, after overlaying
12022 something else on part of it), the @option{hwmap} filter can be used again
12023 in the next mode to retrieve it.
12025 Normal frame input, hardware frame output
12027 If the input is actually a software-mapped hardware frame, then unmap it -
12028 that is, return the original hardware frame.
12030 Otherwise, a device must be provided. Create new hardware surfaces on that
12031 device for the output, then map them back to the software format at the input
12032 and give those frames to the preceding filter. This will then act like the
12033 @option{hwupload} filter, but may be able to avoid an additional copy when
12034 the input is already in a compatible format.
12036 Hardware frame input and output
12038 A device must be supplied for the output, either directly or with the
12039 @option{derive_device} option. The input and output devices must be of
12040 different types and compatible - the exact meaning of this is
12041 system-dependent, but typically it means that they must refer to the same
12042 underlying hardware context (for example, refer to the same graphics card).
12044 If the input frames were originally created on the output device, then unmap
12045 to retrieve the original frames.
12047 Otherwise, map the frames to the output device - create new hardware frames
12048 on the output corresponding to the frames on the input.
12051 The following additional parameters are accepted:
12055 Set the frame mapping mode. Some combination of:
12058 The mapped frame should be readable.
12060 The mapped frame should be writeable.
12062 The mapping will always overwrite the entire frame.
12064 This may improve performance in some cases, as the original contents of the
12065 frame need not be loaded.
12067 The mapping must not involve any copying.
12069 Indirect mappings to copies of frames are created in some cases where either
12070 direct mapping is not possible or it would have unexpected properties.
12071 Setting this flag ensures that the mapping is direct and will fail if that is
12074 Defaults to @var{read+write} if not specified.
12076 @item derive_device @var{type}
12077 Rather than using the device supplied at initialisation, instead derive a new
12078 device of type @var{type} from the device the input frames exist on.
12081 In a hardware to hardware mapping, map in reverse - create frames in the sink
12082 and map them back to the source. This may be necessary in some cases where
12083 a mapping in one direction is required but only the opposite direction is
12084 supported by the devices being used.
12086 This option is dangerous - it may break the preceding filter in undefined
12087 ways if there are any additional constraints on that filter's output.
12088 Do not use it without fully understanding the implications of its use.
12094 Upload system memory frames to hardware surfaces.
12096 The device to upload to must be supplied when the filter is initialised. If
12097 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12098 option or with the @option{derive_device} option. The input and output devices
12099 must be of different types and compatible - the exact meaning of this is
12100 system-dependent, but typically it means that they must refer to the same
12101 underlying hardware context (for example, refer to the same graphics card).
12103 The following additional parameters are accepted:
12106 @item derive_device @var{type}
12107 Rather than using the device supplied at initialisation, instead derive a new
12108 device of type @var{type} from the device the input frames exist on.
12111 @anchor{hwupload_cuda}
12112 @section hwupload_cuda
12114 Upload system memory frames to a CUDA device.
12116 It accepts the following optional parameters:
12120 The number of the CUDA device to use
12125 Apply a high-quality magnification filter designed for pixel art. This filter
12126 was originally created by Maxim Stepin.
12128 It accepts the following option:
12132 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12133 @code{hq3x} and @code{4} for @code{hq4x}.
12134 Default is @code{3}.
12138 Stack input videos horizontally.
12140 All streams must be of same pixel format and of same height.
12142 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12143 to create same output.
12145 The filter accepts the following option:
12149 Set number of input streams. Default is 2.
12152 If set to 1, force the output to terminate when the shortest input
12153 terminates. Default value is 0.
12158 Modify the hue and/or the saturation of the input.
12160 It accepts the following parameters:
12164 Specify the hue angle as a number of degrees. It accepts an expression,
12165 and defaults to "0".
12168 Specify the saturation in the [-10,10] range. It accepts an expression and
12172 Specify the hue angle as a number of radians. It accepts an
12173 expression, and defaults to "0".
12176 Specify the brightness in the [-10,10] range. It accepts an expression and
12180 @option{h} and @option{H} are mutually exclusive, and can't be
12181 specified at the same time.
12183 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12184 expressions containing the following constants:
12188 frame count of the input frame starting from 0
12191 presentation timestamp of the input frame expressed in time base units
12194 frame rate of the input video, NAN if the input frame rate is unknown
12197 timestamp expressed in seconds, NAN if the input timestamp is unknown
12200 time base of the input video
12203 @subsection Examples
12207 Set the hue to 90 degrees and the saturation to 1.0:
12213 Same command but expressing the hue in radians:
12219 Rotate hue and make the saturation swing between 0
12220 and 2 over a period of 1 second:
12222 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12226 Apply a 3 seconds saturation fade-in effect starting at 0:
12228 hue="s=min(t/3\,1)"
12231 The general fade-in expression can be written as:
12233 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12237 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12239 hue="s=max(0\, min(1\, (8-t)/3))"
12242 The general fade-out expression can be written as:
12244 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12249 @subsection Commands
12251 This filter supports the following commands:
12257 Modify the hue and/or the saturation and/or brightness of the input video.
12258 The command accepts the same syntax of the corresponding option.
12260 If the specified expression is not valid, it is kept at its current
12264 @section hysteresis
12266 Grow first stream into second stream by connecting components.
12267 This makes it possible to build more robust edge masks.
12269 This filter accepts the following options:
12273 Set which planes will be processed as bitmap, unprocessed planes will be
12274 copied from first stream.
12275 By default value 0xf, all planes will be processed.
12278 Set threshold which is used in filtering. If pixel component value is higher than
12279 this value filter algorithm for connecting components is activated.
12280 By default value is 0.
12283 The @code{hysteresis} filter also supports the @ref{framesync} options.
12287 Detect video interlacing type.
12289 This filter tries to detect if the input frames are interlaced, progressive,
12290 top or bottom field first. It will also try to detect fields that are
12291 repeated between adjacent frames (a sign of telecine).
12293 Single frame detection considers only immediately adjacent frames when classifying each frame.
12294 Multiple frame detection incorporates the classification history of previous frames.
12296 The filter will log these metadata values:
12299 @item single.current_frame
12300 Detected type of current frame using single-frame detection. One of:
12301 ``tff'' (top field first), ``bff'' (bottom field first),
12302 ``progressive'', or ``undetermined''
12305 Cumulative number of frames detected as top field first using single-frame detection.
12308 Cumulative number of frames detected as top field first using multiple-frame detection.
12311 Cumulative number of frames detected as bottom field first using single-frame detection.
12313 @item multiple.current_frame
12314 Detected type of current frame using multiple-frame detection. One of:
12315 ``tff'' (top field first), ``bff'' (bottom field first),
12316 ``progressive'', or ``undetermined''
12319 Cumulative number of frames detected as bottom field first using multiple-frame detection.
12321 @item single.progressive
12322 Cumulative number of frames detected as progressive using single-frame detection.
12324 @item multiple.progressive
12325 Cumulative number of frames detected as progressive using multiple-frame detection.
12327 @item single.undetermined
12328 Cumulative number of frames that could not be classified using single-frame detection.
12330 @item multiple.undetermined
12331 Cumulative number of frames that could not be classified using multiple-frame detection.
12333 @item repeated.current_frame
12334 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
12336 @item repeated.neither
12337 Cumulative number of frames with no repeated field.
12340 Cumulative number of frames with the top field repeated from the previous frame's top field.
12342 @item repeated.bottom
12343 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
12346 The filter accepts the following options:
12350 Set interlacing threshold.
12352 Set progressive threshold.
12354 Threshold for repeated field detection.
12356 Number of frames after which a given frame's contribution to the
12357 statistics is halved (i.e., it contributes only 0.5 to its
12358 classification). The default of 0 means that all frames seen are given
12359 full weight of 1.0 forever.
12360 @item analyze_interlaced_flag
12361 When this is not 0 then idet will use the specified number of frames to determine
12362 if the interlaced flag is accurate, it will not count undetermined frames.
12363 If the flag is found to be accurate it will be used without any further
12364 computations, if it is found to be inaccurate it will be cleared without any
12365 further computations. This allows inserting the idet filter as a low computational
12366 method to clean up the interlaced flag
12371 Deinterleave or interleave fields.
12373 This filter allows one to process interlaced images fields without
12374 deinterlacing them. Deinterleaving splits the input frame into 2
12375 fields (so called half pictures). Odd lines are moved to the top
12376 half of the output image, even lines to the bottom half.
12377 You can process (filter) them independently and then re-interleave them.
12379 The filter accepts the following options:
12383 @item chroma_mode, c
12384 @item alpha_mode, a
12385 Available values for @var{luma_mode}, @var{chroma_mode} and
12386 @var{alpha_mode} are:
12392 @item deinterleave, d
12393 Deinterleave fields, placing one above the other.
12395 @item interleave, i
12396 Interleave fields. Reverse the effect of deinterleaving.
12398 Default value is @code{none}.
12400 @item luma_swap, ls
12401 @item chroma_swap, cs
12402 @item alpha_swap, as
12403 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
12406 @subsection Commands
12408 This filter supports the all above options as @ref{commands}.
12412 Apply inflate effect to the video.
12414 This filter replaces the pixel by the local(3x3) average by taking into account
12415 only values higher than the pixel.
12417 It accepts the following options:
12424 Limit the maximum change for each plane, default is 65535.
12425 If 0, plane will remain unchanged.
12428 @subsection Commands
12430 This filter supports the all above options as @ref{commands}.
12434 Simple interlacing filter from progressive contents. This interleaves upper (or
12435 lower) lines from odd frames with lower (or upper) lines from even frames,
12436 halving the frame rate and preserving image height.
12439 Original Original New Frame
12440 Frame 'j' Frame 'j+1' (tff)
12441 ========== =========== ==================
12442 Line 0 --------------------> Frame 'j' Line 0
12443 Line 1 Line 1 ----> Frame 'j+1' Line 1
12444 Line 2 ---------------------> Frame 'j' Line 2
12445 Line 3 Line 3 ----> Frame 'j+1' Line 3
12447 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
12450 It accepts the following optional parameters:
12454 This determines whether the interlaced frame is taken from the even
12455 (tff - default) or odd (bff) lines of the progressive frame.
12458 Vertical lowpass filter to avoid twitter interlacing and
12459 reduce moire patterns.
12463 Disable vertical lowpass filter
12466 Enable linear filter (default)
12469 Enable complex filter. This will slightly less reduce twitter and moire
12470 but better retain detail and subjective sharpness impression.
12477 Deinterlace input video by applying Donald Graft's adaptive kernel
12478 deinterling. Work on interlaced parts of a video to produce
12479 progressive frames.
12481 The description of the accepted parameters follows.
12485 Set the threshold which affects the filter's tolerance when
12486 determining if a pixel line must be processed. It must be an integer
12487 in the range [0,255] and defaults to 10. A value of 0 will result in
12488 applying the process on every pixels.
12491 Paint pixels exceeding the threshold value to white if set to 1.
12495 Set the fields order. Swap fields if set to 1, leave fields alone if
12499 Enable additional sharpening if set to 1. Default is 0.
12502 Enable twoway sharpening if set to 1. Default is 0.
12505 @subsection Examples
12509 Apply default values:
12511 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
12515 Enable additional sharpening:
12521 Paint processed pixels in white:
12529 Slowly update darker pixels.
12531 This filter makes short flashes of light appear longer.
12532 This filter accepts the following options:
12536 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
12539 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
12542 @section lenscorrection
12544 Correct radial lens distortion
12546 This filter can be used to correct for radial distortion as can result from the use
12547 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
12548 one can use tools available for example as part of opencv or simply trial-and-error.
12549 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
12550 and extract the k1 and k2 coefficients from the resulting matrix.
12552 Note that effectively the same filter is available in the open-source tools Krita and
12553 Digikam from the KDE project.
12555 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
12556 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
12557 brightness distribution, so you may want to use both filters together in certain
12558 cases, though you will have to take care of ordering, i.e. whether vignetting should
12559 be applied before or after lens correction.
12561 @subsection Options
12563 The filter accepts the following options:
12567 Relative x-coordinate of the focal point of the image, and thereby the center of the
12568 distortion. This value has a range [0,1] and is expressed as fractions of the image
12569 width. Default is 0.5.
12571 Relative y-coordinate of the focal point of the image, and thereby the center of the
12572 distortion. This value has a range [0,1] and is expressed as fractions of the image
12573 height. Default is 0.5.
12575 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
12576 no correction. Default is 0.
12578 Coefficient of the double quadratic correction term. This value has a range [-1,1].
12579 0 means no correction. Default is 0.
12582 The formula that generates the correction is:
12584 @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)
12586 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
12587 distances from the focal point in the source and target images, respectively.
12591 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
12593 The @code{lensfun} filter requires the camera make, camera model, and lens model
12594 to apply the lens correction. The filter will load the lensfun database and
12595 query it to find the corresponding camera and lens entries in the database. As
12596 long as these entries can be found with the given options, the filter can
12597 perform corrections on frames. Note that incomplete strings will result in the
12598 filter choosing the best match with the given options, and the filter will
12599 output the chosen camera and lens models (logged with level "info"). You must
12600 provide the make, camera model, and lens model as they are required.
12602 The filter accepts the following options:
12606 The make of the camera (for example, "Canon"). This option is required.
12609 The model of the camera (for example, "Canon EOS 100D"). This option is
12613 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
12614 option is required.
12617 The type of correction to apply. The following values are valid options:
12621 Enables fixing lens vignetting.
12624 Enables fixing lens geometry. This is the default.
12627 Enables fixing chromatic aberrations.
12630 Enables fixing lens vignetting and lens geometry.
12633 Enables fixing lens vignetting and chromatic aberrations.
12636 Enables fixing both lens geometry and chromatic aberrations.
12639 Enables all possible corrections.
12643 The focal length of the image/video (zoom; expected constant for video). For
12644 example, a 18--55mm lens has focal length range of [18--55], so a value in that
12645 range should be chosen when using that lens. Default 18.
12648 The aperture of the image/video (expected constant for video). Note that
12649 aperture is only used for vignetting correction. Default 3.5.
12651 @item focus_distance
12652 The focus distance of the image/video (expected constant for video). Note that
12653 focus distance is only used for vignetting and only slightly affects the
12654 vignetting correction process. If unknown, leave it at the default value (which
12658 The scale factor which is applied after transformation. After correction the
12659 video is no longer necessarily rectangular. This parameter controls how much of
12660 the resulting image is visible. The value 0 means that a value will be chosen
12661 automatically such that there is little or no unmapped area in the output
12662 image. 1.0 means that no additional scaling is done. Lower values may result
12663 in more of the corrected image being visible, while higher values may avoid
12664 unmapped areas in the output.
12666 @item target_geometry
12667 The target geometry of the output image/video. The following values are valid
12671 @item rectilinear (default)
12674 @item equirectangular
12675 @item fisheye_orthographic
12676 @item fisheye_stereographic
12677 @item fisheye_equisolid
12678 @item fisheye_thoby
12681 Apply the reverse of image correction (instead of correcting distortion, apply
12684 @item interpolation
12685 The type of interpolation used when correcting distortion. The following values
12690 @item linear (default)
12695 @subsection Examples
12699 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
12700 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
12704 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
12708 Apply the same as before, but only for the first 5 seconds of video.
12711 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
12718 Obtain the VMAF (Video Multi-Method Assessment Fusion)
12719 score between two input videos.
12721 The obtained VMAF score is printed through the logging system.
12723 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
12724 After installing the library it can be enabled using:
12725 @code{./configure --enable-libvmaf --enable-version3}.
12726 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
12728 The filter has following options:
12732 Set the model path which is to be used for SVM.
12733 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
12736 Set the file path to be used to store logs.
12739 Set the format of the log file (xml or json).
12741 @item enable_transform
12742 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
12743 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
12744 Default value: @code{false}
12747 Invokes the phone model which will generate VMAF scores higher than in the
12748 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
12749 Default value: @code{false}
12752 Enables computing psnr along with vmaf.
12753 Default value: @code{false}
12756 Enables computing ssim along with vmaf.
12757 Default value: @code{false}
12760 Enables computing ms_ssim along with vmaf.
12761 Default value: @code{false}
12764 Set the pool method to be used for computing vmaf.
12765 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
12768 Set number of threads to be used when computing vmaf.
12769 Default value: @code{0}, which makes use of all available logical processors.
12772 Set interval for frame subsampling used when computing vmaf.
12773 Default value: @code{1}
12775 @item enable_conf_interval
12776 Enables confidence interval.
12777 Default value: @code{false}
12780 This filter also supports the @ref{framesync} options.
12782 @subsection Examples
12785 On the below examples the input file @file{main.mpg} being processed is
12786 compared with the reference file @file{ref.mpg}.
12789 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
12793 Example with options:
12795 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
12799 Example with options and different containers:
12801 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 -
12807 Limits the pixel components values to the specified range [min, max].
12809 The filter accepts the following options:
12813 Lower bound. Defaults to the lowest allowed value for the input.
12816 Upper bound. Defaults to the highest allowed value for the input.
12819 Specify which planes will be processed. Defaults to all available.
12826 The filter accepts the following options:
12830 Set the number of loops. Setting this value to -1 will result in infinite loops.
12834 Set maximal size in number of frames. Default is 0.
12837 Set first frame of loop. Default is 0.
12840 @subsection Examples
12844 Loop single first frame infinitely:
12846 loop=loop=-1:size=1:start=0
12850 Loop single first frame 10 times:
12852 loop=loop=10:size=1:start=0
12856 Loop 10 first frames 5 times:
12858 loop=loop=5:size=10:start=0
12864 Apply a 1D LUT to an input video.
12866 The filter accepts the following options:
12870 Set the 1D LUT file name.
12872 Currently supported formats:
12881 Select interpolation mode.
12883 Available values are:
12887 Use values from the nearest defined point.
12889 Interpolate values using the linear interpolation.
12891 Interpolate values using the cosine interpolation.
12893 Interpolate values using the cubic interpolation.
12895 Interpolate values using the spline interpolation.
12902 Apply a 3D LUT to an input video.
12904 The filter accepts the following options:
12908 Set the 3D LUT file name.
12910 Currently supported formats:
12924 Select interpolation mode.
12926 Available values are:
12930 Use values from the nearest defined point.
12932 Interpolate values using the 8 points defining a cube.
12934 Interpolate values using a tetrahedron.
12940 Turn certain luma values into transparency.
12942 The filter accepts the following options:
12946 Set the luma which will be used as base for transparency.
12947 Default value is @code{0}.
12950 Set the range of luma values to be keyed out.
12951 Default value is @code{0.01}.
12954 Set the range of softness. Default value is @code{0}.
12955 Use this to control gradual transition from zero to full transparency.
12958 @subsection Commands
12959 This filter supports same @ref{commands} as options.
12960 The command accepts the same syntax of the corresponding option.
12962 If the specified expression is not valid, it is kept at its current
12965 @section lut, lutrgb, lutyuv
12967 Compute a look-up table for binding each pixel component input value
12968 to an output value, and apply it to the input video.
12970 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
12971 to an RGB input video.
12973 These filters accept the following parameters:
12976 set first pixel component expression
12978 set second pixel component expression
12980 set third pixel component expression
12982 set fourth pixel component expression, corresponds to the alpha component
12985 set red component expression
12987 set green component expression
12989 set blue component expression
12991 alpha component expression
12994 set Y/luminance component expression
12996 set U/Cb component expression
12998 set V/Cr component expression
13001 Each of them specifies the expression to use for computing the lookup table for
13002 the corresponding pixel component values.
13004 The exact component associated to each of the @var{c*} options depends on the
13007 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13008 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13010 The expressions can contain the following constants and functions:
13015 The input width and height.
13018 The input value for the pixel component.
13021 The input value, clipped to the @var{minval}-@var{maxval} range.
13024 The maximum value for the pixel component.
13027 The minimum value for the pixel component.
13030 The negated value for the pixel component value, clipped to the
13031 @var{minval}-@var{maxval} range; it corresponds to the expression
13032 "maxval-clipval+minval".
13035 The computed value in @var{val}, clipped to the
13036 @var{minval}-@var{maxval} range.
13038 @item gammaval(gamma)
13039 The computed gamma correction value of the pixel component value,
13040 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13042 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13046 All expressions default to "val".
13048 @subsection Examples
13052 Negate input video:
13054 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13055 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13058 The above is the same as:
13060 lutrgb="r=negval:g=negval:b=negval"
13061 lutyuv="y=negval:u=negval:v=negval"
13071 Remove chroma components, turning the video into a graytone image:
13073 lutyuv="u=128:v=128"
13077 Apply a luma burning effect:
13083 Remove green and blue components:
13089 Set a constant alpha channel value on input:
13091 format=rgba,lutrgb=a="maxval-minval/2"
13095 Correct luminance gamma by a factor of 0.5:
13097 lutyuv=y=gammaval(0.5)
13101 Discard least significant bits of luma:
13103 lutyuv=y='bitand(val, 128+64+32)'
13107 Technicolor like effect:
13109 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13113 @section lut2, tlut2
13115 The @code{lut2} filter takes two input streams and outputs one
13118 The @code{tlut2} (time lut2) filter takes two consecutive frames
13119 from one single stream.
13121 This filter accepts the following parameters:
13124 set first pixel component expression
13126 set second pixel component expression
13128 set third pixel component expression
13130 set fourth pixel component expression, corresponds to the alpha component
13133 set output bit depth, only available for @code{lut2} filter. By default is 0,
13134 which means bit depth is automatically picked from first input format.
13137 The @code{lut2} filter also supports the @ref{framesync} options.
13139 Each of them specifies the expression to use for computing the lookup table for
13140 the corresponding pixel component values.
13142 The exact component associated to each of the @var{c*} options depends on the
13145 The expressions can contain the following constants:
13150 The input width and height.
13153 The first input value for the pixel component.
13156 The second input value for the pixel component.
13159 The first input video bit depth.
13162 The second input video bit depth.
13165 All expressions default to "x".
13167 @subsection Examples
13171 Highlight differences between two RGB video streams:
13173 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)'
13177 Highlight differences between two YUV video streams:
13179 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)'
13183 Show max difference between two video streams:
13185 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)))'
13189 @section maskedclamp
13191 Clamp the first input stream with the second input and third input stream.
13193 Returns the value of first stream to be between second input
13194 stream - @code{undershoot} and third input stream + @code{overshoot}.
13196 This filter accepts the following options:
13199 Default value is @code{0}.
13202 Default value is @code{0}.
13205 Set which planes will be processed as bitmap, unprocessed planes will be
13206 copied from first stream.
13207 By default value 0xf, all planes will be processed.
13212 Merge the second and third input stream into output stream using absolute differences
13213 between second input stream and first input stream and absolute difference between
13214 third input stream and first input stream. The picked value will be from second input
13215 stream if second absolute difference is greater than first one or from third input stream
13218 This filter accepts the following options:
13221 Set which planes will be processed as bitmap, unprocessed planes will be
13222 copied from first stream.
13223 By default value 0xf, all planes will be processed.
13226 @section maskedmerge
13228 Merge the first input stream with the second input stream using per pixel
13229 weights in the third input stream.
13231 A value of 0 in the third stream pixel component means that pixel component
13232 from first stream is returned unchanged, while maximum value (eg. 255 for
13233 8-bit videos) means that pixel component from second stream is returned
13234 unchanged. Intermediate values define the amount of merging between both
13235 input stream's pixel components.
13237 This filter accepts the following options:
13240 Set which planes will be processed as bitmap, unprocessed planes will be
13241 copied from first stream.
13242 By default value 0xf, all planes will be processed.
13247 Merge the second and third input stream into output stream using absolute differences
13248 between second input stream and first input stream and absolute difference between
13249 third input stream and first input stream. The picked value will be from second input
13250 stream if second absolute difference is less than first one or from third input stream
13253 This filter accepts the following options:
13256 Set which planes will be processed as bitmap, unprocessed planes will be
13257 copied from first stream.
13258 By default value 0xf, all planes will be processed.
13261 @section maskedthreshold
13262 Pick pixels comparing absolute difference of two video streams with fixed
13265 If absolute difference between pixel component of first and second video
13266 stream is equal or lower than user supplied threshold than pixel component
13267 from first video stream is picked, otherwise pixel component from second
13268 video stream is picked.
13270 This filter accepts the following options:
13273 Set threshold used when picking pixels from absolute difference from two input
13277 Set which planes will be processed as bitmap, unprocessed planes will be
13278 copied from second stream.
13279 By default value 0xf, all planes will be processed.
13283 Create mask from input video.
13285 For example it is useful to create motion masks after @code{tblend} filter.
13287 This filter accepts the following options:
13291 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
13294 Set high threshold. Any pixel component higher than this value will be set to max value
13295 allowed for current pixel format.
13298 Set planes to filter, by default all available planes are filtered.
13301 Fill all frame pixels with this value.
13304 Set max average pixel value for frame. If sum of all pixel components is higher that this
13305 average, output frame will be completely filled with value set by @var{fill} option.
13306 Typically useful for scene changes when used in combination with @code{tblend} filter.
13311 Apply motion-compensation deinterlacing.
13313 It needs one field per frame as input and must thus be used together
13314 with yadif=1/3 or equivalent.
13316 This filter accepts the following options:
13319 Set the deinterlacing mode.
13321 It accepts one of the following values:
13326 use iterative motion estimation
13328 like @samp{slow}, but use multiple reference frames.
13330 Default value is @samp{fast}.
13333 Set the picture field parity assumed for the input video. It must be
13334 one of the following values:
13338 assume top field first
13340 assume bottom field first
13343 Default value is @samp{bff}.
13346 Set per-block quantization parameter (QP) used by the internal
13349 Higher values should result in a smoother motion vector field but less
13350 optimal individual vectors. Default value is 1.
13355 Pick median pixel from certain rectangle defined by radius.
13357 This filter accepts the following options:
13361 Set horizontal radius size. Default value is @code{1}.
13362 Allowed range is integer from 1 to 127.
13365 Set which planes to process. Default is @code{15}, which is all available planes.
13368 Set vertical radius size. Default value is @code{0}.
13369 Allowed range is integer from 0 to 127.
13370 If it is 0, value will be picked from horizontal @code{radius} option.
13373 Set median percentile. Default value is @code{0.5}.
13374 Default value of @code{0.5} will pick always median values, while @code{0} will pick
13375 minimum values, and @code{1} maximum values.
13378 @subsection Commands
13379 This filter supports same @ref{commands} as options.
13380 The command accepts the same syntax of the corresponding option.
13382 If the specified expression is not valid, it is kept at its current
13385 @section mergeplanes
13387 Merge color channel components from several video streams.
13389 The filter accepts up to 4 input streams, and merge selected input
13390 planes to the output video.
13392 This filter accepts the following options:
13395 Set input to output plane mapping. Default is @code{0}.
13397 The mappings is specified as a bitmap. It should be specified as a
13398 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
13399 mapping for the first plane of the output stream. 'A' sets the number of
13400 the input stream to use (from 0 to 3), and 'a' the plane number of the
13401 corresponding input to use (from 0 to 3). The rest of the mappings is
13402 similar, 'Bb' describes the mapping for the output stream second
13403 plane, 'Cc' describes the mapping for the output stream third plane and
13404 'Dd' describes the mapping for the output stream fourth plane.
13407 Set output pixel format. Default is @code{yuva444p}.
13410 @subsection Examples
13414 Merge three gray video streams of same width and height into single video stream:
13416 [a0][a1][a2]mergeplanes=0x001020:yuv444p
13420 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
13422 [a0][a1]mergeplanes=0x00010210:yuva444p
13426 Swap Y and A plane in yuva444p stream:
13428 format=yuva444p,mergeplanes=0x03010200:yuva444p
13432 Swap U and V plane in yuv420p stream:
13434 format=yuv420p,mergeplanes=0x000201:yuv420p
13438 Cast a rgb24 clip to yuv444p:
13440 format=rgb24,mergeplanes=0x000102:yuv444p
13446 Estimate and export motion vectors using block matching algorithms.
13447 Motion vectors are stored in frame side data to be used by other filters.
13449 This filter accepts the following options:
13452 Specify the motion estimation method. Accepts one of the following values:
13456 Exhaustive search algorithm.
13458 Three step search algorithm.
13460 Two dimensional logarithmic search algorithm.
13462 New three step search algorithm.
13464 Four step search algorithm.
13466 Diamond search algorithm.
13468 Hexagon-based search algorithm.
13470 Enhanced predictive zonal search algorithm.
13472 Uneven multi-hexagon search algorithm.
13474 Default value is @samp{esa}.
13477 Macroblock size. Default @code{16}.
13480 Search parameter. Default @code{7}.
13483 @section midequalizer
13485 Apply Midway Image Equalization effect using two video streams.
13487 Midway Image Equalization adjusts a pair of images to have the same
13488 histogram, while maintaining their dynamics as much as possible. It's
13489 useful for e.g. matching exposures from a pair of stereo cameras.
13491 This filter has two inputs and one output, which must be of same pixel format, but
13492 may be of different sizes. The output of filter is first input adjusted with
13493 midway histogram of both inputs.
13495 This filter accepts the following option:
13499 Set which planes to process. Default is @code{15}, which is all available planes.
13502 @section minterpolate
13504 Convert the video to specified frame rate using motion interpolation.
13506 This filter accepts the following options:
13509 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}.
13512 Motion interpolation mode. Following values are accepted:
13515 Duplicate previous or next frame for interpolating new ones.
13517 Blend source frames. Interpolated frame is mean of previous and next frames.
13519 Motion compensated interpolation. Following options are effective when this mode is selected:
13523 Motion compensation mode. Following values are accepted:
13526 Overlapped block motion compensation.
13528 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
13530 Default mode is @samp{obmc}.
13533 Motion estimation mode. Following values are accepted:
13536 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
13538 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
13540 Default mode is @samp{bilat}.
13543 The algorithm to be used for motion estimation. Following values are accepted:
13546 Exhaustive search algorithm.
13548 Three step search algorithm.
13550 Two dimensional logarithmic search algorithm.
13552 New three step search algorithm.
13554 Four step search algorithm.
13556 Diamond search algorithm.
13558 Hexagon-based search algorithm.
13560 Enhanced predictive zonal search algorithm.
13562 Uneven multi-hexagon search algorithm.
13564 Default algorithm is @samp{epzs}.
13567 Macroblock size. Default @code{16}.
13570 Motion estimation search parameter. Default @code{32}.
13573 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).
13578 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:
13581 Disable scene change detection.
13583 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
13585 Default method is @samp{fdiff}.
13587 @item scd_threshold
13588 Scene change detection threshold. Default is @code{5.0}.
13593 Mix several video input streams into one video stream.
13595 A description of the accepted options follows.
13599 The number of inputs. If unspecified, it defaults to 2.
13602 Specify weight of each input video stream as sequence.
13603 Each weight is separated by space. If number of weights
13604 is smaller than number of @var{frames} last specified
13605 weight will be used for all remaining unset weights.
13608 Specify scale, if it is set it will be multiplied with sum
13609 of each weight multiplied with pixel values to give final destination
13610 pixel value. By default @var{scale} is auto scaled to sum of weights.
13613 Specify how end of stream is determined.
13616 The duration of the longest input. (default)
13619 The duration of the shortest input.
13622 The duration of the first input.
13626 @section mpdecimate
13628 Drop frames that do not differ greatly from the previous frame in
13629 order to reduce frame rate.
13631 The main use of this filter is for very-low-bitrate encoding
13632 (e.g. streaming over dialup modem), but it could in theory be used for
13633 fixing movies that were inverse-telecined incorrectly.
13635 A description of the accepted options follows.
13639 Set the maximum number of consecutive frames which can be dropped (if
13640 positive), or the minimum interval between dropped frames (if
13641 negative). If the value is 0, the frame is dropped disregarding the
13642 number of previous sequentially dropped frames.
13644 Default value is 0.
13649 Set the dropping threshold values.
13651 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
13652 represent actual pixel value differences, so a threshold of 64
13653 corresponds to 1 unit of difference for each pixel, or the same spread
13654 out differently over the block.
13656 A frame is a candidate for dropping if no 8x8 blocks differ by more
13657 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
13658 meaning the whole image) differ by more than a threshold of @option{lo}.
13660 Default value for @option{hi} is 64*12, default value for @option{lo} is
13661 64*5, and default value for @option{frac} is 0.33.
13667 Negate (invert) the input video.
13669 It accepts the following option:
13674 With value 1, it negates the alpha component, if present. Default value is 0.
13680 Denoise frames using Non-Local Means algorithm.
13682 Each pixel is adjusted by looking for other pixels with similar contexts. This
13683 context similarity is defined by comparing their surrounding patches of size
13684 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
13687 Note that the research area defines centers for patches, which means some
13688 patches will be made of pixels outside that research area.
13690 The filter accepts the following options.
13694 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
13697 Set patch size. Default is 7. Must be odd number in range [0, 99].
13700 Same as @option{p} but for chroma planes.
13702 The default value is @var{0} and means automatic.
13705 Set research size. Default is 15. Must be odd number in range [0, 99].
13708 Same as @option{r} but for chroma planes.
13710 The default value is @var{0} and means automatic.
13715 Deinterlace video using neural network edge directed interpolation.
13717 This filter accepts the following options:
13721 Mandatory option, without binary file filter can not work.
13722 Currently file can be found here:
13723 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
13726 Set which frames to deinterlace, by default it is @code{all}.
13727 Can be @code{all} or @code{interlaced}.
13730 Set mode of operation.
13732 Can be one of the following:
13736 Use frame flags, both fields.
13738 Use frame flags, single field.
13740 Use top field only.
13742 Use bottom field only.
13744 Use both fields, top first.
13746 Use both fields, bottom first.
13750 Set which planes to process, by default filter process all frames.
13753 Set size of local neighborhood around each pixel, used by the predictor neural
13756 Can be one of the following:
13769 Set the number of neurons in predictor neural network.
13770 Can be one of the following:
13781 Controls the number of different neural network predictions that are blended
13782 together to compute the final output value. Can be @code{fast}, default or
13786 Set which set of weights to use in the predictor.
13787 Can be one of the following:
13791 weights trained to minimize absolute error
13793 weights trained to minimize squared error
13797 Controls whether or not the prescreener neural network is used to decide
13798 which pixels should be processed by the predictor neural network and which
13799 can be handled by simple cubic interpolation.
13800 The prescreener is trained to know whether cubic interpolation will be
13801 sufficient for a pixel or whether it should be predicted by the predictor nn.
13802 The computational complexity of the prescreener nn is much less than that of
13803 the predictor nn. Since most pixels can be handled by cubic interpolation,
13804 using the prescreener generally results in much faster processing.
13805 The prescreener is pretty accurate, so the difference between using it and not
13806 using it is almost always unnoticeable.
13808 Can be one of the following:
13816 Default is @code{new}.
13819 Set various debugging flags.
13824 Force libavfilter not to use any of the specified pixel formats for the
13825 input to the next filter.
13827 It accepts the following parameters:
13831 A '|'-separated list of pixel format names, such as
13832 pix_fmts=yuv420p|monow|rgb24".
13836 @subsection Examples
13840 Force libavfilter to use a format different from @var{yuv420p} for the
13841 input to the vflip filter:
13843 noformat=pix_fmts=yuv420p,vflip
13847 Convert the input video to any of the formats not contained in the list:
13849 noformat=yuv420p|yuv444p|yuv410p
13855 Add noise on video input frame.
13857 The filter accepts the following options:
13865 Set noise seed for specific pixel component or all pixel components in case
13866 of @var{all_seed}. Default value is @code{123457}.
13868 @item all_strength, alls
13869 @item c0_strength, c0s
13870 @item c1_strength, c1s
13871 @item c2_strength, c2s
13872 @item c3_strength, c3s
13873 Set noise strength for specific pixel component or all pixel components in case
13874 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
13876 @item all_flags, allf
13877 @item c0_flags, c0f
13878 @item c1_flags, c1f
13879 @item c2_flags, c2f
13880 @item c3_flags, c3f
13881 Set pixel component flags or set flags for all components if @var{all_flags}.
13882 Available values for component flags are:
13885 averaged temporal noise (smoother)
13887 mix random noise with a (semi)regular pattern
13889 temporal noise (noise pattern changes between frames)
13891 uniform noise (gaussian otherwise)
13895 @subsection Examples
13897 Add temporal and uniform noise to input video:
13899 noise=alls=20:allf=t+u
13904 Normalize RGB video (aka histogram stretching, contrast stretching).
13905 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
13907 For each channel of each frame, the filter computes the input range and maps
13908 it linearly to the user-specified output range. The output range defaults
13909 to the full dynamic range from pure black to pure white.
13911 Temporal smoothing can be used on the input range to reduce flickering (rapid
13912 changes in brightness) caused when small dark or bright objects enter or leave
13913 the scene. This is similar to the auto-exposure (automatic gain control) on a
13914 video camera, and, like a video camera, it may cause a period of over- or
13915 under-exposure of the video.
13917 The R,G,B channels can be normalized independently, which may cause some
13918 color shifting, or linked together as a single channel, which prevents
13919 color shifting. Linked normalization preserves hue. Independent normalization
13920 does not, so it can be used to remove some color casts. Independent and linked
13921 normalization can be combined in any ratio.
13923 The normalize filter accepts the following options:
13928 Colors which define the output range. The minimum input value is mapped to
13929 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
13930 The defaults are black and white respectively. Specifying white for
13931 @var{blackpt} and black for @var{whitept} will give color-inverted,
13932 normalized video. Shades of grey can be used to reduce the dynamic range
13933 (contrast). Specifying saturated colors here can create some interesting
13937 The number of previous frames to use for temporal smoothing. The input range
13938 of each channel is smoothed using a rolling average over the current frame
13939 and the @var{smoothing} previous frames. The default is 0 (no temporal
13943 Controls the ratio of independent (color shifting) channel normalization to
13944 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
13945 independent. Defaults to 1.0 (fully independent).
13948 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
13949 expensive no-op. Defaults to 1.0 (full strength).
13953 @subsection Commands
13954 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
13955 The command accepts the same syntax of the corresponding option.
13957 If the specified expression is not valid, it is kept at its current
13960 @subsection Examples
13962 Stretch video contrast to use the full dynamic range, with no temporal
13963 smoothing; may flicker depending on the source content:
13965 normalize=blackpt=black:whitept=white:smoothing=0
13968 As above, but with 50 frames of temporal smoothing; flicker should be
13969 reduced, depending on the source content:
13971 normalize=blackpt=black:whitept=white:smoothing=50
13974 As above, but with hue-preserving linked channel normalization:
13976 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
13979 As above, but with half strength:
13981 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
13984 Map the darkest input color to red, the brightest input color to cyan:
13986 normalize=blackpt=red:whitept=cyan
13991 Pass the video source unchanged to the output.
13994 Optical Character Recognition
13996 This filter uses Tesseract for optical character recognition. To enable
13997 compilation of this filter, you need to configure FFmpeg with
13998 @code{--enable-libtesseract}.
14000 It accepts the following options:
14004 Set datapath to tesseract data. Default is to use whatever was
14005 set at installation.
14008 Set language, default is "eng".
14011 Set character whitelist.
14014 Set character blacklist.
14017 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14018 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14022 Apply a video transform using libopencv.
14024 To enable this filter, install the libopencv library and headers and
14025 configure FFmpeg with @code{--enable-libopencv}.
14027 It accepts the following parameters:
14032 The name of the libopencv filter to apply.
14034 @item filter_params
14035 The parameters to pass to the libopencv filter. If not specified, the default
14036 values are assumed.
14040 Refer to the official libopencv documentation for more precise
14042 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14044 Several libopencv filters are supported; see the following subsections.
14049 Dilate an image by using a specific structuring element.
14050 It corresponds to the libopencv function @code{cvDilate}.
14052 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14054 @var{struct_el} represents a structuring element, and has the syntax:
14055 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14057 @var{cols} and @var{rows} represent the number of columns and rows of
14058 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14059 point, and @var{shape} the shape for the structuring element. @var{shape}
14060 must be "rect", "cross", "ellipse", or "custom".
14062 If the value for @var{shape} is "custom", it must be followed by a
14063 string of the form "=@var{filename}". The file with name
14064 @var{filename} is assumed to represent a binary image, with each
14065 printable character corresponding to a bright pixel. When a custom
14066 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14067 or columns and rows of the read file are assumed instead.
14069 The default value for @var{struct_el} is "3x3+0x0/rect".
14071 @var{nb_iterations} specifies the number of times the transform is
14072 applied to the image, and defaults to 1.
14076 # Use the default values
14079 # Dilate using a structuring element with a 5x5 cross, iterating two times
14080 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14082 # Read the shape from the file diamond.shape, iterating two times.
14083 # The file diamond.shape may contain a pattern of characters like this
14089 # The specified columns and rows are ignored
14090 # but the anchor point coordinates are not
14091 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14096 Erode an image by using a specific structuring element.
14097 It corresponds to the libopencv function @code{cvErode}.
14099 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14100 with the same syntax and semantics as the @ref{dilate} filter.
14104 Smooth the input video.
14106 The filter takes the following parameters:
14107 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14109 @var{type} is the type of smooth filter to apply, and must be one of
14110 the following values: "blur", "blur_no_scale", "median", "gaussian",
14111 or "bilateral". The default value is "gaussian".
14113 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14114 depends on the smooth type. @var{param1} and
14115 @var{param2} accept integer positive values or 0. @var{param3} and
14116 @var{param4} accept floating point values.
14118 The default value for @var{param1} is 3. The default value for the
14119 other parameters is 0.
14121 These parameters correspond to the parameters assigned to the
14122 libopencv function @code{cvSmooth}.
14124 @section oscilloscope
14126 2D Video Oscilloscope.
14128 Useful to measure spatial impulse, step responses, chroma delays, etc.
14130 It accepts the following parameters:
14134 Set scope center x position.
14137 Set scope center y position.
14140 Set scope size, relative to frame diagonal.
14143 Set scope tilt/rotation.
14149 Set trace center x position.
14152 Set trace center y position.
14155 Set trace width, relative to width of frame.
14158 Set trace height, relative to height of frame.
14161 Set which components to trace. By default it traces first three components.
14164 Draw trace grid. By default is enabled.
14167 Draw some statistics. By default is enabled.
14170 Draw scope. By default is enabled.
14173 @subsection Commands
14174 This filter supports same @ref{commands} as options.
14175 The command accepts the same syntax of the corresponding option.
14177 If the specified expression is not valid, it is kept at its current
14180 @subsection Examples
14184 Inspect full first row of video frame.
14186 oscilloscope=x=0.5:y=0:s=1
14190 Inspect full last row of video frame.
14192 oscilloscope=x=0.5:y=1:s=1
14196 Inspect full 5th line of video frame of height 1080.
14198 oscilloscope=x=0.5:y=5/1080:s=1
14202 Inspect full last column of video frame.
14204 oscilloscope=x=1:y=0.5:s=1:t=1
14212 Overlay one video on top of another.
14214 It takes two inputs and has one output. The first input is the "main"
14215 video on which the second input is overlaid.
14217 It accepts the following parameters:
14219 A description of the accepted options follows.
14224 Set the expression for the x and y coordinates of the overlaid video
14225 on the main video. Default value is "0" for both expressions. In case
14226 the expression is invalid, it is set to a huge value (meaning that the
14227 overlay will not be displayed within the output visible area).
14230 See @ref{framesync}.
14233 Set when the expressions for @option{x}, and @option{y} are evaluated.
14235 It accepts the following values:
14238 only evaluate expressions once during the filter initialization or
14239 when a command is processed
14242 evaluate expressions for each incoming frame
14245 Default value is @samp{frame}.
14248 See @ref{framesync}.
14251 Set the format for the output video.
14253 It accepts the following values:
14256 force YUV420 output
14259 force YUV422 output
14262 force YUV444 output
14265 force packed RGB output
14268 force planar RGB output
14271 automatically pick format
14274 Default value is @samp{yuv420}.
14277 See @ref{framesync}.
14280 Set format of alpha of the overlaid video, it can be @var{straight} or
14281 @var{premultiplied}. Default is @var{straight}.
14284 The @option{x}, and @option{y} expressions can contain the following
14290 The main input width and height.
14294 The overlay input width and height.
14298 The computed values for @var{x} and @var{y}. They are evaluated for
14303 horizontal and vertical chroma subsample values of the output
14304 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
14308 the number of input frame, starting from 0
14311 the position in the file of the input frame, NAN if unknown
14314 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
14318 This filter also supports the @ref{framesync} options.
14320 Note that the @var{n}, @var{pos}, @var{t} variables are available only
14321 when evaluation is done @emph{per frame}, and will evaluate to NAN
14322 when @option{eval} is set to @samp{init}.
14324 Be aware that frames are taken from each input video in timestamp
14325 order, hence, if their initial timestamps differ, it is a good idea
14326 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
14327 have them begin in the same zero timestamp, as the example for
14328 the @var{movie} filter does.
14330 You can chain together more overlays but you should test the
14331 efficiency of such approach.
14333 @subsection Commands
14335 This filter supports the following commands:
14339 Modify the x and y of the overlay input.
14340 The command accepts the same syntax of the corresponding option.
14342 If the specified expression is not valid, it is kept at its current
14346 @subsection Examples
14350 Draw the overlay at 10 pixels from the bottom right corner of the main
14353 overlay=main_w-overlay_w-10:main_h-overlay_h-10
14356 Using named options the example above becomes:
14358 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
14362 Insert a transparent PNG logo in the bottom left corner of the input,
14363 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
14365 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
14369 Insert 2 different transparent PNG logos (second logo on bottom
14370 right corner) using the @command{ffmpeg} tool:
14372 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
14376 Add a transparent color layer on top of the main video; @code{WxH}
14377 must specify the size of the main input to the overlay filter:
14379 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
14383 Play an original video and a filtered version (here with the deshake
14384 filter) side by side using the @command{ffplay} tool:
14386 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
14389 The above command is the same as:
14391 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
14395 Make a sliding overlay appearing from the left to the right top part of the
14396 screen starting since time 2:
14398 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
14402 Compose output by putting two input videos side to side:
14404 ffmpeg -i left.avi -i right.avi -filter_complex "
14405 nullsrc=size=200x100 [background];
14406 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
14407 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
14408 [background][left] overlay=shortest=1 [background+left];
14409 [background+left][right] overlay=shortest=1:x=100 [left+right]
14414 Mask 10-20 seconds of a video by applying the delogo filter to a section
14416 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
14417 -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]'
14422 Chain several overlays in cascade:
14424 nullsrc=s=200x200 [bg];
14425 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
14426 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
14427 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
14428 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
14429 [in3] null, [mid2] overlay=100:100 [out0]
14434 @anchor{overlay_cuda}
14435 @section overlay_cuda
14437 Overlay one video on top of another.
14439 This is the CUDA cariant of the @ref{overlay} filter.
14440 It only accepts CUDA frames. The underlying input pixel formats have to match.
14442 It takes two inputs and has one output. The first input is the "main"
14443 video on which the second input is overlaid.
14445 It accepts the following parameters:
14450 Set the x and y coordinates of the overlaid video on the main video.
14451 Default value is "0" for both expressions.
14454 See @ref{framesync}.
14457 See @ref{framesync}.
14460 See @ref{framesync}.
14464 This filter also supports the @ref{framesync} options.
14468 Apply Overcomplete Wavelet denoiser.
14470 The filter accepts the following options:
14476 Larger depth values will denoise lower frequency components more, but
14477 slow down filtering.
14479 Must be an int in the range 8-16, default is @code{8}.
14481 @item luma_strength, ls
14484 Must be a double value in the range 0-1000, default is @code{1.0}.
14486 @item chroma_strength, cs
14487 Set chroma strength.
14489 Must be a double value in the range 0-1000, default is @code{1.0}.
14495 Add paddings to the input image, and place the original input at the
14496 provided @var{x}, @var{y} coordinates.
14498 It accepts the following parameters:
14503 Specify an expression for the size of the output image with the
14504 paddings added. If the value for @var{width} or @var{height} is 0, the
14505 corresponding input size is used for the output.
14507 The @var{width} expression can reference the value set by the
14508 @var{height} expression, and vice versa.
14510 The default value of @var{width} and @var{height} is 0.
14514 Specify the offsets to place the input image at within the padded area,
14515 with respect to the top/left border of the output image.
14517 The @var{x} expression can reference the value set by the @var{y}
14518 expression, and vice versa.
14520 The default value of @var{x} and @var{y} is 0.
14522 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
14523 so the input image is centered on the padded area.
14526 Specify the color of the padded area. For the syntax of this option,
14527 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
14528 manual,ffmpeg-utils}.
14530 The default value of @var{color} is "black".
14533 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
14535 It accepts the following values:
14539 Only evaluate expressions once during the filter initialization or when
14540 a command is processed.
14543 Evaluate expressions for each incoming frame.
14547 Default value is @samp{init}.
14550 Pad to aspect instead to a resolution.
14554 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
14555 options are expressions containing the following constants:
14560 The input video width and height.
14564 These are the same as @var{in_w} and @var{in_h}.
14568 The output width and height (the size of the padded area), as
14569 specified by the @var{width} and @var{height} expressions.
14573 These are the same as @var{out_w} and @var{out_h}.
14577 The x and y offsets as specified by the @var{x} and @var{y}
14578 expressions, or NAN if not yet specified.
14581 same as @var{iw} / @var{ih}
14584 input sample aspect ratio
14587 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
14591 The horizontal and vertical chroma subsample values. For example for the
14592 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14595 @subsection Examples
14599 Add paddings with the color "violet" to the input video. The output video
14600 size is 640x480, and the top-left corner of the input video is placed at
14603 pad=640:480:0:40:violet
14606 The example above is equivalent to the following command:
14608 pad=width=640:height=480:x=0:y=40:color=violet
14612 Pad the input to get an output with dimensions increased by 3/2,
14613 and put the input video at the center of the padded area:
14615 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
14619 Pad the input to get a squared output with size equal to the maximum
14620 value between the input width and height, and put the input video at
14621 the center of the padded area:
14623 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
14627 Pad the input to get a final w/h ratio of 16:9:
14629 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
14633 In case of anamorphic video, in order to set the output display aspect
14634 correctly, it is necessary to use @var{sar} in the expression,
14635 according to the relation:
14637 (ih * X / ih) * sar = output_dar
14638 X = output_dar / sar
14641 Thus the previous example needs to be modified to:
14643 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
14647 Double the output size and put the input video in the bottom-right
14648 corner of the output padded area:
14650 pad="2*iw:2*ih:ow-iw:oh-ih"
14654 @anchor{palettegen}
14655 @section palettegen
14657 Generate one palette for a whole video stream.
14659 It accepts the following options:
14663 Set the maximum number of colors to quantize in the palette.
14664 Note: the palette will still contain 256 colors; the unused palette entries
14667 @item reserve_transparent
14668 Create a palette of 255 colors maximum and reserve the last one for
14669 transparency. Reserving the transparency color is useful for GIF optimization.
14670 If not set, the maximum of colors in the palette will be 256. You probably want
14671 to disable this option for a standalone image.
14674 @item transparency_color
14675 Set the color that will be used as background for transparency.
14678 Set statistics mode.
14680 It accepts the following values:
14683 Compute full frame histograms.
14685 Compute histograms only for the part that differs from previous frame. This
14686 might be relevant to give more importance to the moving part of your input if
14687 the background is static.
14689 Compute new histogram for each frame.
14692 Default value is @var{full}.
14695 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
14696 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
14697 color quantization of the palette. This information is also visible at
14698 @var{info} logging level.
14700 @subsection Examples
14704 Generate a representative palette of a given video using @command{ffmpeg}:
14706 ffmpeg -i input.mkv -vf palettegen palette.png
14710 @section paletteuse
14712 Use a palette to downsample an input video stream.
14714 The filter takes two inputs: one video stream and a palette. The palette must
14715 be a 256 pixels image.
14717 It accepts the following options:
14721 Select dithering mode. Available algorithms are:
14724 Ordered 8x8 bayer dithering (deterministic)
14726 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
14727 Note: this dithering is sometimes considered "wrong" and is included as a
14729 @item floyd_steinberg
14730 Floyd and Steingberg dithering (error diffusion)
14732 Frankie Sierra dithering v2 (error diffusion)
14734 Frankie Sierra dithering v2 "Lite" (error diffusion)
14737 Default is @var{sierra2_4a}.
14740 When @var{bayer} dithering is selected, this option defines the scale of the
14741 pattern (how much the crosshatch pattern is visible). A low value means more
14742 visible pattern for less banding, and higher value means less visible pattern
14743 at the cost of more banding.
14745 The option must be an integer value in the range [0,5]. Default is @var{2}.
14748 If set, define the zone to process
14752 Only the changing rectangle will be reprocessed. This is similar to GIF
14753 cropping/offsetting compression mechanism. This option can be useful for speed
14754 if only a part of the image is changing, and has use cases such as limiting the
14755 scope of the error diffusal @option{dither} to the rectangle that bounds the
14756 moving scene (it leads to more deterministic output if the scene doesn't change
14757 much, and as a result less moving noise and better GIF compression).
14760 Default is @var{none}.
14763 Take new palette for each output frame.
14765 @item alpha_threshold
14766 Sets the alpha threshold for transparency. Alpha values above this threshold
14767 will be treated as completely opaque, and values below this threshold will be
14768 treated as completely transparent.
14770 The option must be an integer value in the range [0,255]. Default is @var{128}.
14773 @subsection Examples
14777 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
14778 using @command{ffmpeg}:
14780 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
14784 @section perspective
14786 Correct perspective of video not recorded perpendicular to the screen.
14788 A description of the accepted parameters follows.
14799 Set coordinates expression for top left, top right, bottom left and bottom right corners.
14800 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
14801 If the @code{sense} option is set to @code{source}, then the specified points will be sent
14802 to the corners of the destination. If the @code{sense} option is set to @code{destination},
14803 then the corners of the source will be sent to the specified coordinates.
14805 The expressions can use the following variables:
14810 the width and height of video frame.
14814 Output frame count.
14817 @item interpolation
14818 Set interpolation for perspective correction.
14820 It accepts the following values:
14826 Default value is @samp{linear}.
14829 Set interpretation of coordinate options.
14831 It accepts the following values:
14835 Send point in the source specified by the given coordinates to
14836 the corners of the destination.
14838 @item 1, destination
14840 Send the corners of the source to the point in the destination specified
14841 by the given coordinates.
14843 Default value is @samp{source}.
14847 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
14849 It accepts the following values:
14852 only evaluate expressions once during the filter initialization or
14853 when a command is processed
14856 evaluate expressions for each incoming frame
14859 Default value is @samp{init}.
14864 Delay interlaced video by one field time so that the field order changes.
14866 The intended use is to fix PAL movies that have been captured with the
14867 opposite field order to the film-to-video transfer.
14869 A description of the accepted parameters follows.
14875 It accepts the following values:
14878 Capture field order top-first, transfer bottom-first.
14879 Filter will delay the bottom field.
14882 Capture field order bottom-first, transfer top-first.
14883 Filter will delay the top field.
14886 Capture and transfer with the same field order. This mode only exists
14887 for the documentation of the other options to refer to, but if you
14888 actually select it, the filter will faithfully do nothing.
14891 Capture field order determined automatically by field flags, transfer
14893 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
14894 basis using field flags. If no field information is available,
14895 then this works just like @samp{u}.
14898 Capture unknown or varying, transfer opposite.
14899 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
14900 analyzing the images and selecting the alternative that produces best
14901 match between the fields.
14904 Capture top-first, transfer unknown or varying.
14905 Filter selects among @samp{t} and @samp{p} using image analysis.
14908 Capture bottom-first, transfer unknown or varying.
14909 Filter selects among @samp{b} and @samp{p} using image analysis.
14912 Capture determined by field flags, transfer unknown or varying.
14913 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
14914 image analysis. If no field information is available, then this works just
14915 like @samp{U}. This is the default mode.
14918 Both capture and transfer unknown or varying.
14919 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
14923 @section photosensitivity
14924 Reduce various flashes in video, so to help users with epilepsy.
14926 It accepts the following options:
14929 Set how many frames to use when filtering. Default is 30.
14932 Set detection threshold factor. Default is 1.
14936 Set how many pixels to skip when sampling frames. Default is 1.
14937 Allowed range is from 1 to 1024.
14940 Leave frames unchanged. Default is disabled.
14943 @section pixdesctest
14945 Pixel format descriptor test filter, mainly useful for internal
14946 testing. The output video should be equal to the input video.
14950 format=monow, pixdesctest
14953 can be used to test the monowhite pixel format descriptor definition.
14957 Display sample values of color channels. Mainly useful for checking color
14958 and levels. Minimum supported resolution is 640x480.
14960 The filters accept the following options:
14964 Set scope X position, relative offset on X axis.
14967 Set scope Y position, relative offset on Y axis.
14976 Set window opacity. This window also holds statistics about pixel area.
14979 Set window X position, relative offset on X axis.
14982 Set window Y position, relative offset on Y axis.
14987 Enable the specified chain of postprocessing subfilters using libpostproc. This
14988 library should be automatically selected with a GPL build (@code{--enable-gpl}).
14989 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
14990 Each subfilter and some options have a short and a long name that can be used
14991 interchangeably, i.e. dr/dering are the same.
14993 The filters accept the following options:
14997 Set postprocessing subfilters string.
15000 All subfilters share common options to determine their scope:
15004 Honor the quality commands for this subfilter.
15007 Do chrominance filtering, too (default).
15010 Do luminance filtering only (no chrominance).
15013 Do chrominance filtering only (no luminance).
15016 These options can be appended after the subfilter name, separated by a '|'.
15018 Available subfilters are:
15021 @item hb/hdeblock[|difference[|flatness]]
15022 Horizontal deblocking filter
15025 Difference factor where higher values mean more deblocking (default: @code{32}).
15027 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15030 @item vb/vdeblock[|difference[|flatness]]
15031 Vertical deblocking filter
15034 Difference factor where higher values mean more deblocking (default: @code{32}).
15036 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15039 @item ha/hadeblock[|difference[|flatness]]
15040 Accurate horizontal deblocking filter
15043 Difference factor where higher values mean more deblocking (default: @code{32}).
15045 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15048 @item va/vadeblock[|difference[|flatness]]
15049 Accurate vertical deblocking filter
15052 Difference factor where higher values mean more deblocking (default: @code{32}).
15054 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15058 The horizontal and vertical deblocking filters share the difference and
15059 flatness values so you cannot set different horizontal and vertical
15063 @item h1/x1hdeblock
15064 Experimental horizontal deblocking filter
15066 @item v1/x1vdeblock
15067 Experimental vertical deblocking filter
15072 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15075 larger -> stronger filtering
15077 larger -> stronger filtering
15079 larger -> stronger filtering
15082 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15085 Stretch luminance to @code{0-255}.
15088 @item lb/linblenddeint
15089 Linear blend deinterlacing filter that deinterlaces the given block by
15090 filtering all lines with a @code{(1 2 1)} filter.
15092 @item li/linipoldeint
15093 Linear interpolating deinterlacing filter that deinterlaces the given block by
15094 linearly interpolating every second line.
15096 @item ci/cubicipoldeint
15097 Cubic interpolating deinterlacing filter deinterlaces the given block by
15098 cubically interpolating every second line.
15100 @item md/mediandeint
15101 Median deinterlacing filter that deinterlaces the given block by applying a
15102 median filter to every second line.
15104 @item fd/ffmpegdeint
15105 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15106 second line with a @code{(-1 4 2 4 -1)} filter.
15109 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15110 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15112 @item fq/forceQuant[|quantizer]
15113 Overrides the quantizer table from the input with the constant quantizer you
15121 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15124 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15127 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15130 @subsection Examples
15134 Apply horizontal and vertical deblocking, deringing and automatic
15135 brightness/contrast:
15141 Apply default filters without brightness/contrast correction:
15147 Apply default filters and temporal denoiser:
15149 pp=default/tmpnoise|1|2|3
15153 Apply deblocking on luminance only, and switch vertical deblocking on or off
15154 automatically depending on available CPU time:
15161 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15162 similar to spp = 6 with 7 point DCT, where only the center sample is
15165 The filter accepts the following options:
15169 Force a constant quantization parameter. It accepts an integer in range
15170 0 to 63. If not set, the filter will use the QP from the video stream
15174 Set thresholding mode. Available modes are:
15178 Set hard thresholding.
15180 Set soft thresholding (better de-ringing effect, but likely blurrier).
15182 Set medium thresholding (good results, default).
15186 @section premultiply
15187 Apply alpha premultiply effect to input video stream using first plane
15188 of second stream as alpha.
15190 Both streams must have same dimensions and same pixel format.
15192 The filter accepts the following option:
15196 Set which planes will be processed, unprocessed planes will be copied.
15197 By default value 0xf, all planes will be processed.
15200 Do not require 2nd input for processing, instead use alpha plane from input stream.
15204 Apply prewitt operator to input video stream.
15206 The filter accepts the following option:
15210 Set which planes will be processed, unprocessed planes will be copied.
15211 By default value 0xf, all planes will be processed.
15214 Set value which will be multiplied with filtered result.
15217 Set value which will be added to filtered result.
15220 @section pseudocolor
15222 Alter frame colors in video with pseudocolors.
15224 This filter accepts the following options:
15228 set pixel first component expression
15231 set pixel second component expression
15234 set pixel third component expression
15237 set pixel fourth component expression, corresponds to the alpha component
15240 set component to use as base for altering colors
15243 Each of them specifies the expression to use for computing the lookup table for
15244 the corresponding pixel component values.
15246 The expressions can contain the following constants and functions:
15251 The input width and height.
15254 The input value for the pixel component.
15256 @item ymin, umin, vmin, amin
15257 The minimum allowed component value.
15259 @item ymax, umax, vmax, amax
15260 The maximum allowed component value.
15263 All expressions default to "val".
15265 @subsection Examples
15269 Change too high luma values to gradient:
15271 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'"
15277 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
15278 Ratio) between two input videos.
15280 This filter takes in input two input videos, the first input is
15281 considered the "main" source and is passed unchanged to the
15282 output. The second input is used as a "reference" video for computing
15285 Both video inputs must have the same resolution and pixel format for
15286 this filter to work correctly. Also it assumes that both inputs
15287 have the same number of frames, which are compared one by one.
15289 The obtained average PSNR is printed through the logging system.
15291 The filter stores the accumulated MSE (mean squared error) of each
15292 frame, and at the end of the processing it is averaged across all frames
15293 equally, and the following formula is applied to obtain the PSNR:
15296 PSNR = 10*log10(MAX^2/MSE)
15299 Where MAX is the average of the maximum values of each component of the
15302 The description of the accepted parameters follows.
15305 @item stats_file, f
15306 If specified the filter will use the named file to save the PSNR of
15307 each individual frame. When filename equals "-" the data is sent to
15310 @item stats_version
15311 Specifies which version of the stats file format to use. Details of
15312 each format are written below.
15313 Default value is 1.
15315 @item stats_add_max
15316 Determines whether the max value is output to the stats log.
15317 Default value is 0.
15318 Requires stats_version >= 2. If this is set and stats_version < 2,
15319 the filter will return an error.
15322 This filter also supports the @ref{framesync} options.
15324 The file printed if @var{stats_file} is selected, contains a sequence of
15325 key/value pairs of the form @var{key}:@var{value} for each compared
15328 If a @var{stats_version} greater than 1 is specified, a header line precedes
15329 the list of per-frame-pair stats, with key value pairs following the frame
15330 format with the following parameters:
15333 @item psnr_log_version
15334 The version of the log file format. Will match @var{stats_version}.
15337 A comma separated list of the per-frame-pair parameters included in
15341 A description of each shown per-frame-pair parameter follows:
15345 sequential number of the input frame, starting from 1
15348 Mean Square Error pixel-by-pixel average difference of the compared
15349 frames, averaged over all the image components.
15351 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
15352 Mean Square Error pixel-by-pixel average difference of the compared
15353 frames for the component specified by the suffix.
15355 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
15356 Peak Signal to Noise ratio of the compared frames for the component
15357 specified by the suffix.
15359 @item max_avg, max_y, max_u, max_v
15360 Maximum allowed value for each channel, and average over all
15364 @subsection Examples
15369 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
15370 [main][ref] psnr="stats_file=stats.log" [out]
15373 On this example the input file being processed is compared with the
15374 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
15375 is stored in @file{stats.log}.
15378 Another example with different containers:
15380 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 -
15387 Pulldown reversal (inverse telecine) filter, capable of handling mixed
15388 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
15391 The pullup filter is designed to take advantage of future context in making
15392 its decisions. This filter is stateless in the sense that it does not lock
15393 onto a pattern to follow, but it instead looks forward to the following
15394 fields in order to identify matches and rebuild progressive frames.
15396 To produce content with an even framerate, insert the fps filter after
15397 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
15398 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
15400 The filter accepts the following options:
15407 These options set the amount of "junk" to ignore at the left, right, top, and
15408 bottom of the image, respectively. Left and right are in units of 8 pixels,
15409 while top and bottom are in units of 2 lines.
15410 The default is 8 pixels on each side.
15413 Set the strict breaks. Setting this option to 1 will reduce the chances of
15414 filter generating an occasional mismatched frame, but it may also cause an
15415 excessive number of frames to be dropped during high motion sequences.
15416 Conversely, setting it to -1 will make filter match fields more easily.
15417 This may help processing of video where there is slight blurring between
15418 the fields, but may also cause there to be interlaced frames in the output.
15419 Default value is @code{0}.
15422 Set the metric plane to use. It accepts the following values:
15428 Use chroma blue plane.
15431 Use chroma red plane.
15434 This option may be set to use chroma plane instead of the default luma plane
15435 for doing filter's computations. This may improve accuracy on very clean
15436 source material, but more likely will decrease accuracy, especially if there
15437 is chroma noise (rainbow effect) or any grayscale video.
15438 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
15439 load and make pullup usable in realtime on slow machines.
15442 For best results (without duplicated frames in the output file) it is
15443 necessary to change the output frame rate. For example, to inverse
15444 telecine NTSC input:
15446 ffmpeg -i input -vf pullup -r 24000/1001 ...
15451 Change video quantization parameters (QP).
15453 The filter accepts the following option:
15457 Set expression for quantization parameter.
15460 The expression is evaluated through the eval API and can contain, among others,
15461 the following constants:
15465 1 if index is not 129, 0 otherwise.
15468 Sequential index starting from -129 to 128.
15471 @subsection Examples
15475 Some equation like:
15483 Flush video frames from internal cache of frames into a random order.
15484 No frame is discarded.
15485 Inspired by @ref{frei0r} nervous filter.
15489 Set size in number of frames of internal cache, in range from @code{2} to
15490 @code{512}. Default is @code{30}.
15493 Set seed for random number generator, must be an integer included between
15494 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
15495 less than @code{0}, the filter will try to use a good random seed on a
15499 @section readeia608
15501 Read closed captioning (EIA-608) information from the top lines of a video frame.
15503 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
15504 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
15505 with EIA-608 data (starting from 0). A description of each metadata value follows:
15508 @item lavfi.readeia608.X.cc
15509 The two bytes stored as EIA-608 data (printed in hexadecimal).
15511 @item lavfi.readeia608.X.line
15512 The number of the line on which the EIA-608 data was identified and read.
15515 This filter accepts the following options:
15519 Set the line to start scanning for EIA-608 data. Default is @code{0}.
15522 Set the line to end scanning for EIA-608 data. Default is @code{29}.
15525 Set the ratio of width reserved for sync code detection.
15526 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
15529 Enable checking the parity bit. In the event of a parity error, the filter will output
15530 @code{0x00} for that character. Default is false.
15533 Lowpass lines prior to further processing. Default is enabled.
15536 @subsection Examples
15540 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
15542 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
15548 Read vertical interval timecode (VITC) information from the top lines of a
15551 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
15552 timecode value, if a valid timecode has been detected. Further metadata key
15553 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
15554 timecode data has been found or not.
15556 This filter accepts the following options:
15560 Set the maximum number of lines to scan for VITC data. If the value is set to
15561 @code{-1} the full video frame is scanned. Default is @code{45}.
15564 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
15565 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
15568 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
15569 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
15572 @subsection Examples
15576 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
15577 draw @code{--:--:--:--} as a placeholder:
15579 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
15585 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
15587 Destination pixel at position (X, Y) will be picked from source (x, y) position
15588 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
15589 value for pixel will be used for destination pixel.
15591 Xmap and Ymap input video streams must be of same dimensions. Output video stream
15592 will have Xmap/Ymap video stream dimensions.
15593 Xmap and Ymap input video streams are 16bit depth, single channel.
15597 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
15598 Default is @code{color}.
15601 Specify the color of the unmapped pixels. For the syntax of this option,
15602 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15603 manual,ffmpeg-utils}. Default color is @code{black}.
15606 @section removegrain
15608 The removegrain filter is a spatial denoiser for progressive video.
15612 Set mode for the first plane.
15615 Set mode for the second plane.
15618 Set mode for the third plane.
15621 Set mode for the fourth plane.
15624 Range of mode is from 0 to 24. Description of each mode follows:
15628 Leave input plane unchanged. Default.
15631 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
15634 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
15637 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
15640 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
15641 This is equivalent to a median filter.
15644 Line-sensitive clipping giving the minimal change.
15647 Line-sensitive clipping, intermediate.
15650 Line-sensitive clipping, intermediate.
15653 Line-sensitive clipping, intermediate.
15656 Line-sensitive clipping on a line where the neighbours pixels are the closest.
15659 Replaces the target pixel with the closest neighbour.
15662 [1 2 1] horizontal and vertical kernel blur.
15668 Bob mode, interpolates top field from the line where the neighbours
15669 pixels are the closest.
15672 Bob mode, interpolates bottom field from the line where the neighbours
15673 pixels are the closest.
15676 Bob mode, interpolates top field. Same as 13 but with a more complicated
15677 interpolation formula.
15680 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
15681 interpolation formula.
15684 Clips the pixel with the minimum and maximum of respectively the maximum and
15685 minimum of each pair of opposite neighbour pixels.
15688 Line-sensitive clipping using opposite neighbours whose greatest distance from
15689 the current pixel is minimal.
15692 Replaces the pixel with the average of its 8 neighbours.
15695 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
15698 Clips pixels using the averages of opposite neighbour.
15701 Same as mode 21 but simpler and faster.
15704 Small edge and halo removal, but reputed useless.
15710 @section removelogo
15712 Suppress a TV station logo, using an image file to determine which
15713 pixels comprise the logo. It works by filling in the pixels that
15714 comprise the logo with neighboring pixels.
15716 The filter accepts the following options:
15720 Set the filter bitmap file, which can be any image format supported by
15721 libavformat. The width and height of the image file must match those of the
15722 video stream being processed.
15725 Pixels in the provided bitmap image with a value of zero are not
15726 considered part of the logo, non-zero pixels are considered part of
15727 the logo. If you use white (255) for the logo and black (0) for the
15728 rest, you will be safe. For making the filter bitmap, it is
15729 recommended to take a screen capture of a black frame with the logo
15730 visible, and then using a threshold filter followed by the erode
15731 filter once or twice.
15733 If needed, little splotches can be fixed manually. Remember that if
15734 logo pixels are not covered, the filter quality will be much
15735 reduced. Marking too many pixels as part of the logo does not hurt as
15736 much, but it will increase the amount of blurring needed to cover over
15737 the image and will destroy more information than necessary, and extra
15738 pixels will slow things down on a large logo.
15740 @section repeatfields
15742 This filter uses the repeat_field flag from the Video ES headers and hard repeats
15743 fields based on its value.
15747 Reverse a video clip.
15749 Warning: This filter requires memory to buffer the entire clip, so trimming
15752 @subsection Examples
15756 Take the first 5 seconds of a clip, and reverse it.
15763 Shift R/G/B/A pixels horizontally and/or vertically.
15765 The filter accepts the following options:
15768 Set amount to shift red horizontally.
15770 Set amount to shift red vertically.
15772 Set amount to shift green horizontally.
15774 Set amount to shift green vertically.
15776 Set amount to shift blue horizontally.
15778 Set amount to shift blue vertically.
15780 Set amount to shift alpha horizontally.
15782 Set amount to shift alpha vertically.
15784 Set edge mode, can be @var{smear}, default, or @var{warp}.
15787 @subsection Commands
15789 This filter supports the all above options as @ref{commands}.
15792 Apply roberts cross operator to input video stream.
15794 The filter accepts the following option:
15798 Set which planes will be processed, unprocessed planes will be copied.
15799 By default value 0xf, all planes will be processed.
15802 Set value which will be multiplied with filtered result.
15805 Set value which will be added to filtered result.
15810 Rotate video by an arbitrary angle expressed in radians.
15812 The filter accepts the following options:
15814 A description of the optional parameters follows.
15817 Set an expression for the angle by which to rotate the input video
15818 clockwise, expressed as a number of radians. A negative value will
15819 result in a counter-clockwise rotation. By default it is set to "0".
15821 This expression is evaluated for each frame.
15824 Set the output width expression, default value is "iw".
15825 This expression is evaluated just once during configuration.
15828 Set the output height expression, default value is "ih".
15829 This expression is evaluated just once during configuration.
15832 Enable bilinear interpolation if set to 1, a value of 0 disables
15833 it. Default value is 1.
15836 Set the color used to fill the output area not covered by the rotated
15837 image. For the general syntax of this option, check the
15838 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
15839 If the special value "none" is selected then no
15840 background is printed (useful for example if the background is never shown).
15842 Default value is "black".
15845 The expressions for the angle and the output size can contain the
15846 following constants and functions:
15850 sequential number of the input frame, starting from 0. It is always NAN
15851 before the first frame is filtered.
15854 time in seconds of the input frame, it is set to 0 when the filter is
15855 configured. It is always NAN before the first frame is filtered.
15859 horizontal and vertical chroma subsample values. For example for the
15860 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15864 the input video width and height
15868 the output width and height, that is the size of the padded area as
15869 specified by the @var{width} and @var{height} expressions
15873 the minimal width/height required for completely containing the input
15874 video rotated by @var{a} radians.
15876 These are only available when computing the @option{out_w} and
15877 @option{out_h} expressions.
15880 @subsection Examples
15884 Rotate the input by PI/6 radians clockwise:
15890 Rotate the input by PI/6 radians counter-clockwise:
15896 Rotate the input by 45 degrees clockwise:
15902 Apply a constant rotation with period T, starting from an angle of PI/3:
15904 rotate=PI/3+2*PI*t/T
15908 Make the input video rotation oscillating with a period of T
15909 seconds and an amplitude of A radians:
15911 rotate=A*sin(2*PI/T*t)
15915 Rotate the video, output size is chosen so that the whole rotating
15916 input video is always completely contained in the output:
15918 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
15922 Rotate the video, reduce the output size so that no background is ever
15925 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
15929 @subsection Commands
15931 The filter supports the following commands:
15935 Set the angle expression.
15936 The command accepts the same syntax of the corresponding option.
15938 If the specified expression is not valid, it is kept at its current
15944 Apply Shape Adaptive Blur.
15946 The filter accepts the following options:
15949 @item luma_radius, lr
15950 Set luma blur filter strength, must be a value in range 0.1-4.0, default
15951 value is 1.0. A greater value will result in a more blurred image, and
15952 in slower processing.
15954 @item luma_pre_filter_radius, lpfr
15955 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
15958 @item luma_strength, ls
15959 Set luma maximum difference between pixels to still be considered, must
15960 be a value in the 0.1-100.0 range, default value is 1.0.
15962 @item chroma_radius, cr
15963 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
15964 greater value will result in a more blurred image, and in slower
15967 @item chroma_pre_filter_radius, cpfr
15968 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
15970 @item chroma_strength, cs
15971 Set chroma maximum difference between pixels to still be considered,
15972 must be a value in the -0.9-100.0 range.
15975 Each chroma option value, if not explicitly specified, is set to the
15976 corresponding luma option value.
15981 Scale (resize) the input video, using the libswscale library.
15983 The scale filter forces the output display aspect ratio to be the same
15984 of the input, by changing the output sample aspect ratio.
15986 If the input image format is different from the format requested by
15987 the next filter, the scale filter will convert the input to the
15990 @subsection Options
15991 The filter accepts the following options, or any of the options
15992 supported by the libswscale scaler.
15994 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
15995 the complete list of scaler options.
16000 Set the output video dimension expression. Default value is the input
16003 If the @var{width} or @var{w} value is 0, the input width is used for
16004 the output. If the @var{height} or @var{h} value is 0, the input height
16005 is used for the output.
16007 If one and only one of the values is -n with n >= 1, the scale filter
16008 will use a value that maintains the aspect ratio of the input image,
16009 calculated from the other specified dimension. After that it will,
16010 however, make sure that the calculated dimension is divisible by n and
16011 adjust the value if necessary.
16013 If both values are -n with n >= 1, the behavior will be identical to
16014 both values being set to 0 as previously detailed.
16016 See below for the list of accepted constants for use in the dimension
16020 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16024 Only evaluate expressions once during the filter initialization or when a command is processed.
16027 Evaluate expressions for each incoming frame.
16031 Default value is @samp{init}.
16035 Set the interlacing mode. It accepts the following values:
16039 Force interlaced aware scaling.
16042 Do not apply interlaced scaling.
16045 Select interlaced aware scaling depending on whether the source frames
16046 are flagged as interlaced or not.
16049 Default value is @samp{0}.
16052 Set libswscale scaling flags. See
16053 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16054 complete list of values. If not explicitly specified the filter applies
16058 @item param0, param1
16059 Set libswscale input parameters for scaling algorithms that need them. See
16060 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16061 complete documentation. If not explicitly specified the filter applies
16067 Set the video size. For the syntax of this option, check the
16068 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16070 @item in_color_matrix
16071 @item out_color_matrix
16072 Set in/output YCbCr color space type.
16074 This allows the autodetected value to be overridden as well as allows forcing
16075 a specific value used for the output and encoder.
16077 If not specified, the color space type depends on the pixel format.
16083 Choose automatically.
16086 Format conforming to International Telecommunication Union (ITU)
16087 Recommendation BT.709.
16090 Set color space conforming to the United States Federal Communications
16091 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16096 Set color space conforming to:
16100 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16103 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16106 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16111 Set color space conforming to SMPTE ST 240:1999.
16114 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16119 Set in/output YCbCr sample range.
16121 This allows the autodetected value to be overridden as well as allows forcing
16122 a specific value used for the output and encoder. If not specified, the
16123 range depends on the pixel format. Possible values:
16127 Choose automatically.
16130 Set full range (0-255 in case of 8-bit luma).
16132 @item mpeg/limited/tv
16133 Set "MPEG" range (16-235 in case of 8-bit luma).
16136 @item force_original_aspect_ratio
16137 Enable decreasing or increasing output video width or height if necessary to
16138 keep the original aspect ratio. Possible values:
16142 Scale the video as specified and disable this feature.
16145 The output video dimensions will automatically be decreased if needed.
16148 The output video dimensions will automatically be increased if needed.
16152 One useful instance of this option is that when you know a specific device's
16153 maximum allowed resolution, you can use this to limit the output video to
16154 that, while retaining the aspect ratio. For example, device A allows
16155 1280x720 playback, and your video is 1920x800. Using this option (set it to
16156 decrease) and specifying 1280x720 to the command line makes the output
16159 Please note that this is a different thing than specifying -1 for @option{w}
16160 or @option{h}, you still need to specify the output resolution for this option
16163 @item force_divisible_by
16164 Ensures that both the output dimensions, width and height, are divisible by the
16165 given integer when used together with @option{force_original_aspect_ratio}. This
16166 works similar to using @code{-n} in the @option{w} and @option{h} options.
16168 This option respects the value set for @option{force_original_aspect_ratio},
16169 increasing or decreasing the resolution accordingly. The video's aspect ratio
16170 may be slightly modified.
16172 This option can be handy if you need to have a video fit within or exceed
16173 a defined resolution using @option{force_original_aspect_ratio} but also have
16174 encoder restrictions on width or height divisibility.
16178 The values of the @option{w} and @option{h} options are expressions
16179 containing the following constants:
16184 The input width and height
16188 These are the same as @var{in_w} and @var{in_h}.
16192 The output (scaled) width and height
16196 These are the same as @var{out_w} and @var{out_h}
16199 The same as @var{iw} / @var{ih}
16202 input sample aspect ratio
16205 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16209 horizontal and vertical input chroma subsample values. For example for the
16210 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16214 horizontal and vertical output chroma subsample values. For example for the
16215 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16218 The (sequential) number of the input frame, starting from 0.
16219 Only available with @code{eval=frame}.
16222 The presentation timestamp of the input frame, expressed as a number of
16223 seconds. Only available with @code{eval=frame}.
16226 The position (byte offset) of the frame in the input stream, or NaN if
16227 this information is unavailable and/or meaningless (for example in case of synthetic video).
16228 Only available with @code{eval=frame}.
16231 @subsection Examples
16235 Scale the input video to a size of 200x100
16240 This is equivalent to:
16251 Specify a size abbreviation for the output size:
16256 which can also be written as:
16262 Scale the input to 2x:
16264 scale=w=2*iw:h=2*ih
16268 The above is the same as:
16270 scale=2*in_w:2*in_h
16274 Scale the input to 2x with forced interlaced scaling:
16276 scale=2*iw:2*ih:interl=1
16280 Scale the input to half size:
16282 scale=w=iw/2:h=ih/2
16286 Increase the width, and set the height to the same size:
16292 Seek Greek harmony:
16299 Increase the height, and set the width to 3/2 of the height:
16301 scale=w=3/2*oh:h=3/5*ih
16305 Increase the size, making the size a multiple of the chroma
16308 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
16312 Increase the width to a maximum of 500 pixels,
16313 keeping the same aspect ratio as the input:
16315 scale=w='min(500\, iw*3/2):h=-1'
16319 Make pixels square by combining scale and setsar:
16321 scale='trunc(ih*dar):ih',setsar=1/1
16325 Make pixels square by combining scale and setsar,
16326 making sure the resulting resolution is even (required by some codecs):
16328 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
16332 @subsection Commands
16334 This filter supports the following commands:
16338 Set the output video dimension expression.
16339 The command accepts the same syntax of the corresponding option.
16341 If the specified expression is not valid, it is kept at its current
16347 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
16348 format conversion on CUDA video frames. Setting the output width and height
16349 works in the same way as for the @var{scale} filter.
16351 The following additional options are accepted:
16354 The pixel format of the output CUDA frames. If set to the string "same" (the
16355 default), the input format will be kept. Note that automatic format negotiation
16356 and conversion is not yet supported for hardware frames
16359 The interpolation algorithm used for resizing. One of the following:
16366 @item cubic2p_bspline
16367 2-parameter cubic (B=1, C=0)
16369 @item cubic2p_catmullrom
16370 2-parameter cubic (B=0, C=1/2)
16372 @item cubic2p_b05c03
16373 2-parameter cubic (B=1/2, C=3/10)
16381 @item force_original_aspect_ratio
16382 Enable decreasing or increasing output video width or height if necessary to
16383 keep the original aspect ratio. Possible values:
16387 Scale the video as specified and disable this feature.
16390 The output video dimensions will automatically be decreased if needed.
16393 The output video dimensions will automatically be increased if needed.
16397 One useful instance of this option is that when you know a specific device's
16398 maximum allowed resolution, you can use this to limit the output video to
16399 that, while retaining the aspect ratio. For example, device A allows
16400 1280x720 playback, and your video is 1920x800. Using this option (set it to
16401 decrease) and specifying 1280x720 to the command line makes the output
16404 Please note that this is a different thing than specifying -1 for @option{w}
16405 or @option{h}, you still need to specify the output resolution for this option
16408 @item force_divisible_by
16409 Ensures that both the output dimensions, width and height, are divisible by the
16410 given integer when used together with @option{force_original_aspect_ratio}. This
16411 works similar to using @code{-n} in the @option{w} and @option{h} options.
16413 This option respects the value set for @option{force_original_aspect_ratio},
16414 increasing or decreasing the resolution accordingly. The video's aspect ratio
16415 may be slightly modified.
16417 This option can be handy if you need to have a video fit within or exceed
16418 a defined resolution using @option{force_original_aspect_ratio} but also have
16419 encoder restrictions on width or height divisibility.
16425 Scale (resize) the input video, based on a reference video.
16427 See the scale filter for available options, scale2ref supports the same but
16428 uses the reference video instead of the main input as basis. scale2ref also
16429 supports the following additional constants for the @option{w} and
16430 @option{h} options:
16435 The main input video's width and height
16438 The same as @var{main_w} / @var{main_h}
16441 The main input video's sample aspect ratio
16443 @item main_dar, mdar
16444 The main input video's display aspect ratio. Calculated from
16445 @code{(main_w / main_h) * main_sar}.
16449 The main input video's horizontal and vertical chroma subsample values.
16450 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
16454 The (sequential) number of the main input frame, starting from 0.
16455 Only available with @code{eval=frame}.
16458 The presentation timestamp of the main input frame, expressed as a number of
16459 seconds. Only available with @code{eval=frame}.
16462 The position (byte offset) of the frame in the main input stream, or NaN if
16463 this information is unavailable and/or meaningless (for example in case of synthetic video).
16464 Only available with @code{eval=frame}.
16467 @subsection Examples
16471 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
16473 'scale2ref[b][a];[a][b]overlay'
16477 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
16479 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
16483 @subsection Commands
16485 This filter supports the following commands:
16489 Set the output video dimension expression.
16490 The command accepts the same syntax of the corresponding option.
16492 If the specified expression is not valid, it is kept at its current
16497 Scroll input video horizontally and/or vertically by constant speed.
16499 The filter accepts the following options:
16501 @item horizontal, h
16502 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
16503 Negative values changes scrolling direction.
16506 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
16507 Negative values changes scrolling direction.
16510 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
16513 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
16516 @subsection Commands
16518 This filter supports the following @ref{commands}:
16520 @item horizontal, h
16521 Set the horizontal scrolling speed.
16523 Set the vertical scrolling speed.
16526 @anchor{selectivecolor}
16527 @section selectivecolor
16529 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
16530 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
16531 by the "purity" of the color (that is, how saturated it already is).
16533 This filter is similar to the Adobe Photoshop Selective Color tool.
16535 The filter accepts the following options:
16538 @item correction_method
16539 Select color correction method.
16541 Available values are:
16544 Specified adjustments are applied "as-is" (added/subtracted to original pixel
16547 Specified adjustments are relative to the original component value.
16549 Default is @code{absolute}.
16551 Adjustments for red pixels (pixels where the red component is the maximum)
16553 Adjustments for yellow pixels (pixels where the blue component is the minimum)
16555 Adjustments for green pixels (pixels where the green component is the maximum)
16557 Adjustments for cyan pixels (pixels where the red component is the minimum)
16559 Adjustments for blue pixels (pixels where the blue component is the maximum)
16561 Adjustments for magenta pixels (pixels where the green component is the minimum)
16563 Adjustments for white pixels (pixels where all components are greater than 128)
16565 Adjustments for all pixels except pure black and pure white
16567 Adjustments for black pixels (pixels where all components are lesser than 128)
16569 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
16572 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
16573 4 space separated floating point adjustment values in the [-1,1] range,
16574 respectively to adjust the amount of cyan, magenta, yellow and black for the
16575 pixels of its range.
16577 @subsection Examples
16581 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
16582 increase magenta by 27% in blue areas:
16584 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
16588 Use a Photoshop selective color preset:
16590 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
16594 @anchor{separatefields}
16595 @section separatefields
16597 The @code{separatefields} takes a frame-based video input and splits
16598 each frame into its components fields, producing a new half height clip
16599 with twice the frame rate and twice the frame count.
16601 This filter use field-dominance information in frame to decide which
16602 of each pair of fields to place first in the output.
16603 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
16605 @section setdar, setsar
16607 The @code{setdar} filter sets the Display Aspect Ratio for the filter
16610 This is done by changing the specified Sample (aka Pixel) Aspect
16611 Ratio, according to the following equation:
16613 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
16616 Keep in mind that the @code{setdar} filter does not modify the pixel
16617 dimensions of the video frame. Also, the display aspect ratio set by
16618 this filter may be changed by later filters in the filterchain,
16619 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
16622 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
16623 the filter output video.
16625 Note that as a consequence of the application of this filter, the
16626 output display aspect ratio will change according to the equation
16629 Keep in mind that the sample aspect ratio set by the @code{setsar}
16630 filter may be changed by later filters in the filterchain, e.g. if
16631 another "setsar" or a "setdar" filter is applied.
16633 It accepts the following parameters:
16636 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
16637 Set the aspect ratio used by the filter.
16639 The parameter can be a floating point number string, an expression, or
16640 a string of the form @var{num}:@var{den}, where @var{num} and
16641 @var{den} are the numerator and denominator of the aspect ratio. If
16642 the parameter is not specified, it is assumed the value "0".
16643 In case the form "@var{num}:@var{den}" is used, the @code{:} character
16647 Set the maximum integer value to use for expressing numerator and
16648 denominator when reducing the expressed aspect ratio to a rational.
16649 Default value is @code{100}.
16653 The parameter @var{sar} is an expression containing
16654 the following constants:
16658 These are approximated values for the mathematical constants e
16659 (Euler's number), pi (Greek pi), and phi (the golden ratio).
16662 The input width and height.
16665 These are the same as @var{w} / @var{h}.
16668 The input sample aspect ratio.
16671 The input display aspect ratio. It is the same as
16672 (@var{w} / @var{h}) * @var{sar}.
16675 Horizontal and vertical chroma subsample values. For example, for the
16676 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16679 @subsection Examples
16684 To change the display aspect ratio to 16:9, specify one of the following:
16691 To change the sample aspect ratio to 10:11, specify:
16697 To set a display aspect ratio of 16:9, and specify a maximum integer value of
16698 1000 in the aspect ratio reduction, use the command:
16700 setdar=ratio=16/9:max=1000
16708 Force field for the output video frame.
16710 The @code{setfield} filter marks the interlace type field for the
16711 output frames. It does not change the input frame, but only sets the
16712 corresponding property, which affects how the frame is treated by
16713 following filters (e.g. @code{fieldorder} or @code{yadif}).
16715 The filter accepts the following options:
16720 Available values are:
16724 Keep the same field property.
16727 Mark the frame as bottom-field-first.
16730 Mark the frame as top-field-first.
16733 Mark the frame as progressive.
16740 Force frame parameter for the output video frame.
16742 The @code{setparams} filter marks interlace and color range for the
16743 output frames. It does not change the input frame, but only sets the
16744 corresponding property, which affects how the frame is treated by
16749 Available values are:
16753 Keep the same field property (default).
16756 Mark the frame as bottom-field-first.
16759 Mark the frame as top-field-first.
16762 Mark the frame as progressive.
16766 Available values are:
16770 Keep the same color range property (default).
16772 @item unspecified, unknown
16773 Mark the frame as unspecified color range.
16775 @item limited, tv, mpeg
16776 Mark the frame as limited range.
16778 @item full, pc, jpeg
16779 Mark the frame as full range.
16782 @item color_primaries
16783 Set the color primaries.
16784 Available values are:
16788 Keep the same color primaries property (default).
16805 Set the color transfer.
16806 Available values are:
16810 Keep the same color trc property (default).
16832 Set the colorspace.
16833 Available values are:
16837 Keep the same colorspace property (default).
16850 @item chroma-derived-nc
16851 @item chroma-derived-c
16858 Show a line containing various information for each input video frame.
16859 The input video is not modified.
16861 This filter supports the following options:
16865 Calculate checksums of each plane. By default enabled.
16868 The shown line contains a sequence of key/value pairs of the form
16869 @var{key}:@var{value}.
16871 The following values are shown in the output:
16875 The (sequential) number of the input frame, starting from 0.
16878 The Presentation TimeStamp of the input frame, expressed as a number of
16879 time base units. The time base unit depends on the filter input pad.
16882 The Presentation TimeStamp of the input frame, expressed as a number of
16886 The position of the frame in the input stream, or -1 if this information is
16887 unavailable and/or meaningless (for example in case of synthetic video).
16890 The pixel format name.
16893 The sample aspect ratio of the input frame, expressed in the form
16894 @var{num}/@var{den}.
16897 The size of the input frame. For the syntax of this option, check the
16898 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16901 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
16902 for bottom field first).
16905 This is 1 if the frame is a key frame, 0 otherwise.
16908 The picture type of the input frame ("I" for an I-frame, "P" for a
16909 P-frame, "B" for a B-frame, or "?" for an unknown type).
16910 Also refer to the documentation of the @code{AVPictureType} enum and of
16911 the @code{av_get_picture_type_char} function defined in
16912 @file{libavutil/avutil.h}.
16915 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
16917 @item plane_checksum
16918 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
16919 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
16922 The mean value of pixels in each plane of the input frame, expressed in the form
16923 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
16926 The standard deviation of pixel values in each plane of the input frame, expressed
16927 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
16931 @section showpalette
16933 Displays the 256 colors palette of each frame. This filter is only relevant for
16934 @var{pal8} pixel format frames.
16936 It accepts the following option:
16940 Set the size of the box used to represent one palette color entry. Default is
16941 @code{30} (for a @code{30x30} pixel box).
16944 @section shuffleframes
16946 Reorder and/or duplicate and/or drop video frames.
16948 It accepts the following parameters:
16952 Set the destination indexes of input frames.
16953 This is space or '|' separated list of indexes that maps input frames to output
16954 frames. Number of indexes also sets maximal value that each index may have.
16955 '-1' index have special meaning and that is to drop frame.
16958 The first frame has the index 0. The default is to keep the input unchanged.
16960 @subsection Examples
16964 Swap second and third frame of every three frames of the input:
16966 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
16970 Swap 10th and 1st frame of every ten frames of the input:
16972 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
16976 @section shuffleplanes
16978 Reorder and/or duplicate video planes.
16980 It accepts the following parameters:
16985 The index of the input plane to be used as the first output plane.
16988 The index of the input plane to be used as the second output plane.
16991 The index of the input plane to be used as the third output plane.
16994 The index of the input plane to be used as the fourth output plane.
16998 The first plane has the index 0. The default is to keep the input unchanged.
17000 @subsection Examples
17004 Swap the second and third planes of the input:
17006 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17010 @anchor{signalstats}
17011 @section signalstats
17012 Evaluate various visual metrics that assist in determining issues associated
17013 with the digitization of analog video media.
17015 By default the filter will log these metadata values:
17019 Display the minimal Y value contained within the input frame. Expressed in
17023 Display the Y value at the 10% percentile within the input frame. Expressed in
17027 Display the average Y value within the input frame. Expressed in range of
17031 Display the Y value at the 90% percentile within the input frame. Expressed in
17035 Display the maximum Y value contained within the input frame. Expressed in
17039 Display the minimal U value contained within the input frame. Expressed in
17043 Display the U value at the 10% percentile within the input frame. Expressed in
17047 Display the average U value within the input frame. Expressed in range of
17051 Display the U value at the 90% percentile within the input frame. Expressed in
17055 Display the maximum U value contained within the input frame. Expressed in
17059 Display the minimal V value contained within the input frame. Expressed in
17063 Display the V value at the 10% percentile within the input frame. Expressed in
17067 Display the average V value within the input frame. Expressed in range of
17071 Display the V value at the 90% percentile within the input frame. Expressed in
17075 Display the maximum V value contained within the input frame. Expressed in
17079 Display the minimal saturation value contained within the input frame.
17080 Expressed in range of [0-~181.02].
17083 Display the saturation value at the 10% percentile within the input frame.
17084 Expressed in range of [0-~181.02].
17087 Display the average saturation value within the input frame. Expressed in range
17091 Display the saturation value at the 90% percentile within the input frame.
17092 Expressed in range of [0-~181.02].
17095 Display the maximum saturation value contained within the input frame.
17096 Expressed in range of [0-~181.02].
17099 Display the median value for hue within the input frame. Expressed in range of
17103 Display the average value for hue within the input frame. Expressed in range of
17107 Display the average of sample value difference between all values of the Y
17108 plane in the current frame and corresponding values of the previous input frame.
17109 Expressed in range of [0-255].
17112 Display the average of sample value difference between all values of the U
17113 plane in the current frame and corresponding values of the previous input frame.
17114 Expressed in range of [0-255].
17117 Display the average of sample value difference between all values of the V
17118 plane in the current frame and corresponding values of the previous input frame.
17119 Expressed in range of [0-255].
17122 Display bit depth of Y plane in current frame.
17123 Expressed in range of [0-16].
17126 Display bit depth of U plane in current frame.
17127 Expressed in range of [0-16].
17130 Display bit depth of V plane in current frame.
17131 Expressed in range of [0-16].
17134 The filter accepts the following options:
17140 @option{stat} specify an additional form of image analysis.
17141 @option{out} output video with the specified type of pixel highlighted.
17143 Both options accept the following values:
17147 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17148 unlike the neighboring pixels of the same field. Examples of temporal outliers
17149 include the results of video dropouts, head clogs, or tape tracking issues.
17152 Identify @var{vertical line repetition}. Vertical line repetition includes
17153 similar rows of pixels within a frame. In born-digital video vertical line
17154 repetition is common, but this pattern is uncommon in video digitized from an
17155 analog source. When it occurs in video that results from the digitization of an
17156 analog source it can indicate concealment from a dropout compensator.
17159 Identify pixels that fall outside of legal broadcast range.
17163 Set the highlight color for the @option{out} option. The default color is
17167 @subsection Examples
17171 Output data of various video metrics:
17173 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17177 Output specific data about the minimum and maximum values of the Y plane per frame:
17179 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17183 Playback video while highlighting pixels that are outside of broadcast range in red.
17185 ffplay example.mov -vf signalstats="out=brng:color=red"
17189 Playback video with signalstats metadata drawn over the frame.
17191 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17194 The contents of signalstat_drawtext.txt used in the command are:
17197 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17198 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17199 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17200 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17208 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17209 input. In this case the matching between the inputs can be calculated additionally.
17210 The filter always passes through the first input. The signature of each stream can
17211 be written into a file.
17213 It accepts the following options:
17217 Enable or disable the matching process.
17219 Available values are:
17223 Disable the calculation of a matching (default).
17225 Calculate the matching for the whole video and output whether the whole video
17226 matches or only parts.
17228 Calculate only until a matching is found or the video ends. Should be faster in
17233 Set the number of inputs. The option value must be a non negative integer.
17234 Default value is 1.
17237 Set the path to which the output is written. If there is more than one input,
17238 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
17239 integer), that will be replaced with the input number. If no filename is
17240 specified, no output will be written. This is the default.
17243 Choose the output format.
17245 Available values are:
17249 Use the specified binary representation (default).
17251 Use the specified xml representation.
17255 Set threshold to detect one word as similar. The option value must be an integer
17256 greater than zero. The default value is 9000.
17259 Set threshold to detect all words as similar. The option value must be an integer
17260 greater than zero. The default value is 60000.
17263 Set threshold to detect frames as similar. The option value must be an integer
17264 greater than zero. The default value is 116.
17267 Set the minimum length of a sequence in frames to recognize it as matching
17268 sequence. The option value must be a non negative integer value.
17269 The default value is 0.
17272 Set the minimum relation, that matching frames to all frames must have.
17273 The option value must be a double value between 0 and 1. The default value is 0.5.
17276 @subsection Examples
17280 To calculate the signature of an input video and store it in signature.bin:
17282 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
17286 To detect whether two videos match and store the signatures in XML format in
17287 signature0.xml and signature1.xml:
17289 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 -
17297 Blur the input video without impacting the outlines.
17299 It accepts the following options:
17302 @item luma_radius, lr
17303 Set the luma radius. The option value must be a float number in
17304 the range [0.1,5.0] that specifies the variance of the gaussian filter
17305 used to blur the image (slower if larger). Default value is 1.0.
17307 @item luma_strength, ls
17308 Set the luma strength. The option value must be a float number
17309 in the range [-1.0,1.0] that configures the blurring. A value included
17310 in [0.0,1.0] will blur the image whereas a value included in
17311 [-1.0,0.0] will sharpen the image. Default value is 1.0.
17313 @item luma_threshold, lt
17314 Set the luma threshold used as a coefficient to determine
17315 whether a pixel should be blurred or not. The option value must be an
17316 integer in the range [-30,30]. A value of 0 will filter all the image,
17317 a value included in [0,30] will filter flat areas and a value included
17318 in [-30,0] will filter edges. Default value is 0.
17320 @item chroma_radius, cr
17321 Set the chroma radius. The option value must be a float number in
17322 the range [0.1,5.0] that specifies the variance of the gaussian filter
17323 used to blur the image (slower if larger). Default value is @option{luma_radius}.
17325 @item chroma_strength, cs
17326 Set the chroma strength. The option value must be a float number
17327 in the range [-1.0,1.0] that configures the blurring. A value included
17328 in [0.0,1.0] will blur the image whereas a value included in
17329 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
17331 @item chroma_threshold, ct
17332 Set the chroma threshold used as a coefficient to determine
17333 whether a pixel should be blurred or not. The option value must be an
17334 integer in the range [-30,30]. A value of 0 will filter all the image,
17335 a value included in [0,30] will filter flat areas and a value included
17336 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
17339 If a chroma option is not explicitly set, the corresponding luma value
17343 Apply sobel operator to input video stream.
17345 The filter accepts the following option:
17349 Set which planes will be processed, unprocessed planes will be copied.
17350 By default value 0xf, all planes will be processed.
17353 Set value which will be multiplied with filtered result.
17356 Set value which will be added to filtered result.
17362 Apply a simple postprocessing filter that compresses and decompresses the image
17363 at several (or - in the case of @option{quality} level @code{6} - all) shifts
17364 and average the results.
17366 The filter accepts the following options:
17370 Set quality. This option defines the number of levels for averaging. It accepts
17371 an integer in the range 0-6. If set to @code{0}, the filter will have no
17372 effect. A value of @code{6} means the higher quality. For each increment of
17373 that value the speed drops by a factor of approximately 2. Default value is
17377 Force a constant quantization parameter. If not set, the filter will use the QP
17378 from the video stream (if available).
17381 Set thresholding mode. Available modes are:
17385 Set hard thresholding (default).
17387 Set soft thresholding (better de-ringing effect, but likely blurrier).
17390 @item use_bframe_qp
17391 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
17392 option may cause flicker since the B-Frames have often larger QP. Default is
17393 @code{0} (not enabled).
17396 @subsection Commands
17398 This filter supports the following commands:
17400 @item quality, level
17401 Set quality level. The value @code{max} can be used to set the maximum level,
17402 currently @code{6}.
17408 Scale the input by applying one of the super-resolution methods based on
17409 convolutional neural networks. Supported models:
17413 Super-Resolution Convolutional Neural Network model (SRCNN).
17414 See @url{https://arxiv.org/abs/1501.00092}.
17417 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
17418 See @url{https://arxiv.org/abs/1609.05158}.
17421 Training scripts as well as scripts for model file (.pb) saving can be found at
17422 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
17423 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
17425 Native model files (.model) can be generated from TensorFlow model
17426 files (.pb) by using tools/python/convert.py
17428 The filter accepts the following options:
17432 Specify which DNN backend to use for model loading and execution. This option accepts
17433 the following values:
17437 Native implementation of DNN loading and execution.
17440 TensorFlow backend. To enable this backend you
17441 need to install the TensorFlow for C library (see
17442 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
17443 @code{--enable-libtensorflow}
17446 Default value is @samp{native}.
17449 Set path to model file specifying network architecture and its parameters.
17450 Note that different backends use different file formats. TensorFlow backend
17451 can load files for both formats, while native backend can load files for only
17455 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
17456 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
17457 input upscaled using bicubic upscaling with proper scale factor.
17460 This feature can also be finished with @ref{dnn_processing} filter.
17464 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
17466 This filter takes in input two input videos, the first input is
17467 considered the "main" source and is passed unchanged to the
17468 output. The second input is used as a "reference" video for computing
17471 Both video inputs must have the same resolution and pixel format for
17472 this filter to work correctly. Also it assumes that both inputs
17473 have the same number of frames, which are compared one by one.
17475 The filter stores the calculated SSIM of each frame.
17477 The description of the accepted parameters follows.
17480 @item stats_file, f
17481 If specified the filter will use the named file to save the SSIM of
17482 each individual frame. When filename equals "-" the data is sent to
17486 The file printed if @var{stats_file} is selected, contains a sequence of
17487 key/value pairs of the form @var{key}:@var{value} for each compared
17490 A description of each shown parameter follows:
17494 sequential number of the input frame, starting from 1
17496 @item Y, U, V, R, G, B
17497 SSIM of the compared frames for the component specified by the suffix.
17500 SSIM of the compared frames for the whole frame.
17503 Same as above but in dB representation.
17506 This filter also supports the @ref{framesync} options.
17508 @subsection Examples
17513 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
17514 [main][ref] ssim="stats_file=stats.log" [out]
17517 On this example the input file being processed is compared with the
17518 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
17519 is stored in @file{stats.log}.
17522 Another example with both psnr and ssim at same time:
17524 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
17528 Another example with different containers:
17530 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 -
17536 Convert between different stereoscopic image formats.
17538 The filters accept the following options:
17542 Set stereoscopic image format of input.
17544 Available values for input image formats are:
17547 side by side parallel (left eye left, right eye right)
17550 side by side crosseye (right eye left, left eye right)
17553 side by side parallel with half width resolution
17554 (left eye left, right eye right)
17557 side by side crosseye with half width resolution
17558 (right eye left, left eye right)
17562 above-below (left eye above, right eye below)
17566 above-below (right eye above, left eye below)
17570 above-below with half height resolution
17571 (left eye above, right eye below)
17575 above-below with half height resolution
17576 (right eye above, left eye below)
17579 alternating frames (left eye first, right eye second)
17582 alternating frames (right eye first, left eye second)
17585 interleaved rows (left eye has top row, right eye starts on next row)
17588 interleaved rows (right eye has top row, left eye starts on next row)
17591 interleaved columns, left eye first
17594 interleaved columns, right eye first
17596 Default value is @samp{sbsl}.
17600 Set stereoscopic image format of output.
17604 side by side parallel (left eye left, right eye right)
17607 side by side crosseye (right eye left, left eye right)
17610 side by side parallel with half width resolution
17611 (left eye left, right eye right)
17614 side by side crosseye with half width resolution
17615 (right eye left, left eye right)
17619 above-below (left eye above, right eye below)
17623 above-below (right eye above, left eye below)
17627 above-below with half height resolution
17628 (left eye above, right eye below)
17632 above-below with half height resolution
17633 (right eye above, left eye below)
17636 alternating frames (left eye first, right eye second)
17639 alternating frames (right eye first, left eye second)
17642 interleaved rows (left eye has top row, right eye starts on next row)
17645 interleaved rows (right eye has top row, left eye starts on next row)
17648 anaglyph red/blue gray
17649 (red filter on left eye, blue filter on right eye)
17652 anaglyph red/green gray
17653 (red filter on left eye, green filter on right eye)
17656 anaglyph red/cyan gray
17657 (red filter on left eye, cyan filter on right eye)
17660 anaglyph red/cyan half colored
17661 (red filter on left eye, cyan filter on right eye)
17664 anaglyph red/cyan color
17665 (red filter on left eye, cyan filter on right eye)
17668 anaglyph red/cyan color optimized with the least squares projection of dubois
17669 (red filter on left eye, cyan filter on right eye)
17672 anaglyph green/magenta gray
17673 (green filter on left eye, magenta filter on right eye)
17676 anaglyph green/magenta half colored
17677 (green filter on left eye, magenta filter on right eye)
17680 anaglyph green/magenta colored
17681 (green filter on left eye, magenta filter on right eye)
17684 anaglyph green/magenta color optimized with the least squares projection of dubois
17685 (green filter on left eye, magenta filter on right eye)
17688 anaglyph yellow/blue gray
17689 (yellow filter on left eye, blue filter on right eye)
17692 anaglyph yellow/blue half colored
17693 (yellow filter on left eye, blue filter on right eye)
17696 anaglyph yellow/blue colored
17697 (yellow filter on left eye, blue filter on right eye)
17700 anaglyph yellow/blue color optimized with the least squares projection of dubois
17701 (yellow filter on left eye, blue filter on right eye)
17704 mono output (left eye only)
17707 mono output (right eye only)
17710 checkerboard, left eye first
17713 checkerboard, right eye first
17716 interleaved columns, left eye first
17719 interleaved columns, right eye first
17725 Default value is @samp{arcd}.
17728 @subsection Examples
17732 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
17738 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
17744 @section streamselect, astreamselect
17745 Select video or audio streams.
17747 The filter accepts the following options:
17751 Set number of inputs. Default is 2.
17754 Set input indexes to remap to outputs.
17757 @subsection Commands
17759 The @code{streamselect} and @code{astreamselect} filter supports the following
17764 Set input indexes to remap to outputs.
17767 @subsection Examples
17771 Select first 5 seconds 1st stream and rest of time 2nd stream:
17773 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
17777 Same as above, but for audio:
17779 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
17786 Draw subtitles on top of input video using the libass library.
17788 To enable compilation of this filter you need to configure FFmpeg with
17789 @code{--enable-libass}. This filter also requires a build with libavcodec and
17790 libavformat to convert the passed subtitles file to ASS (Advanced Substation
17791 Alpha) subtitles format.
17793 The filter accepts the following options:
17797 Set the filename of the subtitle file to read. It must be specified.
17799 @item original_size
17800 Specify the size of the original video, the video for which the ASS file
17801 was composed. For the syntax of this option, check the
17802 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17803 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
17804 correctly scale the fonts if the aspect ratio has been changed.
17807 Set a directory path containing fonts that can be used by the filter.
17808 These fonts will be used in addition to whatever the font provider uses.
17811 Process alpha channel, by default alpha channel is untouched.
17814 Set subtitles input character encoding. @code{subtitles} filter only. Only
17815 useful if not UTF-8.
17817 @item stream_index, si
17818 Set subtitles stream index. @code{subtitles} filter only.
17821 Override default style or script info parameters of the subtitles. It accepts a
17822 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
17825 If the first key is not specified, it is assumed that the first value
17826 specifies the @option{filename}.
17828 For example, to render the file @file{sub.srt} on top of the input
17829 video, use the command:
17834 which is equivalent to:
17836 subtitles=filename=sub.srt
17839 To render the default subtitles stream from file @file{video.mkv}, use:
17841 subtitles=video.mkv
17844 To render the second subtitles stream from that file, use:
17846 subtitles=video.mkv:si=1
17849 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
17850 @code{DejaVu Serif}, use:
17852 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
17855 @section super2xsai
17857 Scale the input by 2x and smooth using the Super2xSaI (Scale and
17858 Interpolate) pixel art scaling algorithm.
17860 Useful for enlarging pixel art images without reducing sharpness.
17864 Swap two rectangular objects in video.
17866 This filter accepts the following options:
17876 Set 1st rect x coordinate.
17879 Set 1st rect y coordinate.
17882 Set 2nd rect x coordinate.
17885 Set 2nd rect y coordinate.
17887 All expressions are evaluated once for each frame.
17890 The all options are expressions containing the following constants:
17895 The input width and height.
17898 same as @var{w} / @var{h}
17901 input sample aspect ratio
17904 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
17907 The number of the input frame, starting from 0.
17910 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
17913 the position in the file of the input frame, NAN if unknown
17920 Blend successive video frames.
17926 Apply telecine process to the video.
17928 This filter accepts the following options:
17937 The default value is @code{top}.
17941 A string of numbers representing the pulldown pattern you wish to apply.
17942 The default value is @code{23}.
17946 Some typical patterns:
17951 24p: 2332 (preferred)
17958 24p: 222222222223 ("Euro pulldown")
17963 @section thistogram
17965 Compute and draw a color distribution histogram for the input video across time.
17967 Unlike @ref{histogram} video filter which only shows histogram of single input frame
17968 at certain time, this filter shows also past histograms of number of frames defined
17969 by @code{width} option.
17971 The computed histogram is a representation of the color component
17972 distribution in an image.
17974 The filter accepts the following options:
17978 Set width of single color component output. Default value is @code{0}.
17979 Value of @code{0} means width will be picked from input video.
17980 This also set number of passed histograms to keep.
17981 Allowed range is [0, 8192].
17983 @item display_mode, d
17985 It accepts the following values:
17988 Per color component graphs are placed below each other.
17991 Per color component graphs are placed side by side.
17994 Presents information identical to that in the @code{parade}, except
17995 that the graphs representing color components are superimposed directly
17998 Default is @code{stack}.
18000 @item levels_mode, m
18001 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18002 Default is @code{linear}.
18004 @item components, c
18005 Set what color components to display.
18006 Default is @code{7}.
18009 Set background opacity. Default is @code{0.9}.
18012 Show envelope. Default is disabled.
18015 Set envelope color. Default is @code{gold}.
18020 Apply threshold effect to video stream.
18022 This filter needs four video streams to perform thresholding.
18023 First stream is stream we are filtering.
18024 Second stream is holding threshold values, third stream is holding min values,
18025 and last, fourth stream is holding max values.
18027 The filter accepts the following option:
18031 Set which planes will be processed, unprocessed planes will be copied.
18032 By default value 0xf, all planes will be processed.
18035 For example if first stream pixel's component value is less then threshold value
18036 of pixel component from 2nd threshold stream, third stream value will picked,
18037 otherwise fourth stream pixel component value will be picked.
18039 Using color source filter one can perform various types of thresholding:
18041 @subsection Examples
18045 Binary threshold, using gray color as threshold:
18047 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18051 Inverted binary threshold, using gray color as threshold:
18053 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18057 Truncate binary threshold, using gray color as threshold:
18059 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18063 Threshold to zero, using gray color as threshold:
18065 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18069 Inverted threshold to zero, using gray color as threshold:
18071 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18076 Select the most representative frame in a given sequence of consecutive frames.
18078 The filter accepts the following options:
18082 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18083 will pick one of them, and then handle the next batch of @var{n} frames until
18084 the end. Default is @code{100}.
18087 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18088 value will result in a higher memory usage, so a high value is not recommended.
18090 @subsection Examples
18094 Extract one picture each 50 frames:
18100 Complete example of a thumbnail creation with @command{ffmpeg}:
18102 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18108 Tile several successive frames together.
18110 The filter accepts the following options:
18115 Set the grid size (i.e. the number of lines and columns). For the syntax of
18116 this option, check the
18117 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18120 Set the maximum number of frames to render in the given area. It must be less
18121 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18122 the area will be used.
18125 Set the outer border margin in pixels.
18128 Set the inner border thickness (i.e. the number of pixels between frames). For
18129 more advanced padding options (such as having different values for the edges),
18130 refer to the pad video filter.
18133 Specify the color of the unused area. For the syntax of this option, check the
18134 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18135 The default value of @var{color} is "black".
18138 Set the number of frames to overlap when tiling several successive frames together.
18139 The value must be between @code{0} and @var{nb_frames - 1}.
18142 Set the number of frames to initially be empty before displaying first output frame.
18143 This controls how soon will one get first output frame.
18144 The value must be between @code{0} and @var{nb_frames - 1}.
18147 @subsection Examples
18151 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18153 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18155 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18156 duplicating each output frame to accommodate the originally detected frame
18160 Display @code{5} pictures in an area of @code{3x2} frames,
18161 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18162 mixed flat and named options:
18164 tile=3x2:nb_frames=5:padding=7:margin=2
18168 @section tinterlace
18170 Perform various types of temporal field interlacing.
18172 Frames are counted starting from 1, so the first input frame is
18175 The filter accepts the following options:
18180 Specify the mode of the interlacing. This option can also be specified
18181 as a value alone. See below for a list of values for this option.
18183 Available values are:
18187 Move odd frames into the upper field, even into the lower field,
18188 generating a double height frame at half frame rate.
18192 Frame 1 Frame 2 Frame 3 Frame 4
18194 11111 22222 33333 44444
18195 11111 22222 33333 44444
18196 11111 22222 33333 44444
18197 11111 22222 33333 44444
18211 Only output odd frames, even frames are dropped, generating a frame with
18212 unchanged height at half frame rate.
18217 Frame 1 Frame 2 Frame 3 Frame 4
18219 11111 22222 33333 44444
18220 11111 22222 33333 44444
18221 11111 22222 33333 44444
18222 11111 22222 33333 44444
18232 Only output even frames, odd frames are dropped, generating a frame with
18233 unchanged height at half frame rate.
18238 Frame 1 Frame 2 Frame 3 Frame 4
18240 11111 22222 33333 44444
18241 11111 22222 33333 44444
18242 11111 22222 33333 44444
18243 11111 22222 33333 44444
18253 Expand each frame to full height, but pad alternate lines with black,
18254 generating a frame with double height at the same input frame rate.
18259 Frame 1 Frame 2 Frame 3 Frame 4
18261 11111 22222 33333 44444
18262 11111 22222 33333 44444
18263 11111 22222 33333 44444
18264 11111 22222 33333 44444
18267 11111 ..... 33333 .....
18268 ..... 22222 ..... 44444
18269 11111 ..... 33333 .....
18270 ..... 22222 ..... 44444
18271 11111 ..... 33333 .....
18272 ..... 22222 ..... 44444
18273 11111 ..... 33333 .....
18274 ..... 22222 ..... 44444
18278 @item interleave_top, 4
18279 Interleave the upper field from odd frames with the lower field from
18280 even frames, generating a frame with unchanged height at half frame rate.
18285 Frame 1 Frame 2 Frame 3 Frame 4
18287 11111<- 22222 33333<- 44444
18288 11111 22222<- 33333 44444<-
18289 11111<- 22222 33333<- 44444
18290 11111 22222<- 33333 44444<-
18300 @item interleave_bottom, 5
18301 Interleave the lower field from odd frames with the upper field from
18302 even frames, generating a frame with unchanged height at half frame rate.
18307 Frame 1 Frame 2 Frame 3 Frame 4
18309 11111 22222<- 33333 44444<-
18310 11111<- 22222 33333<- 44444
18311 11111 22222<- 33333 44444<-
18312 11111<- 22222 33333<- 44444
18322 @item interlacex2, 6
18323 Double frame rate with unchanged height. Frames are inserted each
18324 containing the second temporal field from the previous input frame and
18325 the first temporal field from the next input frame. This mode relies on
18326 the top_field_first flag. Useful for interlaced video displays with no
18327 field synchronisation.
18332 Frame 1 Frame 2 Frame 3 Frame 4
18334 11111 22222 33333 44444
18335 11111 22222 33333 44444
18336 11111 22222 33333 44444
18337 11111 22222 33333 44444
18340 11111 22222 22222 33333 33333 44444 44444
18341 11111 11111 22222 22222 33333 33333 44444
18342 11111 22222 22222 33333 33333 44444 44444
18343 11111 11111 22222 22222 33333 33333 44444
18348 Move odd frames into the upper field, even into the lower field,
18349 generating a double height frame at same frame rate.
18354 Frame 1 Frame 2 Frame 3 Frame 4
18356 11111 22222 33333 44444
18357 11111 22222 33333 44444
18358 11111 22222 33333 44444
18359 11111 22222 33333 44444
18362 11111 33333 33333 55555
18363 22222 22222 44444 44444
18364 11111 33333 33333 55555
18365 22222 22222 44444 44444
18366 11111 33333 33333 55555
18367 22222 22222 44444 44444
18368 11111 33333 33333 55555
18369 22222 22222 44444 44444
18374 Numeric values are deprecated but are accepted for backward
18375 compatibility reasons.
18377 Default mode is @code{merge}.
18380 Specify flags influencing the filter process.
18382 Available value for @var{flags} is:
18385 @item low_pass_filter, vlpf
18386 Enable linear vertical low-pass filtering in the filter.
18387 Vertical low-pass filtering is required when creating an interlaced
18388 destination from a progressive source which contains high-frequency
18389 vertical detail. Filtering will reduce interlace 'twitter' and Moire
18392 @item complex_filter, cvlpf
18393 Enable complex vertical low-pass filtering.
18394 This will slightly less reduce interlace 'twitter' and Moire
18395 patterning but better retain detail and subjective sharpness impression.
18398 Bypass already interlaced frames, only adjust the frame rate.
18401 Vertical low-pass filtering and bypassing already interlaced frames can only be
18402 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
18407 Pick median pixels from several successive input video frames.
18409 The filter accepts the following options:
18413 Set radius of median filter.
18414 Default is 1. Allowed range is from 1 to 127.
18417 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
18420 Set median percentile. Default value is @code{0.5}.
18421 Default value of @code{0.5} will pick always median values, while @code{0} will pick
18422 minimum values, and @code{1} maximum values.
18427 Mix successive video frames.
18429 A description of the accepted options follows.
18433 The number of successive frames to mix. If unspecified, it defaults to 3.
18436 Specify weight of each input video frame.
18437 Each weight is separated by space. If number of weights is smaller than
18438 number of @var{frames} last specified weight will be used for all remaining
18442 Specify scale, if it is set it will be multiplied with sum
18443 of each weight multiplied with pixel values to give final destination
18444 pixel value. By default @var{scale} is auto scaled to sum of weights.
18447 @subsection Examples
18451 Average 7 successive frames:
18453 tmix=frames=7:weights="1 1 1 1 1 1 1"
18457 Apply simple temporal convolution:
18459 tmix=frames=3:weights="-1 3 -1"
18463 Similar as above but only showing temporal differences:
18465 tmix=frames=3:weights="-1 2 -1":scale=1
18471 Tone map colors from different dynamic ranges.
18473 This filter expects data in single precision floating point, as it needs to
18474 operate on (and can output) out-of-range values. Another filter, such as
18475 @ref{zscale}, is needed to convert the resulting frame to a usable format.
18477 The tonemapping algorithms implemented only work on linear light, so input
18478 data should be linearized beforehand (and possibly correctly tagged).
18481 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
18484 @subsection Options
18485 The filter accepts the following options.
18489 Set the tone map algorithm to use.
18491 Possible values are:
18494 Do not apply any tone map, only desaturate overbright pixels.
18497 Hard-clip any out-of-range values. Use it for perfect color accuracy for
18498 in-range values, while distorting out-of-range values.
18501 Stretch the entire reference gamut to a linear multiple of the display.
18504 Fit a logarithmic transfer between the tone curves.
18507 Preserve overall image brightness with a simple curve, using nonlinear
18508 contrast, which results in flattening details and degrading color accuracy.
18511 Preserve both dark and bright details better than @var{reinhard}, at the cost
18512 of slightly darkening everything. Use it when detail preservation is more
18513 important than color and brightness accuracy.
18516 Smoothly map out-of-range values, while retaining contrast and colors for
18517 in-range material as much as possible. Use it when color accuracy is more
18518 important than detail preservation.
18524 Tune the tone mapping algorithm.
18526 This affects the following algorithms:
18532 Specifies the scale factor to use while stretching.
18536 Specifies the exponent of the function.
18540 Specify an extra linear coefficient to multiply into the signal before clipping.
18544 Specify the local contrast coefficient at the display peak.
18545 Default to 0.5, which means that in-gamut values will be about half as bright
18552 Specify the transition point from linear to mobius transform. Every value
18553 below this point is guaranteed to be mapped 1:1. The higher the value, the
18554 more accurate the result will be, at the cost of losing bright details.
18555 Default to 0.3, which due to the steep initial slope still preserves in-range
18556 colors fairly accurately.
18560 Apply desaturation for highlights that exceed this level of brightness. The
18561 higher the parameter, the more color information will be preserved. This
18562 setting helps prevent unnaturally blown-out colors for super-highlights, by
18563 (smoothly) turning into white instead. This makes images feel more natural,
18564 at the cost of reducing information about out-of-range colors.
18566 The default of 2.0 is somewhat conservative and will mostly just apply to
18567 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
18569 This option works only if the input frame has a supported color tag.
18572 Override signal/nominal/reference peak with this value. Useful when the
18573 embedded peak information in display metadata is not reliable or when tone
18574 mapping from a lower range to a higher range.
18579 Temporarily pad video frames.
18581 The filter accepts the following options:
18585 Specify number of delay frames before input video stream. Default is 0.
18588 Specify number of padding frames after input video stream.
18589 Set to -1 to pad indefinitely. Default is 0.
18592 Set kind of frames added to beginning of stream.
18593 Can be either @var{add} or @var{clone}.
18594 With @var{add} frames of solid-color are added.
18595 With @var{clone} frames are clones of first frame.
18596 Default is @var{add}.
18599 Set kind of frames added to end of stream.
18600 Can be either @var{add} or @var{clone}.
18601 With @var{add} frames of solid-color are added.
18602 With @var{clone} frames are clones of last frame.
18603 Default is @var{add}.
18605 @item start_duration, stop_duration
18606 Specify the duration of the start/stop delay. See
18607 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18608 for the accepted syntax.
18609 These options override @var{start} and @var{stop}. Default is 0.
18612 Specify the color of the padded area. For the syntax of this option,
18613 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
18614 manual,ffmpeg-utils}.
18616 The default value of @var{color} is "black".
18622 Transpose rows with columns in the input video and optionally flip it.
18624 It accepts the following parameters:
18629 Specify the transposition direction.
18631 Can assume the following values:
18633 @item 0, 4, cclock_flip
18634 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
18642 Rotate by 90 degrees clockwise, that is:
18650 Rotate by 90 degrees counterclockwise, that is:
18657 @item 3, 7, clock_flip
18658 Rotate by 90 degrees clockwise and vertically flip, that is:
18666 For values between 4-7, the transposition is only done if the input
18667 video geometry is portrait and not landscape. These values are
18668 deprecated, the @code{passthrough} option should be used instead.
18670 Numerical values are deprecated, and should be dropped in favor of
18671 symbolic constants.
18674 Do not apply the transposition if the input geometry matches the one
18675 specified by the specified value. It accepts the following values:
18678 Always apply transposition.
18680 Preserve portrait geometry (when @var{height} >= @var{width}).
18682 Preserve landscape geometry (when @var{width} >= @var{height}).
18685 Default value is @code{none}.
18688 For example to rotate by 90 degrees clockwise and preserve portrait
18691 transpose=dir=1:passthrough=portrait
18694 The command above can also be specified as:
18696 transpose=1:portrait
18699 @section transpose_npp
18701 Transpose rows with columns in the input video and optionally flip it.
18702 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
18704 It accepts the following parameters:
18709 Specify the transposition direction.
18711 Can assume the following values:
18714 Rotate by 90 degrees counterclockwise and vertically flip. (default)
18717 Rotate by 90 degrees clockwise.
18720 Rotate by 90 degrees counterclockwise.
18723 Rotate by 90 degrees clockwise and vertically flip.
18727 Do not apply the transposition if the input geometry matches the one
18728 specified by the specified value. It accepts the following values:
18731 Always apply transposition. (default)
18733 Preserve portrait geometry (when @var{height} >= @var{width}).
18735 Preserve landscape geometry (when @var{width} >= @var{height}).
18741 Trim the input so that the output contains one continuous subpart of the input.
18743 It accepts the following parameters:
18746 Specify the time of the start of the kept section, i.e. the frame with the
18747 timestamp @var{start} will be the first frame in the output.
18750 Specify the time of the first frame that will be dropped, i.e. the frame
18751 immediately preceding the one with the timestamp @var{end} will be the last
18752 frame in the output.
18755 This is the same as @var{start}, except this option sets the start timestamp
18756 in timebase units instead of seconds.
18759 This is the same as @var{end}, except this option sets the end timestamp
18760 in timebase units instead of seconds.
18763 The maximum duration of the output in seconds.
18766 The number of the first frame that should be passed to the output.
18769 The number of the first frame that should be dropped.
18772 @option{start}, @option{end}, and @option{duration} are expressed as time
18773 duration specifications; see
18774 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18775 for the accepted syntax.
18777 Note that the first two sets of the start/end options and the @option{duration}
18778 option look at the frame timestamp, while the _frame variants simply count the
18779 frames that pass through the filter. Also note that this filter does not modify
18780 the timestamps. If you wish for the output timestamps to start at zero, insert a
18781 setpts filter after the trim filter.
18783 If multiple start or end options are set, this filter tries to be greedy and
18784 keep all the frames that match at least one of the specified constraints. To keep
18785 only the part that matches all the constraints at once, chain multiple trim
18788 The defaults are such that all the input is kept. So it is possible to set e.g.
18789 just the end values to keep everything before the specified time.
18794 Drop everything except the second minute of input:
18796 ffmpeg -i INPUT -vf trim=60:120
18800 Keep only the first second:
18802 ffmpeg -i INPUT -vf trim=duration=1
18807 @section unpremultiply
18808 Apply alpha unpremultiply effect to input video stream using first plane
18809 of second stream as alpha.
18811 Both streams must have same dimensions and same pixel format.
18813 The filter accepts the following option:
18817 Set which planes will be processed, unprocessed planes will be copied.
18818 By default value 0xf, all planes will be processed.
18820 If the format has 1 or 2 components, then luma is bit 0.
18821 If the format has 3 or 4 components:
18822 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
18823 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
18824 If present, the alpha channel is always the last bit.
18827 Do not require 2nd input for processing, instead use alpha plane from input stream.
18833 Sharpen or blur the input video.
18835 It accepts the following parameters:
18838 @item luma_msize_x, lx
18839 Set the luma matrix horizontal size. It must be an odd integer between
18840 3 and 23. The default value is 5.
18842 @item luma_msize_y, ly
18843 Set the luma matrix vertical size. It must be an odd integer between 3
18844 and 23. The default value is 5.
18846 @item luma_amount, la
18847 Set the luma effect strength. It must be a floating point number, reasonable
18848 values lay between -1.5 and 1.5.
18850 Negative values will blur the input video, while positive values will
18851 sharpen it, a value of zero will disable the effect.
18853 Default value is 1.0.
18855 @item chroma_msize_x, cx
18856 Set the chroma matrix horizontal size. It must be an odd integer
18857 between 3 and 23. The default value is 5.
18859 @item chroma_msize_y, cy
18860 Set the chroma matrix vertical size. It must be an odd integer
18861 between 3 and 23. The default value is 5.
18863 @item chroma_amount, ca
18864 Set the chroma effect strength. It must be a floating point number, reasonable
18865 values lay between -1.5 and 1.5.
18867 Negative values will blur the input video, while positive values will
18868 sharpen it, a value of zero will disable the effect.
18870 Default value is 0.0.
18874 All parameters are optional and default to the equivalent of the
18875 string '5:5:1.0:5:5:0.0'.
18877 @subsection Examples
18881 Apply strong luma sharpen effect:
18883 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
18887 Apply a strong blur of both luma and chroma parameters:
18889 unsharp=7:7:-2:7:7:-2
18895 Apply ultra slow/simple postprocessing filter that compresses and decompresses
18896 the image at several (or - in the case of @option{quality} level @code{8} - all)
18897 shifts and average the results.
18899 The way this differs from the behavior of spp is that uspp actually encodes &
18900 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
18901 DCT similar to MJPEG.
18903 The filter accepts the following options:
18907 Set quality. This option defines the number of levels for averaging. It accepts
18908 an integer in the range 0-8. If set to @code{0}, the filter will have no
18909 effect. A value of @code{8} means the higher quality. For each increment of
18910 that value the speed drops by a factor of approximately 2. Default value is
18914 Force a constant quantization parameter. If not set, the filter will use the QP
18915 from the video stream (if available).
18920 Convert 360 videos between various formats.
18922 The filter accepts the following options:
18928 Set format of the input/output video.
18936 Equirectangular projection.
18941 Cubemap with 3x2/6x1/1x6 layout.
18943 Format specific options:
18948 Set padding proportion for the input/output cubemap. Values in decimals.
18955 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)
18958 Default value is @b{@samp{0}}.
18962 Set fixed padding for the input/output cubemap. Values in pixels.
18964 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
18968 Set order of faces for the input/output cubemap. Choose one direction for each position.
18970 Designation of directions:
18986 Default value is @b{@samp{rludfb}}.
18990 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
18992 Designation of angles:
18995 0 degrees clockwise
18997 90 degrees clockwise
18999 180 degrees clockwise
19001 270 degrees clockwise
19004 Default value is @b{@samp{000000}}.
19008 Equi-Angular Cubemap.
19015 Format specific options:
19020 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19022 If diagonal field of view is set it overrides horizontal and vertical field of view.
19027 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19029 If diagonal field of view is set it overrides horizontal and vertical field of view.
19035 Format specific options:
19039 Set padding proportion. Values in decimals.
19049 Default value is @b{@samp{0}}.
19055 Facebook's 360 formats.
19058 Stereographic format.
19060 Format specific options:
19065 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19067 If diagonal field of view is set it overrides horizontal and vertical field of view.
19072 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19074 If diagonal field of view is set it overrides horizontal and vertical field of view.
19081 Ball format, gives significant distortion toward the back.
19084 Hammer-Aitoff map projection format.
19087 Sinusoidal map projection format.
19090 Fisheye projection.
19092 Format specific options:
19097 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19099 If diagonal field of view is set it overrides horizontal and vertical field of view.
19104 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19106 If diagonal field of view is set it overrides horizontal and vertical field of view.
19110 Pannini projection.
19112 Format specific options:
19115 Set output pannini parameter.
19118 Set input pannini parameter.
19122 Cylindrical projection.
19124 Format specific options:
19129 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19131 If diagonal field of view is set it overrides horizontal and vertical field of view.
19136 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19138 If diagonal field of view is set it overrides horizontal and vertical field of view.
19142 Perspective projection. @i{(output only)}
19144 Format specific options:
19147 Set perspective parameter.
19151 Tetrahedron projection.
19154 Truncated square pyramid projection.
19158 Half equirectangular projection.
19162 Set interpolation method.@*
19163 @i{Note: more complex interpolation methods require much more memory to run.}
19173 Bilinear interpolation.
19175 Lagrange9 interpolation.
19178 Bicubic interpolation.
19181 Lanczos interpolation.
19184 Spline16 interpolation.
19187 Gaussian interpolation.
19190 Default value is @b{@samp{line}}.
19194 Set the output video resolution.
19196 Default resolution depends on formats.
19200 Set the input/output stereo format.
19211 Default value is @b{@samp{2d}} for input and output format.
19216 Set rotation for the output video. Values in degrees.
19219 Set rotation order for the output video. Choose one item for each position.
19230 Default value is @b{@samp{ypr}}.
19235 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
19239 Set if input video is flipped horizontally/vertically. Boolean values.
19242 Set if input video is transposed. Boolean value, by default disabled.
19245 Set if output video needs to be transposed. Boolean value, by default disabled.
19248 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
19251 @subsection Examples
19255 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
19257 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
19260 Extract back view of Equi-Angular Cubemap:
19262 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
19265 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
19267 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
19271 @subsection Commands
19273 This filter supports subset of above options as @ref{commands}.
19275 @section vaguedenoiser
19277 Apply a wavelet based denoiser.
19279 It transforms each frame from the video input into the wavelet domain,
19280 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
19281 the obtained coefficients. It does an inverse wavelet transform after.
19282 Due to wavelet properties, it should give a nice smoothed result, and
19283 reduced noise, without blurring picture features.
19285 This filter accepts the following options:
19289 The filtering strength. The higher, the more filtered the video will be.
19290 Hard thresholding can use a higher threshold than soft thresholding
19291 before the video looks overfiltered. Default value is 2.
19294 The filtering method the filter will use.
19296 It accepts the following values:
19299 All values under the threshold will be zeroed.
19302 All values under the threshold will be zeroed. All values above will be
19303 reduced by the threshold.
19306 Scales or nullifies coefficients - intermediary between (more) soft and
19307 (less) hard thresholding.
19310 Default is garrote.
19313 Number of times, the wavelet will decompose the picture. Picture can't
19314 be decomposed beyond a particular point (typically, 8 for a 640x480
19315 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
19318 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
19321 A list of the planes to process. By default all planes are processed.
19324 @section vectorscope
19326 Display 2 color component values in the two dimensional graph (which is called
19329 This filter accepts the following options:
19333 Set vectorscope mode.
19335 It accepts the following values:
19339 Gray values are displayed on graph, higher brightness means more pixels have
19340 same component color value on location in graph. This is the default mode.
19343 Gray values are displayed on graph. Surrounding pixels values which are not
19344 present in video frame are drawn in gradient of 2 color components which are
19345 set by option @code{x} and @code{y}. The 3rd color component is static.
19348 Actual color components values present in video frame are displayed on graph.
19351 Similar as color2 but higher frequency of same values @code{x} and @code{y}
19352 on graph increases value of another color component, which is luminance by
19353 default values of @code{x} and @code{y}.
19356 Actual colors present in video frame are displayed on graph. If two different
19357 colors map to same position on graph then color with higher value of component
19358 not present in graph is picked.
19361 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
19362 component picked from radial gradient.
19366 Set which color component will be represented on X-axis. Default is @code{1}.
19369 Set which color component will be represented on Y-axis. Default is @code{2}.
19372 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
19373 of color component which represents frequency of (X, Y) location in graph.
19378 No envelope, this is default.
19381 Instant envelope, even darkest single pixel will be clearly highlighted.
19384 Hold maximum and minimum values presented in graph over time. This way you
19385 can still spot out of range values without constantly looking at vectorscope.
19388 Peak and instant envelope combined together.
19392 Set what kind of graticule to draw.
19401 Set graticule opacity.
19404 Set graticule flags.
19408 Draw graticule for white point.
19411 Draw graticule for black point.
19414 Draw color points short names.
19418 Set background opacity.
19420 @item lthreshold, l
19421 Set low threshold for color component not represented on X or Y axis.
19422 Values lower than this value will be ignored. Default is 0.
19423 Note this value is multiplied with actual max possible value one pixel component
19424 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
19427 @item hthreshold, h
19428 Set high threshold for color component not represented on X or Y axis.
19429 Values higher than this value will be ignored. Default is 1.
19430 Note this value is multiplied with actual max possible value one pixel component
19431 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
19432 is 0.9 * 255 = 230.
19434 @item colorspace, c
19435 Set what kind of colorspace to use when drawing graticule.
19445 Set color tint for gray/tint vectorscope mode. By default both options are zero.
19446 This means no tint, and output will remain gray.
19449 @anchor{vidstabdetect}
19450 @section vidstabdetect
19452 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
19453 @ref{vidstabtransform} for pass 2.
19455 This filter generates a file with relative translation and rotation
19456 transform information about subsequent frames, which is then used by
19457 the @ref{vidstabtransform} filter.
19459 To enable compilation of this filter you need to configure FFmpeg with
19460 @code{--enable-libvidstab}.
19462 This filter accepts the following options:
19466 Set the path to the file used to write the transforms information.
19467 Default value is @file{transforms.trf}.
19470 Set how shaky the video is and how quick the camera is. It accepts an
19471 integer in the range 1-10, a value of 1 means little shakiness, a
19472 value of 10 means strong shakiness. Default value is 5.
19475 Set the accuracy of the detection process. It must be a value in the
19476 range 1-15. A value of 1 means low accuracy, a value of 15 means high
19477 accuracy. Default value is 15.
19480 Set stepsize of the search process. The region around minimum is
19481 scanned with 1 pixel resolution. Default value is 6.
19484 Set minimum contrast. Below this value a local measurement field is
19485 discarded. Must be a floating point value in the range 0-1. Default
19489 Set reference frame number for tripod mode.
19491 If enabled, the motion of the frames is compared to a reference frame
19492 in the filtered stream, identified by the specified number. The idea
19493 is to compensate all movements in a more-or-less static scene and keep
19494 the camera view absolutely still.
19496 If set to 0, it is disabled. The frames are counted starting from 1.
19499 Show fields and transforms in the resulting frames. It accepts an
19500 integer in the range 0-2. Default value is 0, which disables any
19504 @subsection Examples
19508 Use default values:
19514 Analyze strongly shaky movie and put the results in file
19515 @file{mytransforms.trf}:
19517 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
19521 Visualize the result of internal transformations in the resulting
19524 vidstabdetect=show=1
19528 Analyze a video with medium shakiness using @command{ffmpeg}:
19530 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
19534 @anchor{vidstabtransform}
19535 @section vidstabtransform
19537 Video stabilization/deshaking: pass 2 of 2,
19538 see @ref{vidstabdetect} for pass 1.
19540 Read a file with transform information for each frame and
19541 apply/compensate them. Together with the @ref{vidstabdetect}
19542 filter this can be used to deshake videos. See also
19543 @url{http://public.hronopik.de/vid.stab}. It is important to also use
19544 the @ref{unsharp} filter, see below.
19546 To enable compilation of this filter you need to configure FFmpeg with
19547 @code{--enable-libvidstab}.
19549 @subsection Options
19553 Set path to the file used to read the transforms. Default value is
19554 @file{transforms.trf}.
19557 Set the number of frames (value*2 + 1) used for lowpass filtering the
19558 camera movements. Default value is 10.
19560 For example a number of 10 means that 21 frames are used (10 in the
19561 past and 10 in the future) to smoothen the motion in the video. A
19562 larger value leads to a smoother video, but limits the acceleration of
19563 the camera (pan/tilt movements). 0 is a special case where a static
19564 camera is simulated.
19567 Set the camera path optimization algorithm.
19569 Accepted values are:
19572 gaussian kernel low-pass filter on camera motion (default)
19574 averaging on transformations
19578 Set maximal number of pixels to translate frames. Default value is -1,
19582 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
19583 value is -1, meaning no limit.
19586 Specify how to deal with borders that may be visible due to movement
19589 Available values are:
19592 keep image information from previous frame (default)
19594 fill the border black
19598 Invert transforms if set to 1. Default value is 0.
19601 Consider transforms as relative to previous frame if set to 1,
19602 absolute if set to 0. Default value is 0.
19605 Set percentage to zoom. A positive value will result in a zoom-in
19606 effect, a negative value in a zoom-out effect. Default value is 0 (no
19610 Set optimal zooming to avoid borders.
19612 Accepted values are:
19617 optimal static zoom value is determined (only very strong movements
19618 will lead to visible borders) (default)
19620 optimal adaptive zoom value is determined (no borders will be
19621 visible), see @option{zoomspeed}
19624 Note that the value given at zoom is added to the one calculated here.
19627 Set percent to zoom maximally each frame (enabled when
19628 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
19632 Specify type of interpolation.
19634 Available values are:
19639 linear only horizontal
19641 linear in both directions (default)
19643 cubic in both directions (slow)
19647 Enable virtual tripod mode if set to 1, which is equivalent to
19648 @code{relative=0:smoothing=0}. Default value is 0.
19650 Use also @code{tripod} option of @ref{vidstabdetect}.
19653 Increase log verbosity if set to 1. Also the detected global motions
19654 are written to the temporary file @file{global_motions.trf}. Default
19658 @subsection Examples
19662 Use @command{ffmpeg} for a typical stabilization with default values:
19664 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
19667 Note the use of the @ref{unsharp} filter which is always recommended.
19670 Zoom in a bit more and load transform data from a given file:
19672 vidstabtransform=zoom=5:input="mytransforms.trf"
19676 Smoothen the video even more:
19678 vidstabtransform=smoothing=30
19684 Flip the input video vertically.
19686 For example, to vertically flip a video with @command{ffmpeg}:
19688 ffmpeg -i in.avi -vf "vflip" out.avi
19693 Detect variable frame rate video.
19695 This filter tries to detect if the input is variable or constant frame rate.
19697 At end it will output number of frames detected as having variable delta pts,
19698 and ones with constant delta pts.
19699 If there was frames with variable delta, than it will also show min, max and
19700 average delta encountered.
19704 Boost or alter saturation.
19706 The filter accepts the following options:
19709 Set strength of boost if positive value or strength of alter if negative value.
19710 Default is 0. Allowed range is from -2 to 2.
19713 Set the red balance. Default is 1. Allowed range is from -10 to 10.
19716 Set the green balance. Default is 1. Allowed range is from -10 to 10.
19719 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
19722 Set the red luma coefficient.
19725 Set the green luma coefficient.
19728 Set the blue luma coefficient.
19731 If @code{intensity} is negative and this is set to 1, colors will change,
19732 otherwise colors will be less saturated, more towards gray.
19735 @subsection Commands
19737 This filter supports the all above options as @ref{commands}.
19742 Make or reverse a natural vignetting effect.
19744 The filter accepts the following options:
19748 Set lens angle expression as a number of radians.
19750 The value is clipped in the @code{[0,PI/2]} range.
19752 Default value: @code{"PI/5"}
19756 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
19760 Set forward/backward mode.
19762 Available modes are:
19765 The larger the distance from the central point, the darker the image becomes.
19768 The larger the distance from the central point, the brighter the image becomes.
19769 This can be used to reverse a vignette effect, though there is no automatic
19770 detection to extract the lens @option{angle} and other settings (yet). It can
19771 also be used to create a burning effect.
19774 Default value is @samp{forward}.
19777 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
19779 It accepts the following values:
19782 Evaluate expressions only once during the filter initialization.
19785 Evaluate expressions for each incoming frame. This is way slower than the
19786 @samp{init} mode since it requires all the scalers to be re-computed, but it
19787 allows advanced dynamic expressions.
19790 Default value is @samp{init}.
19793 Set dithering to reduce the circular banding effects. Default is @code{1}
19797 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
19798 Setting this value to the SAR of the input will make a rectangular vignetting
19799 following the dimensions of the video.
19801 Default is @code{1/1}.
19804 @subsection Expressions
19806 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
19807 following parameters.
19812 input width and height
19815 the number of input frame, starting from 0
19818 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
19819 @var{TB} units, NAN if undefined
19822 frame rate of the input video, NAN if the input frame rate is unknown
19825 the PTS (Presentation TimeStamp) of the filtered video frame,
19826 expressed in seconds, NAN if undefined
19829 time base of the input video
19833 @subsection Examples
19837 Apply simple strong vignetting effect:
19843 Make a flickering vignetting:
19845 vignette='PI/4+random(1)*PI/50':eval=frame
19850 @section vmafmotion
19852 Obtain the average VMAF motion score of a video.
19853 It is one of the component metrics of VMAF.
19855 The obtained average motion score is printed through the logging system.
19857 The filter accepts the following options:
19861 If specified, the filter will use the named file to save the motion score of
19862 each frame with respect to the previous frame.
19863 When filename equals "-" the data is sent to standard output.
19868 ffmpeg -i ref.mpg -vf vmafmotion -f null -
19872 Stack input videos vertically.
19874 All streams must be of same pixel format and of same width.
19876 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
19877 to create same output.
19879 The filter accepts the following options:
19883 Set number of input streams. Default is 2.
19886 If set to 1, force the output to terminate when the shortest input
19887 terminates. Default value is 0.
19892 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
19893 Deinterlacing Filter").
19895 Based on the process described by Martin Weston for BBC R&D, and
19896 implemented based on the de-interlace algorithm written by Jim
19897 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
19898 uses filter coefficients calculated by BBC R&D.
19900 This filter uses field-dominance information in frame to decide which
19901 of each pair of fields to place first in the output.
19902 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
19904 There are two sets of filter coefficients, so called "simple"
19905 and "complex". Which set of filter coefficients is used can
19906 be set by passing an optional parameter:
19910 Set the interlacing filter coefficients. Accepts one of the following values:
19914 Simple filter coefficient set.
19916 More-complex filter coefficient set.
19918 Default value is @samp{complex}.
19921 Specify which frames to deinterlace. Accepts one of the following values:
19925 Deinterlace all frames,
19927 Only deinterlace frames marked as interlaced.
19930 Default value is @samp{all}.
19934 Video waveform monitor.
19936 The waveform monitor plots color component intensity. By default luminance
19937 only. Each column of the waveform corresponds to a column of pixels in the
19940 It accepts the following options:
19944 Can be either @code{row}, or @code{column}. Default is @code{column}.
19945 In row mode, the graph on the left side represents color component value 0 and
19946 the right side represents value = 255. In column mode, the top side represents
19947 color component value = 0 and bottom side represents value = 255.
19950 Set intensity. Smaller values are useful to find out how many values of the same
19951 luminance are distributed across input rows/columns.
19952 Default value is @code{0.04}. Allowed range is [0, 1].
19955 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
19956 In mirrored mode, higher values will be represented on the left
19957 side for @code{row} mode and at the top for @code{column} mode. Default is
19958 @code{1} (mirrored).
19962 It accepts the following values:
19965 Presents information identical to that in the @code{parade}, except
19966 that the graphs representing color components are superimposed directly
19969 This display mode makes it easier to spot relative differences or similarities
19970 in overlapping areas of the color components that are supposed to be identical,
19971 such as neutral whites, grays, or blacks.
19974 Display separate graph for the color components side by side in
19975 @code{row} mode or one below the other in @code{column} mode.
19978 Display separate graph for the color components side by side in
19979 @code{column} mode or one below the other in @code{row} mode.
19981 Using this display mode makes it easy to spot color casts in the highlights
19982 and shadows of an image, by comparing the contours of the top and the bottom
19983 graphs of each waveform. Since whites, grays, and blacks are characterized
19984 by exactly equal amounts of red, green, and blue, neutral areas of the picture
19985 should display three waveforms of roughly equal width/height. If not, the
19986 correction is easy to perform by making level adjustments the three waveforms.
19988 Default is @code{stack}.
19990 @item components, c
19991 Set which color components to display. Default is 1, which means only luminance
19992 or red color component if input is in RGB colorspace. If is set for example to
19993 7 it will display all 3 (if) available color components.
19998 No envelope, this is default.
20001 Instant envelope, minimum and maximum values presented in graph will be easily
20002 visible even with small @code{step} value.
20005 Hold minimum and maximum values presented in graph across time. This way you
20006 can still spot out of range values without constantly looking at waveforms.
20009 Peak and instant envelope combined together.
20015 No filtering, this is default.
20018 Luma and chroma combined together.
20021 Similar as above, but shows difference between blue and red chroma.
20024 Similar as above, but use different colors.
20027 Similar as above, but again with different colors.
20030 Displays only chroma.
20033 Displays actual color value on waveform.
20036 Similar as above, but with luma showing frequency of chroma values.
20040 Set which graticule to display.
20044 Do not display graticule.
20047 Display green graticule showing legal broadcast ranges.
20050 Display orange graticule showing legal broadcast ranges.
20053 Display invert graticule showing legal broadcast ranges.
20057 Set graticule opacity.
20060 Set graticule flags.
20064 Draw numbers above lines. By default enabled.
20067 Draw dots instead of lines.
20071 Set scale used for displaying graticule.
20078 Default is digital.
20081 Set background opacity.
20085 Set tint for output.
20086 Only used with lowpass filter and when display is not overlay and input
20087 pixel formats are not RGB.
20090 @section weave, doubleweave
20092 The @code{weave} takes a field-based video input and join
20093 each two sequential fields into single frame, producing a new double
20094 height clip with half the frame rate and half the frame count.
20096 The @code{doubleweave} works same as @code{weave} but without
20097 halving frame rate and frame count.
20099 It accepts the following option:
20103 Set first field. Available values are:
20107 Set the frame as top-field-first.
20110 Set the frame as bottom-field-first.
20114 @subsection Examples
20118 Interlace video using @ref{select} and @ref{separatefields} filter:
20120 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
20125 Apply the xBR high-quality magnification filter which is designed for pixel
20126 art. It follows a set of edge-detection rules, see
20127 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
20129 It accepts the following option:
20133 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
20134 @code{3xBR} and @code{4} for @code{4xBR}.
20135 Default is @code{3}.
20140 Apply cross fade from one input video stream to another input video stream.
20141 The cross fade is applied for specified duration.
20143 The filter accepts the following options:
20147 Set one of available transition effects:
20187 Default transition effect is fade.
20190 Set cross fade duration in seconds.
20191 Default duration is 1 second.
20194 Set cross fade start relative to first input stream in seconds.
20195 Default offset is 0.
20198 Set expression for custom transition effect.
20200 The expressions can use the following variables and functions:
20205 The coordinates of the current sample.
20209 The width and height of the image.
20212 Progress of transition effect.
20215 Currently processed plane.
20218 Return value of first input at current location and plane.
20221 Return value of second input at current location and plane.
20227 Return the value of the pixel at location (@var{x},@var{y}) of the
20228 first/second/third/fourth component of first input.
20234 Return the value of the pixel at location (@var{x},@var{y}) of the
20235 first/second/third/fourth component of second input.
20239 @subsection Examples
20243 Cross fade from one input video to another input video, with fade transition and duration of transition
20244 of 2 seconds starting at offset of 5 seconds:
20246 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
20251 Pick median pixels from several input videos.
20253 The filter accepts the following options:
20257 Set number of inputs.
20258 Default is 3. Allowed range is from 3 to 255.
20259 If number of inputs is even number, than result will be mean value between two median values.
20262 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
20265 Set median percentile. Default value is @code{0.5}.
20266 Default value of @code{0.5} will pick always median values, while @code{0} will pick
20267 minimum values, and @code{1} maximum values.
20271 Stack video inputs into custom layout.
20273 All streams must be of same pixel format.
20275 The filter accepts the following options:
20279 Set number of input streams. Default is 2.
20282 Specify layout of inputs.
20283 This option requires the desired layout configuration to be explicitly set by the user.
20284 This sets position of each video input in output. Each input
20285 is separated by '|'.
20286 The first number represents the column, and the second number represents the row.
20287 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
20288 where X is video input from which to take width or height.
20289 Multiple values can be used when separated by '+'. In such
20290 case values are summed together.
20292 Note that if inputs are of different sizes gaps may appear, as not all of
20293 the output video frame will be filled. Similarly, videos can overlap each
20294 other if their position doesn't leave enough space for the full frame of
20297 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
20298 a layout must be set by the user.
20301 If set to 1, force the output to terminate when the shortest input
20302 terminates. Default value is 0.
20305 If set to valid color, all unused pixels will be filled with that color.
20306 By default fill is set to none, so it is disabled.
20309 @subsection Examples
20313 Display 4 inputs into 2x2 grid.
20317 input1(0, 0) | input3(w0, 0)
20318 input2(0, h0) | input4(w0, h0)
20322 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
20325 Note that if inputs are of different sizes, gaps or overlaps may occur.
20328 Display 4 inputs into 1x4 grid.
20335 input4(0, h0+h1+h2)
20339 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
20342 Note that if inputs are of different widths, unused space will appear.
20345 Display 9 inputs into 3x3 grid.
20349 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
20350 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
20351 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
20355 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
20358 Note that if inputs are of different sizes, gaps or overlaps may occur.
20361 Display 16 inputs into 4x4 grid.
20365 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
20366 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
20367 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
20368 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
20372 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|
20373 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
20376 Note that if inputs are of different sizes, gaps or overlaps may occur.
20383 Deinterlace the input video ("yadif" means "yet another deinterlacing
20386 It accepts the following parameters:
20392 The interlacing mode to adopt. It accepts one of the following values:
20395 @item 0, send_frame
20396 Output one frame for each frame.
20397 @item 1, send_field
20398 Output one frame for each field.
20399 @item 2, send_frame_nospatial
20400 Like @code{send_frame}, but it skips the spatial interlacing check.
20401 @item 3, send_field_nospatial
20402 Like @code{send_field}, but it skips the spatial interlacing check.
20405 The default value is @code{send_frame}.
20408 The picture field parity assumed for the input interlaced video. It accepts one
20409 of the following values:
20413 Assume the top field is first.
20415 Assume the bottom field is first.
20417 Enable automatic detection of field parity.
20420 The default value is @code{auto}.
20421 If the interlacing is unknown or the decoder does not export this information,
20422 top field first will be assumed.
20425 Specify which frames to deinterlace. Accepts one of the following
20430 Deinterlace all frames.
20431 @item 1, interlaced
20432 Only deinterlace frames marked as interlaced.
20435 The default value is @code{all}.
20438 @section yadif_cuda
20440 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
20441 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
20444 It accepts the following parameters:
20450 The interlacing mode to adopt. It accepts one of the following values:
20453 @item 0, send_frame
20454 Output one frame for each frame.
20455 @item 1, send_field
20456 Output one frame for each field.
20457 @item 2, send_frame_nospatial
20458 Like @code{send_frame}, but it skips the spatial interlacing check.
20459 @item 3, send_field_nospatial
20460 Like @code{send_field}, but it skips the spatial interlacing check.
20463 The default value is @code{send_frame}.
20466 The picture field parity assumed for the input interlaced video. It accepts one
20467 of the following values:
20471 Assume the top field is first.
20473 Assume the bottom field is first.
20475 Enable automatic detection of field parity.
20478 The default value is @code{auto}.
20479 If the interlacing is unknown or the decoder does not export this information,
20480 top field first will be assumed.
20483 Specify which frames to deinterlace. Accepts one of the following
20488 Deinterlace all frames.
20489 @item 1, interlaced
20490 Only deinterlace frames marked as interlaced.
20493 The default value is @code{all}.
20498 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
20499 The algorithm is described in
20500 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
20502 It accepts the following parameters:
20506 Set the window radius. Default value is 3.
20509 Set which planes to filter. Default is only the first plane.
20512 Set blur strength. Default value is 128.
20515 @subsection Commands
20516 This filter supports same @ref{commands} as options.
20520 Apply Zoom & Pan effect.
20522 This filter accepts the following options:
20526 Set the zoom expression. Range is 1-10. Default is 1.
20530 Set the x and y expression. Default is 0.
20533 Set the duration expression in number of frames.
20534 This sets for how many number of frames effect will last for
20535 single input image.
20538 Set the output image size, default is 'hd720'.
20541 Set the output frame rate, default is '25'.
20544 Each expression can contain the following constants:
20563 Output frame count.
20567 Last calculated 'x' and 'y' position from 'x' and 'y' expression
20568 for current input frame.
20572 'x' and 'y' of last output frame of previous input frame or 0 when there was
20573 not yet such frame (first input frame).
20576 Last calculated zoom from 'z' expression for current input frame.
20579 Last calculated zoom of last output frame of previous input frame.
20582 Number of output frames for current input frame. Calculated from 'd' expression
20583 for each input frame.
20586 number of output frames created for previous input frame
20589 Rational number: input width / input height
20592 sample aspect ratio
20595 display aspect ratio
20599 @subsection Examples
20603 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
20605 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
20609 Zoom-in up to 1.5 and pan always at center of picture:
20611 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
20615 Same as above but without pausing:
20617 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
20623 Scale (resize) the input video, using the z.lib library:
20624 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
20625 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
20627 The zscale filter forces the output display aspect ratio to be the same
20628 as the input, by changing the output sample aspect ratio.
20630 If the input image format is different from the format requested by
20631 the next filter, the zscale filter will convert the input to the
20634 @subsection Options
20635 The filter accepts the following options.
20640 Set the output video dimension expression. Default value is the input
20643 If the @var{width} or @var{w} value is 0, the input width is used for
20644 the output. If the @var{height} or @var{h} value is 0, the input height
20645 is used for the output.
20647 If one and only one of the values is -n with n >= 1, the zscale filter
20648 will use a value that maintains the aspect ratio of the input image,
20649 calculated from the other specified dimension. After that it will,
20650 however, make sure that the calculated dimension is divisible by n and
20651 adjust the value if necessary.
20653 If both values are -n with n >= 1, the behavior will be identical to
20654 both values being set to 0 as previously detailed.
20656 See below for the list of accepted constants for use in the dimension
20660 Set the video size. For the syntax of this option, check the
20661 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20664 Set the dither type.
20666 Possible values are:
20671 @item error_diffusion
20677 Set the resize filter type.
20679 Possible values are:
20689 Default is bilinear.
20692 Set the color range.
20694 Possible values are:
20701 Default is same as input.
20704 Set the color primaries.
20706 Possible values are:
20716 Default is same as input.
20719 Set the transfer characteristics.
20721 Possible values are:
20735 Default is same as input.
20738 Set the colorspace matrix.
20740 Possible value are:
20751 Default is same as input.
20754 Set the input color range.
20756 Possible values are:
20763 Default is same as input.
20765 @item primariesin, pin
20766 Set the input color primaries.
20768 Possible values are:
20778 Default is same as input.
20780 @item transferin, tin
20781 Set the input transfer characteristics.
20783 Possible values are:
20794 Default is same as input.
20796 @item matrixin, min
20797 Set the input colorspace matrix.
20799 Possible value are:
20811 Set the output chroma location.
20813 Possible values are:
20824 @item chromalin, cin
20825 Set the input chroma location.
20827 Possible values are:
20839 Set the nominal peak luminance.
20842 The values of the @option{w} and @option{h} options are expressions
20843 containing the following constants:
20848 The input width and height
20852 These are the same as @var{in_w} and @var{in_h}.
20856 The output (scaled) width and height
20860 These are the same as @var{out_w} and @var{out_h}
20863 The same as @var{iw} / @var{ih}
20866 input sample aspect ratio
20869 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
20873 horizontal and vertical input chroma subsample values. For example for the
20874 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
20878 horizontal and vertical output chroma subsample values. For example for the
20879 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
20882 @subsection Commands
20884 This filter supports the following commands:
20888 Set the output video dimension expression.
20889 The command accepts the same syntax of the corresponding option.
20891 If the specified expression is not valid, it is kept at its current
20895 @c man end VIDEO FILTERS
20897 @chapter OpenCL Video Filters
20898 @c man begin OPENCL VIDEO FILTERS
20900 Below is a description of the currently available OpenCL video filters.
20902 To enable compilation of these filters you need to configure FFmpeg with
20903 @code{--enable-opencl}.
20905 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
20908 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
20909 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
20910 given device parameters.
20912 @item -filter_hw_device @var{name}
20913 Pass the hardware device called @var{name} to all filters in any filter graph.
20917 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
20921 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
20923 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
20927 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.
20929 @section avgblur_opencl
20931 Apply average blur filter.
20933 The filter accepts the following options:
20937 Set horizontal radius size.
20938 Range is @code{[1, 1024]} and default value is @code{1}.
20941 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20944 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
20947 @subsection Example
20951 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.
20953 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
20957 @section boxblur_opencl
20959 Apply a boxblur algorithm to the input video.
20961 It accepts the following parameters:
20965 @item luma_radius, lr
20966 @item luma_power, lp
20967 @item chroma_radius, cr
20968 @item chroma_power, cp
20969 @item alpha_radius, ar
20970 @item alpha_power, ap
20974 A description of the accepted options follows.
20977 @item luma_radius, lr
20978 @item chroma_radius, cr
20979 @item alpha_radius, ar
20980 Set an expression for the box radius in pixels used for blurring the
20981 corresponding input plane.
20983 The radius value must be a non-negative number, and must not be
20984 greater than the value of the expression @code{min(w,h)/2} for the
20985 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
20988 Default value for @option{luma_radius} is "2". If not specified,
20989 @option{chroma_radius} and @option{alpha_radius} default to the
20990 corresponding value set for @option{luma_radius}.
20992 The expressions can contain the following constants:
20996 The input width and height in pixels.
21000 The input chroma image width and height in pixels.
21004 The horizontal and vertical chroma subsample values. For example, for the
21005 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
21008 @item luma_power, lp
21009 @item chroma_power, cp
21010 @item alpha_power, ap
21011 Specify how many times the boxblur filter is applied to the
21012 corresponding plane.
21014 Default value for @option{luma_power} is 2. If not specified,
21015 @option{chroma_power} and @option{alpha_power} default to the
21016 corresponding value set for @option{luma_power}.
21018 A value of 0 will disable the effect.
21021 @subsection Examples
21023 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.
21027 Apply a boxblur filter with the luma, chroma, and alpha radius
21028 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.
21030 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
21031 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
21035 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.
21037 For the luma plane, a 2x2 box radius will be run once.
21039 For the chroma plane, a 4x4 box radius will be run 5 times.
21041 For the alpha plane, a 3x3 box radius will be run 7 times.
21043 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
21047 @section colorkey_opencl
21048 RGB colorspace color keying.
21050 The filter accepts the following options:
21054 The color which will be replaced with transparency.
21057 Similarity percentage with the key color.
21059 0.01 matches only the exact key color, while 1.0 matches everything.
21064 0.0 makes pixels either fully transparent, or not transparent at all.
21066 Higher values result in semi-transparent pixels, with a higher transparency
21067 the more similar the pixels color is to the key color.
21070 @subsection Examples
21074 Make every semi-green pixel in the input transparent with some slight blending:
21076 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
21080 @section convolution_opencl
21082 Apply convolution of 3x3, 5x5, 7x7 matrix.
21084 The filter accepts the following options:
21091 Set matrix for each plane.
21092 Matrix is sequence of 9, 25 or 49 signed numbers.
21093 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
21099 Set multiplier for calculated value for each plane.
21100 If unset or 0, it will be sum of all matrix elements.
21101 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
21107 Set bias for each plane. This value is added to the result of the multiplication.
21108 Useful for making the overall image brighter or darker.
21109 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
21113 @subsection Examples
21119 -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
21125 -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
21129 Apply edge enhance:
21131 -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
21137 -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
21141 Apply laplacian edge detector which includes diagonals:
21143 -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
21149 -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
21153 @section erosion_opencl
21155 Apply erosion effect to the video.
21157 This filter replaces the pixel by the local(3x3) minimum.
21159 It accepts the following options:
21166 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21167 If @code{0}, plane will remain unchanged.
21170 Flag which specifies the pixel to refer to.
21171 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21173 Flags to local 3x3 coordinates region centered on @code{x}:
21182 @subsection Example
21186 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.
21188 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21192 @section deshake_opencl
21193 Feature-point based video stabilization filter.
21195 The filter accepts the following options:
21199 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
21202 Whether or not additional debug info should be displayed, both in the processed output and in the console.
21204 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
21206 Viewing point matches in the output video is only supported for RGB input.
21208 Defaults to @code{0}.
21210 @item adaptive_crop
21211 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
21213 Defaults to @code{1}.
21215 @item refine_features
21216 Whether or not feature points should be refined at a sub-pixel level.
21218 This can be turned off for a slight performance gain at the cost of precision.
21220 Defaults to @code{1}.
21222 @item smooth_strength
21223 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
21225 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
21227 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
21229 Defaults to @code{0.0}.
21231 @item smooth_window_multiplier
21232 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
21234 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
21236 Acceptable values range from @code{0.1} to @code{10.0}.
21238 Larger values increase the amount of motion data available for determining how to smooth the camera path,
21239 potentially improving smoothness, but also increase latency and memory usage.
21241 Defaults to @code{2.0}.
21245 @subsection Examples
21249 Stabilize a video with a fixed, medium smoothing strength:
21251 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
21255 Stabilize a video with debugging (both in console and in rendered video):
21257 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
21261 @section dilation_opencl
21263 Apply dilation effect to the video.
21265 This filter replaces the pixel by the local(3x3) maximum.
21267 It accepts the following options:
21274 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21275 If @code{0}, plane will remain unchanged.
21278 Flag which specifies the pixel to refer to.
21279 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21281 Flags to local 3x3 coordinates region centered on @code{x}:
21290 @subsection Example
21294 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.
21296 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21300 @section nlmeans_opencl
21302 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
21304 @section overlay_opencl
21306 Overlay one video on top of another.
21308 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
21309 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
21311 The filter accepts the following options:
21316 Set the x coordinate of the overlaid video on the main video.
21317 Default value is @code{0}.
21320 Set the y coordinate of the overlaid video on the main video.
21321 Default value is @code{0}.
21325 @subsection Examples
21329 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
21331 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21334 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
21336 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21341 @section pad_opencl
21343 Add paddings to the input image, and place the original input at the
21344 provided @var{x}, @var{y} coordinates.
21346 It accepts the following options:
21351 Specify an expression for the size of the output image with the
21352 paddings added. If the value for @var{width} or @var{height} is 0, the
21353 corresponding input size is used for the output.
21355 The @var{width} expression can reference the value set by the
21356 @var{height} expression, and vice versa.
21358 The default value of @var{width} and @var{height} is 0.
21362 Specify the offsets to place the input image at within the padded area,
21363 with respect to the top/left border of the output image.
21365 The @var{x} expression can reference the value set by the @var{y}
21366 expression, and vice versa.
21368 The default value of @var{x} and @var{y} is 0.
21370 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
21371 so the input image is centered on the padded area.
21374 Specify the color of the padded area. For the syntax of this option,
21375 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
21376 manual,ffmpeg-utils}.
21379 Pad to an aspect instead to a resolution.
21382 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
21383 options are expressions containing the following constants:
21388 The input video width and height.
21392 These are the same as @var{in_w} and @var{in_h}.
21396 The output width and height (the size of the padded area), as
21397 specified by the @var{width} and @var{height} expressions.
21401 These are the same as @var{out_w} and @var{out_h}.
21405 The x and y offsets as specified by the @var{x} and @var{y}
21406 expressions, or NAN if not yet specified.
21409 same as @var{iw} / @var{ih}
21412 input sample aspect ratio
21415 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
21418 @section prewitt_opencl
21420 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
21422 The filter accepts the following option:
21426 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21429 Set value which will be multiplied with filtered result.
21430 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21433 Set value which will be added to filtered result.
21434 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21437 @subsection Example
21441 Apply the Prewitt operator with scale set to 2 and delta set to 10.
21443 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
21447 @anchor{program_opencl}
21448 @section program_opencl
21450 Filter video using an OpenCL program.
21455 OpenCL program source file.
21458 Kernel name in program.
21461 Number of inputs to the filter. Defaults to 1.
21464 Size of output frames. Defaults to the same as the first input.
21468 The @code{program_opencl} filter also supports the @ref{framesync} options.
21470 The program source file must contain a kernel function with the given name,
21471 which will be run once for each plane of the output. Each run on a plane
21472 gets enqueued as a separate 2D global NDRange with one work-item for each
21473 pixel to be generated. The global ID offset for each work-item is therefore
21474 the coordinates of a pixel in the destination image.
21476 The kernel function needs to take the following arguments:
21479 Destination image, @var{__write_only image2d_t}.
21481 This image will become the output; the kernel should write all of it.
21483 Frame index, @var{unsigned int}.
21485 This is a counter starting from zero and increasing by one for each frame.
21487 Source images, @var{__read_only image2d_t}.
21489 These are the most recent images on each input. The kernel may read from
21490 them to generate the output, but they can't be written to.
21497 Copy the input to the output (output must be the same size as the input).
21499 __kernel void copy(__write_only image2d_t destination,
21500 unsigned int index,
21501 __read_only image2d_t source)
21503 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
21505 int2 location = (int2)(get_global_id(0), get_global_id(1));
21507 float4 value = read_imagef(source, sampler, location);
21509 write_imagef(destination, location, value);
21514 Apply a simple transformation, rotating the input by an amount increasing
21515 with the index counter. Pixel values are linearly interpolated by the
21516 sampler, and the output need not have the same dimensions as the input.
21518 __kernel void rotate_image(__write_only image2d_t dst,
21519 unsigned int index,
21520 __read_only image2d_t src)
21522 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21523 CLK_FILTER_LINEAR);
21525 float angle = (float)index / 100.0f;
21527 float2 dst_dim = convert_float2(get_image_dim(dst));
21528 float2 src_dim = convert_float2(get_image_dim(src));
21530 float2 dst_cen = dst_dim / 2.0f;
21531 float2 src_cen = src_dim / 2.0f;
21533 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
21535 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
21537 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
21538 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
21540 src_pos = src_pos * src_dim / dst_dim;
21542 float2 src_loc = src_pos + src_cen;
21544 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
21545 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
21546 write_imagef(dst, dst_loc, 0.5f);
21548 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
21553 Blend two inputs together, with the amount of each input used varying
21554 with the index counter.
21556 __kernel void blend_images(__write_only image2d_t dst,
21557 unsigned int index,
21558 __read_only image2d_t src1,
21559 __read_only image2d_t src2)
21561 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21562 CLK_FILTER_LINEAR);
21564 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
21566 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
21567 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
21568 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
21570 float4 val1 = read_imagef(src1, sampler, src1_loc);
21571 float4 val2 = read_imagef(src2, sampler, src2_loc);
21573 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
21579 @section roberts_opencl
21580 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
21582 The filter accepts the following option:
21586 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21589 Set value which will be multiplied with filtered result.
21590 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21593 Set value which will be added to filtered result.
21594 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21597 @subsection Example
21601 Apply the Roberts cross operator with scale set to 2 and delta set to 10
21603 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
21607 @section sobel_opencl
21609 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
21611 The filter accepts the following option:
21615 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21618 Set value which will be multiplied with filtered result.
21619 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21622 Set value which will be added to filtered result.
21623 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21626 @subsection Example
21630 Apply sobel operator with scale set to 2 and delta set to 10
21632 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
21636 @section tonemap_opencl
21638 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
21640 It accepts the following parameters:
21644 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
21647 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
21650 Apply desaturation for highlights that exceed this level of brightness. The
21651 higher the parameter, the more color information will be preserved. This
21652 setting helps prevent unnaturally blown-out colors for super-highlights, by
21653 (smoothly) turning into white instead. This makes images feel more natural,
21654 at the cost of reducing information about out-of-range colors.
21656 The default value is 0.5, and the algorithm here is a little different from
21657 the cpu version tonemap currently. A setting of 0.0 disables this option.
21660 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
21661 is used to detect whether the scene has changed or not. If the distance between
21662 the current frame average brightness and the current running average exceeds
21663 a threshold value, we would re-calculate scene average and peak brightness.
21664 The default value is 0.2.
21667 Specify the output pixel format.
21669 Currently supported formats are:
21676 Set the output color range.
21678 Possible values are:
21684 Default is same as input.
21687 Set the output color primaries.
21689 Possible values are:
21695 Default is same as input.
21698 Set the output transfer characteristics.
21700 Possible values are:
21709 Set the output colorspace matrix.
21711 Possible value are:
21717 Default is same as input.
21721 @subsection Example
21725 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
21727 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
21731 @section unsharp_opencl
21733 Sharpen or blur the input video.
21735 It accepts the following parameters:
21738 @item luma_msize_x, lx
21739 Set the luma matrix horizontal size.
21740 Range is @code{[1, 23]} and default value is @code{5}.
21742 @item luma_msize_y, ly
21743 Set the luma matrix vertical size.
21744 Range is @code{[1, 23]} and default value is @code{5}.
21746 @item luma_amount, la
21747 Set the luma effect strength.
21748 Range is @code{[-10, 10]} and default value is @code{1.0}.
21750 Negative values will blur the input video, while positive values will
21751 sharpen it, a value of zero will disable the effect.
21753 @item chroma_msize_x, cx
21754 Set the chroma matrix horizontal size.
21755 Range is @code{[1, 23]} and default value is @code{5}.
21757 @item chroma_msize_y, cy
21758 Set the chroma matrix vertical size.
21759 Range is @code{[1, 23]} and default value is @code{5}.
21761 @item chroma_amount, ca
21762 Set the chroma effect strength.
21763 Range is @code{[-10, 10]} and default value is @code{0.0}.
21765 Negative values will blur the input video, while positive values will
21766 sharpen it, a value of zero will disable the effect.
21770 All parameters are optional and default to the equivalent of the
21771 string '5:5:1.0:5:5:0.0'.
21773 @subsection Examples
21777 Apply strong luma sharpen effect:
21779 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
21783 Apply a strong blur of both luma and chroma parameters:
21785 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
21789 @section xfade_opencl
21791 Cross fade two videos with custom transition effect by using OpenCL.
21793 It accepts the following options:
21797 Set one of possible transition effects.
21801 Select custom transition effect, the actual transition description
21802 will be picked from source and kernel options.
21814 Default transition is fade.
21818 OpenCL program source file for custom transition.
21821 Set name of kernel to use for custom transition from program source file.
21824 Set duration of video transition.
21827 Set time of start of transition relative to first video.
21830 The program source file must contain a kernel function with the given name,
21831 which will be run once for each plane of the output. Each run on a plane
21832 gets enqueued as a separate 2D global NDRange with one work-item for each
21833 pixel to be generated. The global ID offset for each work-item is therefore
21834 the coordinates of a pixel in the destination image.
21836 The kernel function needs to take the following arguments:
21839 Destination image, @var{__write_only image2d_t}.
21841 This image will become the output; the kernel should write all of it.
21844 First Source image, @var{__read_only image2d_t}.
21845 Second Source image, @var{__read_only image2d_t}.
21847 These are the most recent images on each input. The kernel may read from
21848 them to generate the output, but they can't be written to.
21851 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
21858 Apply dots curtain transition effect:
21860 __kernel void blend_images(__write_only image2d_t dst,
21861 __read_only image2d_t src1,
21862 __read_only image2d_t src2,
21865 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21866 CLK_FILTER_LINEAR);
21867 int2 p = (int2)(get_global_id(0), get_global_id(1));
21868 float2 rp = (float2)(get_global_id(0), get_global_id(1));
21869 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
21872 float2 dots = (float2)(20.0, 20.0);
21873 float2 center = (float2)(0,0);
21876 float4 val1 = read_imagef(src1, sampler, p);
21877 float4 val2 = read_imagef(src2, sampler, p);
21878 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
21880 write_imagef(dst, p, next ? val1 : val2);
21886 @c man end OPENCL VIDEO FILTERS
21888 @chapter VAAPI Video Filters
21889 @c man begin VAAPI VIDEO FILTERS
21891 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
21893 To enable compilation of these filters you need to configure FFmpeg with
21894 @code{--enable-vaapi}.
21896 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}
21898 @section tonemap_vaapi
21900 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
21901 It maps the dynamic range of HDR10 content to the SDR content.
21902 It currently only accepts HDR10 as input.
21904 It accepts the following parameters:
21908 Specify the output pixel format.
21910 Currently supported formats are:
21919 Set the output color primaries.
21921 Default is same as input.
21924 Set the output transfer characteristics.
21929 Set the output colorspace matrix.
21931 Default is same as input.
21935 @subsection Example
21939 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
21941 tonemap_vaapi=format=p010:t=bt2020-10
21945 @c man end VAAPI VIDEO FILTERS
21947 @chapter Video Sources
21948 @c man begin VIDEO SOURCES
21950 Below is a description of the currently available video sources.
21954 Buffer video frames, and make them available to the filter chain.
21956 This source is mainly intended for a programmatic use, in particular
21957 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
21959 It accepts the following parameters:
21964 Specify the size (width and height) of the buffered video frames. For the
21965 syntax of this option, check the
21966 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21969 The input video width.
21972 The input video height.
21975 A string representing the pixel format of the buffered video frames.
21976 It may be a number corresponding to a pixel format, or a pixel format
21980 Specify the timebase assumed by the timestamps of the buffered frames.
21983 Specify the frame rate expected for the video stream.
21985 @item pixel_aspect, sar
21986 The sample (pixel) aspect ratio of the input video.
21989 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
21990 to the filtergraph description to specify swscale flags for automatically
21991 inserted scalers. See @ref{Filtergraph syntax}.
21993 @item hw_frames_ctx
21994 When using a hardware pixel format, this should be a reference to an
21995 AVHWFramesContext describing input frames.
22000 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
22003 will instruct the source to accept video frames with size 320x240 and
22004 with format "yuv410p", assuming 1/24 as the timestamps timebase and
22005 square pixels (1:1 sample aspect ratio).
22006 Since the pixel format with name "yuv410p" corresponds to the number 6
22007 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
22008 this example corresponds to:
22010 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
22013 Alternatively, the options can be specified as a flat string, but this
22014 syntax is deprecated:
22016 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
22020 Create a pattern generated by an elementary cellular automaton.
22022 The initial state of the cellular automaton can be defined through the
22023 @option{filename} and @option{pattern} options. If such options are
22024 not specified an initial state is created randomly.
22026 At each new frame a new row in the video is filled with the result of
22027 the cellular automaton next generation. The behavior when the whole
22028 frame is filled is defined by the @option{scroll} option.
22030 This source accepts the following options:
22034 Read the initial cellular automaton state, i.e. the starting row, from
22035 the specified file.
22036 In the file, each non-whitespace character is considered an alive
22037 cell, a newline will terminate the row, and further characters in the
22038 file will be ignored.
22041 Read the initial cellular automaton state, i.e. the starting row, from
22042 the specified string.
22044 Each non-whitespace character in the string is considered an alive
22045 cell, a newline will terminate the row, and further characters in the
22046 string will be ignored.
22049 Set the video rate, that is the number of frames generated per second.
22052 @item random_fill_ratio, ratio
22053 Set the random fill ratio for the initial cellular automaton row. It
22054 is a floating point number value ranging from 0 to 1, defaults to
22057 This option is ignored when a file or a pattern is specified.
22059 @item random_seed, seed
22060 Set the seed for filling randomly the initial row, must be an integer
22061 included between 0 and UINT32_MAX. If not specified, or if explicitly
22062 set to -1, the filter will try to use a good random seed on a best
22066 Set the cellular automaton rule, it is a number ranging from 0 to 255.
22067 Default value is 110.
22070 Set the size of the output video. For the syntax of this option, check the
22071 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22073 If @option{filename} or @option{pattern} is specified, the size is set
22074 by default to the width of the specified initial state row, and the
22075 height is set to @var{width} * PHI.
22077 If @option{size} is set, it must contain the width of the specified
22078 pattern string, and the specified pattern will be centered in the
22081 If a filename or a pattern string is not specified, the size value
22082 defaults to "320x518" (used for a randomly generated initial state).
22085 If set to 1, scroll the output upward when all the rows in the output
22086 have been already filled. If set to 0, the new generated row will be
22087 written over the top row just after the bottom row is filled.
22090 @item start_full, full
22091 If set to 1, completely fill the output with generated rows before
22092 outputting the first frame.
22093 This is the default behavior, for disabling set the value to 0.
22096 If set to 1, stitch the left and right row edges together.
22097 This is the default behavior, for disabling set the value to 0.
22100 @subsection Examples
22104 Read the initial state from @file{pattern}, and specify an output of
22107 cellauto=f=pattern:s=200x400
22111 Generate a random initial row with a width of 200 cells, with a fill
22114 cellauto=ratio=2/3:s=200x200
22118 Create a pattern generated by rule 18 starting by a single alive cell
22119 centered on an initial row with width 100:
22121 cellauto=p=@@:s=100x400:full=0:rule=18
22125 Specify a more elaborated initial pattern:
22127 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
22132 @anchor{coreimagesrc}
22133 @section coreimagesrc
22134 Video source generated on GPU using Apple's CoreImage API on OSX.
22136 This video source is a specialized version of the @ref{coreimage} video filter.
22137 Use a core image generator at the beginning of the applied filterchain to
22138 generate the content.
22140 The coreimagesrc video source accepts the following options:
22142 @item list_generators
22143 List all available generators along with all their respective options as well as
22144 possible minimum and maximum values along with the default values.
22146 list_generators=true
22150 Specify the size of the sourced video. For the syntax of this option, check the
22151 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22152 The default value is @code{320x240}.
22155 Specify the frame rate of the sourced video, as the number of frames
22156 generated per second. It has to be a string in the format
22157 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22158 number or a valid video frame rate abbreviation. The default value is
22162 Set the sample aspect ratio of the sourced video.
22165 Set the duration of the sourced video. See
22166 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22167 for the accepted syntax.
22169 If not specified, or the expressed duration is negative, the video is
22170 supposed to be generated forever.
22173 Additionally, all options of the @ref{coreimage} video filter are accepted.
22174 A complete filterchain can be used for further processing of the
22175 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
22176 and examples for details.
22178 @subsection Examples
22183 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
22184 given as complete and escaped command-line for Apple's standard bash shell:
22186 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
22188 This example is equivalent to the QRCode example of @ref{coreimage} without the
22189 need for a nullsrc video source.
22193 @section mandelbrot
22195 Generate a Mandelbrot set fractal, and progressively zoom towards the
22196 point specified with @var{start_x} and @var{start_y}.
22198 This source accepts the following options:
22203 Set the terminal pts value. Default value is 400.
22206 Set the terminal scale value.
22207 Must be a floating point value. Default value is 0.3.
22210 Set the inner coloring mode, that is the algorithm used to draw the
22211 Mandelbrot fractal internal region.
22213 It shall assume one of the following values:
22218 Show time until convergence.
22220 Set color based on point closest to the origin of the iterations.
22225 Default value is @var{mincol}.
22228 Set the bailout value. Default value is 10.0.
22231 Set the maximum of iterations performed by the rendering
22232 algorithm. Default value is 7189.
22235 Set outer coloring mode.
22236 It shall assume one of following values:
22238 @item iteration_count
22239 Set iteration count mode.
22240 @item normalized_iteration_count
22241 set normalized iteration count mode.
22243 Default value is @var{normalized_iteration_count}.
22246 Set frame rate, expressed as number of frames per second. Default
22250 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22251 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22254 Set the initial scale value. Default value is 3.0.
22257 Set the initial x position. Must be a floating point value between
22258 -100 and 100. Default value is -0.743643887037158704752191506114774.
22261 Set the initial y position. Must be a floating point value between
22262 -100 and 100. Default value is -0.131825904205311970493132056385139.
22267 Generate various test patterns, as generated by the MPlayer test filter.
22269 The size of the generated video is fixed, and is 256x256.
22270 This source is useful in particular for testing encoding features.
22272 This source accepts the following options:
22277 Specify the frame rate of the sourced video, as the number of frames
22278 generated per second. It has to be a string in the format
22279 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22280 number or a valid video frame rate abbreviation. The default value is
22284 Set the duration of the sourced video. See
22285 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22286 for the accepted syntax.
22288 If not specified, or the expressed duration is negative, the video is
22289 supposed to be generated forever.
22293 Set the number or the name of the test to perform. Supported tests are:
22307 @item max_frames, m
22308 Set the maximum number of frames generated for each test, default value is 30.
22312 Default value is "all", which will cycle through the list of all tests.
22317 mptestsrc=t=dc_luma
22320 will generate a "dc_luma" test pattern.
22322 @section frei0r_src
22324 Provide a frei0r source.
22326 To enable compilation of this filter you need to install the frei0r
22327 header and configure FFmpeg with @code{--enable-frei0r}.
22329 This source accepts the following parameters:
22334 The size of the video to generate. For the syntax of this option, check the
22335 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22338 The framerate of the generated video. It may be a string of the form
22339 @var{num}/@var{den} or a frame rate abbreviation.
22342 The name to the frei0r source to load. For more information regarding frei0r and
22343 how to set the parameters, read the @ref{frei0r} section in the video filters
22346 @item filter_params
22347 A '|'-separated list of parameters to pass to the frei0r source.
22351 For example, to generate a frei0r partik0l source with size 200x200
22352 and frame rate 10 which is overlaid on the overlay filter main input:
22354 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
22359 Generate a life pattern.
22361 This source is based on a generalization of John Conway's life game.
22363 The sourced input represents a life grid, each pixel represents a cell
22364 which can be in one of two possible states, alive or dead. Every cell
22365 interacts with its eight neighbours, which are the cells that are
22366 horizontally, vertically, or diagonally adjacent.
22368 At each interaction the grid evolves according to the adopted rule,
22369 which specifies the number of neighbor alive cells which will make a
22370 cell stay alive or born. The @option{rule} option allows one to specify
22373 This source accepts the following options:
22377 Set the file from which to read the initial grid state. In the file,
22378 each non-whitespace character is considered an alive cell, and newline
22379 is used to delimit the end of each row.
22381 If this option is not specified, the initial grid is generated
22385 Set the video rate, that is the number of frames generated per second.
22388 @item random_fill_ratio, ratio
22389 Set the random fill ratio for the initial random grid. It is a
22390 floating point number value ranging from 0 to 1, defaults to 1/PHI.
22391 It is ignored when a file is specified.
22393 @item random_seed, seed
22394 Set the seed for filling the initial random grid, must be an integer
22395 included between 0 and UINT32_MAX. If not specified, or if explicitly
22396 set to -1, the filter will try to use a good random seed on a best
22402 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
22403 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
22404 @var{NS} specifies the number of alive neighbor cells which make a
22405 live cell stay alive, and @var{NB} the number of alive neighbor cells
22406 which make a dead cell to become alive (i.e. to "born").
22407 "s" and "b" can be used in place of "S" and "B", respectively.
22409 Alternatively a rule can be specified by an 18-bits integer. The 9
22410 high order bits are used to encode the next cell state if it is alive
22411 for each number of neighbor alive cells, the low order bits specify
22412 the rule for "borning" new cells. Higher order bits encode for an
22413 higher number of neighbor cells.
22414 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
22415 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
22417 Default value is "S23/B3", which is the original Conway's game of life
22418 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
22419 cells, and will born a new cell if there are three alive cells around
22423 Set the size of the output video. For the syntax of this option, check the
22424 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22426 If @option{filename} is specified, the size is set by default to the
22427 same size of the input file. If @option{size} is set, it must contain
22428 the size specified in the input file, and the initial grid defined in
22429 that file is centered in the larger resulting area.
22431 If a filename is not specified, the size value defaults to "320x240"
22432 (used for a randomly generated initial grid).
22435 If set to 1, stitch the left and right grid edges together, and the
22436 top and bottom edges also. Defaults to 1.
22439 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
22440 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
22441 value from 0 to 255.
22444 Set the color of living (or new born) cells.
22447 Set the color of dead cells. If @option{mold} is set, this is the first color
22448 used to represent a dead cell.
22451 Set mold color, for definitely dead and moldy cells.
22453 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
22454 ffmpeg-utils manual,ffmpeg-utils}.
22457 @subsection Examples
22461 Read a grid from @file{pattern}, and center it on a grid of size
22464 life=f=pattern:s=300x300
22468 Generate a random grid of size 200x200, with a fill ratio of 2/3:
22470 life=ratio=2/3:s=200x200
22474 Specify a custom rule for evolving a randomly generated grid:
22480 Full example with slow death effect (mold) using @command{ffplay}:
22482 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
22489 @anchor{haldclutsrc}
22492 @anchor{pal100bars}
22493 @anchor{rgbtestsrc}
22495 @anchor{smptehdbars}
22498 @anchor{yuvtestsrc}
22499 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
22501 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
22503 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
22505 The @code{color} source provides an uniformly colored input.
22507 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
22508 @ref{haldclut} filter.
22510 The @code{nullsrc} source returns unprocessed video frames. It is
22511 mainly useful to be employed in analysis / debugging tools, or as the
22512 source for filters which ignore the input data.
22514 The @code{pal75bars} source generates a color bars pattern, based on
22515 EBU PAL recommendations with 75% color levels.
22517 The @code{pal100bars} source generates a color bars pattern, based on
22518 EBU PAL recommendations with 100% color levels.
22520 The @code{rgbtestsrc} source generates an RGB test pattern useful for
22521 detecting RGB vs BGR issues. You should see a red, green and blue
22522 stripe from top to bottom.
22524 The @code{smptebars} source generates a color bars pattern, based on
22525 the SMPTE Engineering Guideline EG 1-1990.
22527 The @code{smptehdbars} source generates a color bars pattern, based on
22528 the SMPTE RP 219-2002.
22530 The @code{testsrc} source generates a test video pattern, showing a
22531 color pattern, a scrolling gradient and a timestamp. This is mainly
22532 intended for testing purposes.
22534 The @code{testsrc2} source is similar to testsrc, but supports more
22535 pixel formats instead of just @code{rgb24}. This allows using it as an
22536 input for other tests without requiring a format conversion.
22538 The @code{yuvtestsrc} source generates an YUV test pattern. You should
22539 see a y, cb and cr stripe from top to bottom.
22541 The sources accept the following parameters:
22546 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
22547 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
22548 pixels to be used as identity matrix for 3D lookup tables. Each component is
22549 coded on a @code{1/(N*N)} scale.
22552 Specify the color of the source, only available in the @code{color}
22553 source. For the syntax of this option, check the
22554 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
22557 Specify the size of the sourced video. For the syntax of this option, check the
22558 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22559 The default value is @code{320x240}.
22561 This option is not available with the @code{allrgb}, @code{allyuv}, and
22562 @code{haldclutsrc} filters.
22565 Specify the frame rate of the sourced video, as the number of frames
22566 generated per second. It has to be a string in the format
22567 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22568 number or a valid video frame rate abbreviation. The default value is
22572 Set the duration of the sourced video. See
22573 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22574 for the accepted syntax.
22576 If not specified, or the expressed duration is negative, the video is
22577 supposed to be generated forever.
22580 Set the sample aspect ratio of the sourced video.
22583 Specify the alpha (opacity) of the background, only available in the
22584 @code{testsrc2} source. The value must be between 0 (fully transparent) and
22585 255 (fully opaque, the default).
22588 Set the number of decimals to show in the timestamp, only available in the
22589 @code{testsrc} source.
22591 The displayed timestamp value will correspond to the original
22592 timestamp value multiplied by the power of 10 of the specified
22593 value. Default value is 0.
22596 @subsection Examples
22600 Generate a video with a duration of 5.3 seconds, with size
22601 176x144 and a frame rate of 10 frames per second:
22603 testsrc=duration=5.3:size=qcif:rate=10
22607 The following graph description will generate a red source
22608 with an opacity of 0.2, with size "qcif" and a frame rate of 10
22611 color=c=red@@0.2:s=qcif:r=10
22615 If the input content is to be ignored, @code{nullsrc} can be used. The
22616 following command generates noise in the luminance plane by employing
22617 the @code{geq} filter:
22619 nullsrc=s=256x256, geq=random(1)*255:128:128
22623 @subsection Commands
22625 The @code{color} source supports the following commands:
22629 Set the color of the created image. Accepts the same syntax of the
22630 corresponding @option{color} option.
22635 Generate video using an OpenCL program.
22640 OpenCL program source file.
22643 Kernel name in program.
22646 Size of frames to generate. This must be set.
22649 Pixel format to use for the generated frames. This must be set.
22652 Number of frames generated every second. Default value is '25'.
22656 For details of how the program loading works, see the @ref{program_opencl}
22663 Generate a colour ramp by setting pixel values from the position of the pixel
22664 in the output image. (Note that this will work with all pixel formats, but
22665 the generated output will not be the same.)
22667 __kernel void ramp(__write_only image2d_t dst,
22668 unsigned int index)
22670 int2 loc = (int2)(get_global_id(0), get_global_id(1));
22673 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
22675 write_imagef(dst, loc, val);
22680 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
22682 __kernel void sierpinski_carpet(__write_only image2d_t dst,
22683 unsigned int index)
22685 int2 loc = (int2)(get_global_id(0), get_global_id(1));
22687 float4 value = 0.0f;
22688 int x = loc.x + index;
22689 int y = loc.y + index;
22690 while (x > 0 || y > 0) {
22691 if (x % 3 == 1 && y % 3 == 1) {
22699 write_imagef(dst, loc, value);
22705 @section sierpinski
22707 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
22709 This source accepts the following options:
22713 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22714 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22717 Set frame rate, expressed as number of frames per second. Default
22721 Set seed which is used for random panning.
22724 Set max jump for single pan destination. Allowed range is from 1 to 10000.
22727 Set fractal type, can be default @code{carpet} or @code{triangle}.
22730 @c man end VIDEO SOURCES
22732 @chapter Video Sinks
22733 @c man begin VIDEO SINKS
22735 Below is a description of the currently available video sinks.
22737 @section buffersink
22739 Buffer video frames, and make them available to the end of the filter
22742 This sink is mainly intended for programmatic use, in particular
22743 through the interface defined in @file{libavfilter/buffersink.h}
22744 or the options system.
22746 It accepts a pointer to an AVBufferSinkContext structure, which
22747 defines the incoming buffers' formats, to be passed as the opaque
22748 parameter to @code{avfilter_init_filter} for initialization.
22752 Null video sink: do absolutely nothing with the input video. It is
22753 mainly useful as a template and for use in analysis / debugging
22756 @c man end VIDEO SINKS
22758 @chapter Multimedia Filters
22759 @c man begin MULTIMEDIA FILTERS
22761 Below is a description of the currently available multimedia filters.
22765 Convert input audio to a video output, displaying the audio bit scope.
22767 The filter accepts the following options:
22771 Set frame rate, expressed as number of frames per second. Default
22775 Specify the video size for the output. For the syntax of this option, check the
22776 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22777 Default value is @code{1024x256}.
22780 Specify list of colors separated by space or by '|' which will be used to
22781 draw channels. Unrecognized or missing colors will be replaced
22785 @section adrawgraph
22786 Draw a graph using input audio metadata.
22788 See @ref{drawgraph}
22790 @section agraphmonitor
22792 See @ref{graphmonitor}.
22794 @section ahistogram
22796 Convert input audio to a video output, displaying the volume histogram.
22798 The filter accepts the following options:
22802 Specify how histogram is calculated.
22804 It accepts the following values:
22807 Use single histogram for all channels.
22809 Use separate histogram for each channel.
22811 Default is @code{single}.
22814 Set frame rate, expressed as number of frames per second. Default
22818 Specify the video size for the output. For the syntax of this option, check the
22819 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22820 Default value is @code{hd720}.
22825 It accepts the following values:
22836 reverse logarithmic
22838 Default is @code{log}.
22841 Set amplitude scale.
22843 It accepts the following values:
22850 Default is @code{log}.
22853 Set how much frames to accumulate in histogram.
22854 Default is 1. Setting this to -1 accumulates all frames.
22857 Set histogram ratio of window height.
22860 Set sonogram sliding.
22862 It accepts the following values:
22865 replace old rows with new ones.
22867 scroll from top to bottom.
22869 Default is @code{replace}.
22872 @section aphasemeter
22874 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
22875 representing mean phase of current audio frame. A video output can also be produced and is
22876 enabled by default. The audio is passed through as first output.
22878 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
22879 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
22880 and @code{1} means channels are in phase.
22882 The filter accepts the following options, all related to its video output:
22886 Set the output frame rate. Default value is @code{25}.
22889 Set the video size for the output. For the syntax of this option, check the
22890 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22891 Default value is @code{800x400}.
22896 Specify the red, green, blue contrast. Default values are @code{2},
22897 @code{7} and @code{1}.
22898 Allowed range is @code{[0, 255]}.
22901 Set color which will be used for drawing median phase. If color is
22902 @code{none} which is default, no median phase value will be drawn.
22905 Enable video output. Default is enabled.
22908 @section avectorscope
22910 Convert input audio to a video output, representing the audio vector
22913 The filter is used to measure the difference between channels of stereo
22914 audio stream. A monaural signal, consisting of identical left and right
22915 signal, results in straight vertical line. Any stereo separation is visible
22916 as a deviation from this line, creating a Lissajous figure.
22917 If the straight (or deviation from it) but horizontal line appears this
22918 indicates that the left and right channels are out of phase.
22920 The filter accepts the following options:
22924 Set the vectorscope mode.
22926 Available values are:
22929 Lissajous rotated by 45 degrees.
22932 Same as above but not rotated.
22935 Shape resembling half of circle.
22938 Default value is @samp{lissajous}.
22941 Set the video size for the output. For the syntax of this option, check the
22942 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22943 Default value is @code{400x400}.
22946 Set the output frame rate. Default value is @code{25}.
22952 Specify the red, green, blue and alpha contrast. Default values are @code{40},
22953 @code{160}, @code{80} and @code{255}.
22954 Allowed range is @code{[0, 255]}.
22960 Specify the red, green, blue and alpha fade. Default values are @code{15},
22961 @code{10}, @code{5} and @code{5}.
22962 Allowed range is @code{[0, 255]}.
22965 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
22966 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
22969 Set the vectorscope drawing mode.
22971 Available values are:
22974 Draw dot for each sample.
22977 Draw line between previous and current sample.
22980 Default value is @samp{dot}.
22983 Specify amplitude scale of audio samples.
22985 Available values are:
23001 Swap left channel axis with right channel axis.
23011 Mirror only x axis.
23014 Mirror only y axis.
23022 @subsection Examples
23026 Complete example using @command{ffplay}:
23028 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
23029 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
23033 @section bench, abench
23035 Benchmark part of a filtergraph.
23037 The filter accepts the following options:
23041 Start or stop a timer.
23043 Available values are:
23046 Get the current time, set it as frame metadata (using the key
23047 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
23050 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
23051 the input frame metadata to get the time difference. Time difference, average,
23052 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
23053 @code{min}) are then printed. The timestamps are expressed in seconds.
23057 @subsection Examples
23061 Benchmark @ref{selectivecolor} filter:
23063 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
23069 Concatenate audio and video streams, joining them together one after the
23072 The filter works on segments of synchronized video and audio streams. All
23073 segments must have the same number of streams of each type, and that will
23074 also be the number of streams at output.
23076 The filter accepts the following options:
23081 Set the number of segments. Default is 2.
23084 Set the number of output video streams, that is also the number of video
23085 streams in each segment. Default is 1.
23088 Set the number of output audio streams, that is also the number of audio
23089 streams in each segment. Default is 0.
23092 Activate unsafe mode: do not fail if segments have a different format.
23096 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
23097 @var{a} audio outputs.
23099 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
23100 segment, in the same order as the outputs, then the inputs for the second
23103 Related streams do not always have exactly the same duration, for various
23104 reasons including codec frame size or sloppy authoring. For that reason,
23105 related synchronized streams (e.g. a video and its audio track) should be
23106 concatenated at once. The concat filter will use the duration of the longest
23107 stream in each segment (except the last one), and if necessary pad shorter
23108 audio streams with silence.
23110 For this filter to work correctly, all segments must start at timestamp 0.
23112 All corresponding streams must have the same parameters in all segments; the
23113 filtering system will automatically select a common pixel format for video
23114 streams, and a common sample format, sample rate and channel layout for
23115 audio streams, but other settings, such as resolution, must be converted
23116 explicitly by the user.
23118 Different frame rates are acceptable but will result in variable frame rate
23119 at output; be sure to configure the output file to handle it.
23121 @subsection Examples
23125 Concatenate an opening, an episode and an ending, all in bilingual version
23126 (video in stream 0, audio in streams 1 and 2):
23128 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
23129 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
23130 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
23131 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
23135 Concatenate two parts, handling audio and video separately, using the
23136 (a)movie sources, and adjusting the resolution:
23138 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
23139 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
23140 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
23142 Note that a desync will happen at the stitch if the audio and video streams
23143 do not have exactly the same duration in the first file.
23147 @subsection Commands
23149 This filter supports the following commands:
23152 Close the current segment and step to the next one
23158 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
23159 level. By default, it logs a message at a frequency of 10Hz with the
23160 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
23161 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
23163 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
23164 sample format is double-precision floating point. The input stream will be converted to
23165 this specification, if needed. Users may need to insert aformat and/or aresample filters
23166 after this filter to obtain the original parameters.
23168 The filter also has a video output (see the @var{video} option) with a real
23169 time graph to observe the loudness evolution. The graphic contains the logged
23170 message mentioned above, so it is not printed anymore when this option is set,
23171 unless the verbose logging is set. The main graphing area contains the
23172 short-term loudness (3 seconds of analysis), and the gauge on the right is for
23173 the momentary loudness (400 milliseconds), but can optionally be configured
23174 to instead display short-term loudness (see @var{gauge}).
23176 The green area marks a +/- 1LU target range around the target loudness
23177 (-23LUFS by default, unless modified through @var{target}).
23179 More information about the Loudness Recommendation EBU R128 on
23180 @url{http://tech.ebu.ch/loudness}.
23182 The filter accepts the following options:
23187 Activate the video output. The audio stream is passed unchanged whether this
23188 option is set or no. The video stream will be the first output stream if
23189 activated. Default is @code{0}.
23192 Set the video size. This option is for video only. For the syntax of this
23194 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23195 Default and minimum resolution is @code{640x480}.
23198 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
23199 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
23200 other integer value between this range is allowed.
23203 Set metadata injection. If set to @code{1}, the audio input will be segmented
23204 into 100ms output frames, each of them containing various loudness information
23205 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
23207 Default is @code{0}.
23210 Force the frame logging level.
23212 Available values are:
23215 information logging level
23217 verbose logging level
23220 By default, the logging level is set to @var{info}. If the @option{video} or
23221 the @option{metadata} options are set, it switches to @var{verbose}.
23226 Available modes can be cumulated (the option is a @code{flag} type). Possible
23230 Disable any peak mode (default).
23232 Enable sample-peak mode.
23234 Simple peak mode looking for the higher sample value. It logs a message
23235 for sample-peak (identified by @code{SPK}).
23237 Enable true-peak mode.
23239 If enabled, the peak lookup is done on an over-sampled version of the input
23240 stream for better peak accuracy. It logs a message for true-peak.
23241 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
23242 This mode requires a build with @code{libswresample}.
23246 Treat mono input files as "dual mono". If a mono file is intended for playback
23247 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
23248 If set to @code{true}, this option will compensate for this effect.
23249 Multi-channel input files are not affected by this option.
23252 Set a specific pan law to be used for the measurement of dual mono files.
23253 This parameter is optional, and has a default value of -3.01dB.
23256 Set a specific target level (in LUFS) used as relative zero in the visualization.
23257 This parameter is optional and has a default value of -23LUFS as specified
23258 by EBU R128. However, material published online may prefer a level of -16LUFS
23259 (e.g. for use with podcasts or video platforms).
23262 Set the value displayed by the gauge. Valid values are @code{momentary} and s
23263 @code{shortterm}. By default the momentary value will be used, but in certain
23264 scenarios it may be more useful to observe the short term value instead (e.g.
23268 Sets the display scale for the loudness. Valid parameters are @code{absolute}
23269 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
23270 video output, not the summary or continuous log output.
23273 @subsection Examples
23277 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
23279 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
23283 Run an analysis with @command{ffmpeg}:
23285 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
23289 @section interleave, ainterleave
23291 Temporally interleave frames from several inputs.
23293 @code{interleave} works with video inputs, @code{ainterleave} with audio.
23295 These filters read frames from several inputs and send the oldest
23296 queued frame to the output.
23298 Input streams must have well defined, monotonically increasing frame
23301 In order to submit one frame to output, these filters need to enqueue
23302 at least one frame for each input, so they cannot work in case one
23303 input is not yet terminated and will not receive incoming frames.
23305 For example consider the case when one input is a @code{select} filter
23306 which always drops input frames. The @code{interleave} filter will keep
23307 reading from that input, but it will never be able to send new frames
23308 to output until the input sends an end-of-stream signal.
23310 Also, depending on inputs synchronization, the filters will drop
23311 frames in case one input receives more frames than the other ones, and
23312 the queue is already filled.
23314 These filters accept the following options:
23318 Set the number of different inputs, it is 2 by default.
23321 How to determine the end-of-stream.
23325 The duration of the longest input. (default)
23328 The duration of the shortest input.
23331 The duration of the first input.
23336 @subsection Examples
23340 Interleave frames belonging to different streams using @command{ffmpeg}:
23342 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
23346 Add flickering blur effect:
23348 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
23352 @section metadata, ametadata
23354 Manipulate frame metadata.
23356 This filter accepts the following options:
23360 Set mode of operation of the filter.
23362 Can be one of the following:
23366 If both @code{value} and @code{key} is set, select frames
23367 which have such metadata. If only @code{key} is set, select
23368 every frame that has such key in metadata.
23371 Add new metadata @code{key} and @code{value}. If key is already available
23375 Modify value of already present key.
23378 If @code{value} is set, delete only keys that have such value.
23379 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
23383 Print key and its value if metadata was found. If @code{key} is not set print all
23384 metadata values available in frame.
23388 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
23391 Set metadata value which will be used. This option is mandatory for
23392 @code{modify} and @code{add} mode.
23395 Which function to use when comparing metadata value and @code{value}.
23397 Can be one of following:
23401 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
23404 Values are interpreted as strings, returns true if metadata value starts with
23405 the @code{value} option string.
23408 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
23411 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
23414 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
23417 Values are interpreted as floats, returns true if expression from option @code{expr}
23421 Values are interpreted as strings, returns true if metadata value ends with
23422 the @code{value} option string.
23426 Set expression which is used when @code{function} is set to @code{expr}.
23427 The expression is evaluated through the eval API and can contain the following
23432 Float representation of @code{value} from metadata key.
23435 Float representation of @code{value} as supplied by user in @code{value} option.
23439 If specified in @code{print} mode, output is written to the named file. Instead of
23440 plain filename any writable url can be specified. Filename ``-'' is a shorthand
23441 for standard output. If @code{file} option is not set, output is written to the log
23442 with AV_LOG_INFO loglevel.
23445 Reduces buffering in print mode when output is written to a URL set using @var{file}.
23449 @subsection Examples
23453 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
23456 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
23459 Print silencedetect output to file @file{metadata.txt}.
23461 silencedetect,ametadata=mode=print:file=metadata.txt
23464 Direct all metadata to a pipe with file descriptor 4.
23466 metadata=mode=print:file='pipe\:4'
23470 @section perms, aperms
23472 Set read/write permissions for the output frames.
23474 These filters are mainly aimed at developers to test direct path in the
23475 following filter in the filtergraph.
23477 The filters accept the following options:
23481 Select the permissions mode.
23483 It accepts the following values:
23486 Do nothing. This is the default.
23488 Set all the output frames read-only.
23490 Set all the output frames directly writable.
23492 Make the frame read-only if writable, and writable if read-only.
23494 Set each output frame read-only or writable randomly.
23498 Set the seed for the @var{random} mode, must be an integer included between
23499 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
23500 @code{-1}, the filter will try to use a good random seed on a best effort
23504 Note: in case of auto-inserted filter between the permission filter and the
23505 following one, the permission might not be received as expected in that
23506 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
23507 perms/aperms filter can avoid this problem.
23509 @section realtime, arealtime
23511 Slow down filtering to match real time approximately.
23513 These filters will pause the filtering for a variable amount of time to
23514 match the output rate with the input timestamps.
23515 They are similar to the @option{re} option to @code{ffmpeg}.
23517 They accept the following options:
23521 Time limit for the pauses. Any pause longer than that will be considered
23522 a timestamp discontinuity and reset the timer. Default is 2 seconds.
23524 Speed factor for processing. The value must be a float larger than zero.
23525 Values larger than 1.0 will result in faster than realtime processing,
23526 smaller will slow processing down. The @var{limit} is automatically adapted
23527 accordingly. Default is 1.0.
23529 A processing speed faster than what is possible without these filters cannot
23534 @section select, aselect
23536 Select frames to pass in output.
23538 This filter accepts the following options:
23543 Set expression, which is evaluated for each input frame.
23545 If the expression is evaluated to zero, the frame is discarded.
23547 If the evaluation result is negative or NaN, the frame is sent to the
23548 first output; otherwise it is sent to the output with index
23549 @code{ceil(val)-1}, assuming that the input index starts from 0.
23551 For example a value of @code{1.2} corresponds to the output with index
23552 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
23555 Set the number of outputs. The output to which to send the selected
23556 frame is based on the result of the evaluation. Default value is 1.
23559 The expression can contain the following constants:
23563 The (sequential) number of the filtered frame, starting from 0.
23566 The (sequential) number of the selected frame, starting from 0.
23568 @item prev_selected_n
23569 The sequential number of the last selected frame. It's NAN if undefined.
23572 The timebase of the input timestamps.
23575 The PTS (Presentation TimeStamp) of the filtered video frame,
23576 expressed in @var{TB} units. It's NAN if undefined.
23579 The PTS of the filtered video frame,
23580 expressed in seconds. It's NAN if undefined.
23583 The PTS of the previously filtered video frame. It's NAN if undefined.
23585 @item prev_selected_pts
23586 The PTS of the last previously filtered video frame. It's NAN if undefined.
23588 @item prev_selected_t
23589 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
23592 The PTS of the first video frame in the video. It's NAN if undefined.
23595 The time of the first video frame in the video. It's NAN if undefined.
23597 @item pict_type @emph{(video only)}
23598 The type of the filtered frame. It can assume one of the following
23610 @item interlace_type @emph{(video only)}
23611 The frame interlace type. It can assume one of the following values:
23614 The frame is progressive (not interlaced).
23616 The frame is top-field-first.
23618 The frame is bottom-field-first.
23621 @item consumed_sample_n @emph{(audio only)}
23622 the number of selected samples before the current frame
23624 @item samples_n @emph{(audio only)}
23625 the number of samples in the current frame
23627 @item sample_rate @emph{(audio only)}
23628 the input sample rate
23631 This is 1 if the filtered frame is a key-frame, 0 otherwise.
23634 the position in the file of the filtered frame, -1 if the information
23635 is not available (e.g. for synthetic video)
23637 @item scene @emph{(video only)}
23638 value between 0 and 1 to indicate a new scene; a low value reflects a low
23639 probability for the current frame to introduce a new scene, while a higher
23640 value means the current frame is more likely to be one (see the example below)
23642 @item concatdec_select
23643 The concat demuxer can select only part of a concat input file by setting an
23644 inpoint and an outpoint, but the output packets may not be entirely contained
23645 in the selected interval. By using this variable, it is possible to skip frames
23646 generated by the concat demuxer which are not exactly contained in the selected
23649 This works by comparing the frame pts against the @var{lavf.concat.start_time}
23650 and the @var{lavf.concat.duration} packet metadata values which are also
23651 present in the decoded frames.
23653 The @var{concatdec_select} variable is -1 if the frame pts is at least
23654 start_time and either the duration metadata is missing or the frame pts is less
23655 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
23658 That basically means that an input frame is selected if its pts is within the
23659 interval set by the concat demuxer.
23663 The default value of the select expression is "1".
23665 @subsection Examples
23669 Select all frames in input:
23674 The example above is the same as:
23686 Select only I-frames:
23688 select='eq(pict_type\,I)'
23692 Select one frame every 100:
23694 select='not(mod(n\,100))'
23698 Select only frames contained in the 10-20 time interval:
23700 select=between(t\,10\,20)
23704 Select only I-frames contained in the 10-20 time interval:
23706 select=between(t\,10\,20)*eq(pict_type\,I)
23710 Select frames with a minimum distance of 10 seconds:
23712 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
23716 Use aselect to select only audio frames with samples number > 100:
23718 aselect='gt(samples_n\,100)'
23722 Create a mosaic of the first scenes:
23724 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
23727 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
23731 Send even and odd frames to separate outputs, and compose them:
23733 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
23737 Select useful frames from an ffconcat file which is using inpoints and
23738 outpoints but where the source files are not intra frame only.
23740 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
23744 @section sendcmd, asendcmd
23746 Send commands to filters in the filtergraph.
23748 These filters read commands to be sent to other filters in the
23751 @code{sendcmd} must be inserted between two video filters,
23752 @code{asendcmd} must be inserted between two audio filters, but apart
23753 from that they act the same way.
23755 The specification of commands can be provided in the filter arguments
23756 with the @var{commands} option, or in a file specified by the
23757 @var{filename} option.
23759 These filters accept the following options:
23762 Set the commands to be read and sent to the other filters.
23764 Set the filename of the commands to be read and sent to the other
23768 @subsection Commands syntax
23770 A commands description consists of a sequence of interval
23771 specifications, comprising a list of commands to be executed when a
23772 particular event related to that interval occurs. The occurring event
23773 is typically the current frame time entering or leaving a given time
23776 An interval is specified by the following syntax:
23778 @var{START}[-@var{END}] @var{COMMANDS};
23781 The time interval is specified by the @var{START} and @var{END} times.
23782 @var{END} is optional and defaults to the maximum time.
23784 The current frame time is considered within the specified interval if
23785 it is included in the interval [@var{START}, @var{END}), that is when
23786 the time is greater or equal to @var{START} and is lesser than
23789 @var{COMMANDS} consists of a sequence of one or more command
23790 specifications, separated by ",", relating to that interval. The
23791 syntax of a command specification is given by:
23793 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
23796 @var{FLAGS} is optional and specifies the type of events relating to
23797 the time interval which enable sending the specified command, and must
23798 be a non-null sequence of identifier flags separated by "+" or "|" and
23799 enclosed between "[" and "]".
23801 The following flags are recognized:
23804 The command is sent when the current frame timestamp enters the
23805 specified interval. In other words, the command is sent when the
23806 previous frame timestamp was not in the given interval, and the
23810 The command is sent when the current frame timestamp leaves the
23811 specified interval. In other words, the command is sent when the
23812 previous frame timestamp was in the given interval, and the
23816 The command @var{ARG} is interpreted as expression and result of
23817 expression is passed as @var{ARG}.
23819 The expression is evaluated through the eval API and can contain the following
23824 Original position in the file of the frame, or undefined if undefined
23825 for the current frame.
23828 The presentation timestamp in input.
23831 The count of the input frame for video or audio, starting from 0.
23834 The time in seconds of the current frame.
23837 The start time in seconds of the current command interval.
23840 The end time in seconds of the current command interval.
23843 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
23848 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
23851 @var{TARGET} specifies the target of the command, usually the name of
23852 the filter class or a specific filter instance name.
23854 @var{COMMAND} specifies the name of the command for the target filter.
23856 @var{ARG} is optional and specifies the optional list of argument for
23857 the given @var{COMMAND}.
23859 Between one interval specification and another, whitespaces, or
23860 sequences of characters starting with @code{#} until the end of line,
23861 are ignored and can be used to annotate comments.
23863 A simplified BNF description of the commands specification syntax
23866 @var{COMMAND_FLAG} ::= "enter" | "leave"
23867 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
23868 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
23869 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
23870 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
23871 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
23874 @subsection Examples
23878 Specify audio tempo change at second 4:
23880 asendcmd=c='4.0 atempo tempo 1.5',atempo
23884 Target a specific filter instance:
23886 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
23890 Specify a list of drawtext and hue commands in a file.
23892 # show text in the interval 5-10
23893 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
23894 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
23896 # desaturate the image in the interval 15-20
23897 15.0-20.0 [enter] hue s 0,
23898 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
23900 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
23902 # apply an exponential saturation fade-out effect, starting from time 25
23903 25 [enter] hue s exp(25-t)
23906 A filtergraph allowing to read and process the above command list
23907 stored in a file @file{test.cmd}, can be specified with:
23909 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
23914 @section setpts, asetpts
23916 Change the PTS (presentation timestamp) of the input frames.
23918 @code{setpts} works on video frames, @code{asetpts} on audio frames.
23920 This filter accepts the following options:
23925 The expression which is evaluated for each frame to construct its timestamp.
23929 The expression is evaluated through the eval API and can contain the following
23933 @item FRAME_RATE, FR
23934 frame rate, only defined for constant frame-rate video
23937 The presentation timestamp in input
23940 The count of the input frame for video or the number of consumed samples,
23941 not including the current frame for audio, starting from 0.
23943 @item NB_CONSUMED_SAMPLES
23944 The number of consumed samples, not including the current frame (only
23947 @item NB_SAMPLES, S
23948 The number of samples in the current frame (only audio)
23950 @item SAMPLE_RATE, SR
23951 The audio sample rate.
23954 The PTS of the first frame.
23957 the time in seconds of the first frame
23960 State whether the current frame is interlaced.
23963 the time in seconds of the current frame
23966 original position in the file of the frame, or undefined if undefined
23967 for the current frame
23970 The previous input PTS.
23973 previous input time in seconds
23976 The previous output PTS.
23979 previous output time in seconds
23982 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
23986 The wallclock (RTC) time at the start of the movie in microseconds.
23989 The timebase of the input timestamps.
23993 @subsection Examples
23997 Start counting PTS from zero
23999 setpts=PTS-STARTPTS
24003 Apply fast motion effect:
24009 Apply slow motion effect:
24015 Set fixed rate of 25 frames per second:
24021 Set fixed rate 25 fps with some jitter:
24023 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
24027 Apply an offset of 10 seconds to the input PTS:
24033 Generate timestamps from a "live source" and rebase onto the current timebase:
24035 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
24039 Generate timestamps by counting samples:
24048 Force color range for the output video frame.
24050 The @code{setrange} filter marks the color range property for the
24051 output frames. It does not change the input frame, but only sets the
24052 corresponding property, which affects how the frame is treated by
24055 The filter accepts the following options:
24060 Available values are:
24064 Keep the same color range property.
24066 @item unspecified, unknown
24067 Set the color range as unspecified.
24069 @item limited, tv, mpeg
24070 Set the color range as limited.
24072 @item full, pc, jpeg
24073 Set the color range as full.
24077 @section settb, asettb
24079 Set the timebase to use for the output frames timestamps.
24080 It is mainly useful for testing timebase configuration.
24082 It accepts the following parameters:
24087 The expression which is evaluated into the output timebase.
24091 The value for @option{tb} is an arithmetic expression representing a
24092 rational. The expression can contain the constants "AVTB" (the default
24093 timebase), "intb" (the input timebase) and "sr" (the sample rate,
24094 audio only). Default value is "intb".
24096 @subsection Examples
24100 Set the timebase to 1/25:
24106 Set the timebase to 1/10:
24112 Set the timebase to 1001/1000:
24118 Set the timebase to 2*intb:
24124 Set the default timebase value:
24131 Convert input audio to a video output representing frequency spectrum
24132 logarithmically using Brown-Puckette constant Q transform algorithm with
24133 direct frequency domain coefficient calculation (but the transform itself
24134 is not really constant Q, instead the Q factor is actually variable/clamped),
24135 with musical tone scale, from E0 to D#10.
24137 The filter accepts the following options:
24141 Specify the video size for the output. It must be even. For the syntax of this option,
24142 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24143 Default value is @code{1920x1080}.
24146 Set the output frame rate. Default value is @code{25}.
24149 Set the bargraph height. It must be even. Default value is @code{-1} which
24150 computes the bargraph height automatically.
24153 Set the axis height. It must be even. Default value is @code{-1} which computes
24154 the axis height automatically.
24157 Set the sonogram height. It must be even. Default value is @code{-1} which
24158 computes the sonogram height automatically.
24161 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
24162 instead. Default value is @code{1}.
24164 @item sono_v, volume
24165 Specify the sonogram volume expression. It can contain variables:
24168 the @var{bar_v} evaluated expression
24169 @item frequency, freq, f
24170 the frequency where it is evaluated
24171 @item timeclamp, tc
24172 the value of @var{timeclamp} option
24176 @item a_weighting(f)
24177 A-weighting of equal loudness
24178 @item b_weighting(f)
24179 B-weighting of equal loudness
24180 @item c_weighting(f)
24181 C-weighting of equal loudness.
24183 Default value is @code{16}.
24185 @item bar_v, volume2
24186 Specify the bargraph volume expression. It can contain variables:
24189 the @var{sono_v} evaluated expression
24190 @item frequency, freq, f
24191 the frequency where it is evaluated
24192 @item timeclamp, tc
24193 the value of @var{timeclamp} option
24197 @item a_weighting(f)
24198 A-weighting of equal loudness
24199 @item b_weighting(f)
24200 B-weighting of equal loudness
24201 @item c_weighting(f)
24202 C-weighting of equal loudness.
24204 Default value is @code{sono_v}.
24206 @item sono_g, gamma
24207 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
24208 higher gamma makes the spectrum having more range. Default value is @code{3}.
24209 Acceptable range is @code{[1, 7]}.
24211 @item bar_g, gamma2
24212 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
24216 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
24217 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
24219 @item timeclamp, tc
24220 Specify the transform timeclamp. At low frequency, there is trade-off between
24221 accuracy in time domain and frequency domain. If timeclamp is lower,
24222 event in time domain is represented more accurately (such as fast bass drum),
24223 otherwise event in frequency domain is represented more accurately
24224 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
24227 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
24228 limits future samples by applying asymmetric windowing in time domain, useful
24229 when low latency is required. Accepted range is @code{[0, 1]}.
24232 Specify the transform base frequency. Default value is @code{20.01523126408007475},
24233 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
24236 Specify the transform end frequency. Default value is @code{20495.59681441799654},
24237 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
24240 This option is deprecated and ignored.
24243 Specify the transform length in time domain. Use this option to control accuracy
24244 trade-off between time domain and frequency domain at every frequency sample.
24245 It can contain variables:
24247 @item frequency, freq, f
24248 the frequency where it is evaluated
24249 @item timeclamp, tc
24250 the value of @var{timeclamp} option.
24252 Default value is @code{384*tc/(384+tc*f)}.
24255 Specify the transform count for every video frame. Default value is @code{6}.
24256 Acceptable range is @code{[1, 30]}.
24259 Specify the transform count for every single pixel. Default value is @code{0},
24260 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
24263 Specify font file for use with freetype to draw the axis. If not specified,
24264 use embedded font. Note that drawing with font file or embedded font is not
24265 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
24269 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
24270 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
24274 Specify font color expression. This is arithmetic expression that should return
24275 integer value 0xRRGGBB. It can contain variables:
24277 @item frequency, freq, f
24278 the frequency where it is evaluated
24279 @item timeclamp, tc
24280 the value of @var{timeclamp} option
24285 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
24286 @item r(x), g(x), b(x)
24287 red, green, and blue value of intensity x.
24289 Default value is @code{st(0, (midi(f)-59.5)/12);
24290 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
24291 r(1-ld(1)) + b(ld(1))}.
24294 Specify image file to draw the axis. This option override @var{fontfile} and
24295 @var{fontcolor} option.
24298 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
24299 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
24300 Default value is @code{1}.
24303 Set colorspace. The accepted values are:
24306 Unspecified (default)
24315 BT.470BG or BT.601-6 625
24318 SMPTE-170M or BT.601-6 525
24324 BT.2020 with non-constant luminance
24329 Set spectrogram color scheme. This is list of floating point values with format
24330 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
24331 The default is @code{1|0.5|0|0|0.5|1}.
24335 @subsection Examples
24339 Playing audio while showing the spectrum:
24341 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
24345 Same as above, but with frame rate 30 fps:
24347 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
24351 Playing at 1280x720:
24353 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
24357 Disable sonogram display:
24363 A1 and its harmonics: A1, A2, (near)E3, A3:
24365 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),
24366 asplit[a][out1]; [a] showcqt [out0]'
24370 Same as above, but with more accuracy in frequency domain:
24372 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),
24373 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
24379 bar_v=10:sono_v=bar_v*a_weighting(f)
24383 Custom gamma, now spectrum is linear to the amplitude.
24389 Custom tlength equation:
24391 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)))'
24395 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
24397 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
24401 Custom font using fontconfig:
24403 font='Courier New,Monospace,mono|bold'
24407 Custom frequency range with custom axis using image file:
24409 axisfile=myaxis.png:basefreq=40:endfreq=10000
24415 Convert input audio to video output representing the audio power spectrum.
24416 Audio amplitude is on Y-axis while frequency is on X-axis.
24418 The filter accepts the following options:
24422 Specify size of video. For the syntax of this option, check the
24423 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24424 Default is @code{1024x512}.
24428 This set how each frequency bin will be represented.
24430 It accepts the following values:
24436 Default is @code{bar}.
24439 Set amplitude scale.
24441 It accepts the following values:
24455 Default is @code{log}.
24458 Set frequency scale.
24460 It accepts the following values:
24469 Reverse logarithmic scale.
24471 Default is @code{lin}.
24474 Set window size. Allowed range is from 16 to 65536.
24476 Default is @code{2048}
24479 Set windowing function.
24481 It accepts the following values:
24504 Default is @code{hanning}.
24507 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
24508 which means optimal overlap for selected window function will be picked.
24511 Set time averaging. Setting this to 0 will display current maximal peaks.
24512 Default is @code{1}, which means time averaging is disabled.
24515 Specify list of colors separated by space or by '|' which will be used to
24516 draw channel frequencies. Unrecognized or missing colors will be replaced
24520 Set channel display mode.
24522 It accepts the following values:
24527 Default is @code{combined}.
24530 Set minimum amplitude used in @code{log} amplitude scaler.
24534 @section showspatial
24536 Convert stereo input audio to a video output, representing the spatial relationship
24537 between two channels.
24539 The filter accepts the following options:
24543 Specify the video size for the output. For the syntax of this option, check the
24544 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24545 Default value is @code{512x512}.
24548 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
24551 Set window function.
24553 It accepts the following values:
24578 Default value is @code{hann}.
24581 Set ratio of overlap window. Default value is @code{0.5}.
24582 When value is @code{1} overlap is set to recommended size for specific
24583 window function currently used.
24586 @anchor{showspectrum}
24587 @section showspectrum
24589 Convert input audio to a video output, representing the audio frequency
24592 The filter accepts the following options:
24596 Specify the video size for the output. For the syntax of this option, check the
24597 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24598 Default value is @code{640x512}.
24601 Specify how the spectrum should slide along the window.
24603 It accepts the following values:
24606 the samples start again on the left when they reach the right
24608 the samples scroll from right to left
24610 frames are only produced when the samples reach the right
24612 the samples scroll from left to right
24615 Default value is @code{replace}.
24618 Specify display mode.
24620 It accepts the following values:
24623 all channels are displayed in the same row
24625 all channels are displayed in separate rows
24628 Default value is @samp{combined}.
24631 Specify display color mode.
24633 It accepts the following values:
24636 each channel is displayed in a separate color
24638 each channel is displayed using the same color scheme
24640 each channel is displayed using the rainbow color scheme
24642 each channel is displayed using the moreland color scheme
24644 each channel is displayed using the nebulae color scheme
24646 each channel is displayed using the fire color scheme
24648 each channel is displayed using the fiery color scheme
24650 each channel is displayed using the fruit color scheme
24652 each channel is displayed using the cool color scheme
24654 each channel is displayed using the magma color scheme
24656 each channel is displayed using the green color scheme
24658 each channel is displayed using the viridis color scheme
24660 each channel is displayed using the plasma color scheme
24662 each channel is displayed using the cividis color scheme
24664 each channel is displayed using the terrain color scheme
24667 Default value is @samp{channel}.
24670 Specify scale used for calculating intensity color values.
24672 It accepts the following values:
24677 square root, default
24688 Default value is @samp{sqrt}.
24691 Specify frequency scale.
24693 It accepts the following values:
24701 Default value is @samp{lin}.
24704 Set saturation modifier for displayed colors. Negative values provide
24705 alternative color scheme. @code{0} is no saturation at all.
24706 Saturation must be in [-10.0, 10.0] range.
24707 Default value is @code{1}.
24710 Set window function.
24712 It accepts the following values:
24737 Default value is @code{hann}.
24740 Set orientation of time vs frequency axis. Can be @code{vertical} or
24741 @code{horizontal}. Default is @code{vertical}.
24744 Set ratio of overlap window. Default value is @code{0}.
24745 When value is @code{1} overlap is set to recommended size for specific
24746 window function currently used.
24749 Set scale gain for calculating intensity color values.
24750 Default value is @code{1}.
24753 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
24756 Set color rotation, must be in [-1.0, 1.0] range.
24757 Default value is @code{0}.
24760 Set start frequency from which to display spectrogram. Default is @code{0}.
24763 Set stop frequency to which to display spectrogram. Default is @code{0}.
24766 Set upper frame rate limit. Default is @code{auto}, unlimited.
24769 Draw time and frequency axes and legends. Default is disabled.
24772 The usage is very similar to the showwaves filter; see the examples in that
24775 @subsection Examples
24779 Large window with logarithmic color scaling:
24781 showspectrum=s=1280x480:scale=log
24785 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
24787 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24788 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
24792 @section showspectrumpic
24794 Convert input audio to a single video frame, representing the audio frequency
24797 The filter accepts the following options:
24801 Specify the video size for the output. For the syntax of this option, check the
24802 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24803 Default value is @code{4096x2048}.
24806 Specify display mode.
24808 It accepts the following values:
24811 all channels are displayed in the same row
24813 all channels are displayed in separate rows
24815 Default value is @samp{combined}.
24818 Specify display color mode.
24820 It accepts the following values:
24823 each channel is displayed in a separate color
24825 each channel is displayed using the same color scheme
24827 each channel is displayed using the rainbow color scheme
24829 each channel is displayed using the moreland color scheme
24831 each channel is displayed using the nebulae color scheme
24833 each channel is displayed using the fire color scheme
24835 each channel is displayed using the fiery color scheme
24837 each channel is displayed using the fruit color scheme
24839 each channel is displayed using the cool color scheme
24841 each channel is displayed using the magma color scheme
24843 each channel is displayed using the green color scheme
24845 each channel is displayed using the viridis color scheme
24847 each channel is displayed using the plasma color scheme
24849 each channel is displayed using the cividis color scheme
24851 each channel is displayed using the terrain color scheme
24853 Default value is @samp{intensity}.
24856 Specify scale used for calculating intensity color values.
24858 It accepts the following values:
24863 square root, default
24873 Default value is @samp{log}.
24876 Specify frequency scale.
24878 It accepts the following values:
24886 Default value is @samp{lin}.
24889 Set saturation modifier for displayed colors. Negative values provide
24890 alternative color scheme. @code{0} is no saturation at all.
24891 Saturation must be in [-10.0, 10.0] range.
24892 Default value is @code{1}.
24895 Set window function.
24897 It accepts the following values:
24921 Default value is @code{hann}.
24924 Set orientation of time vs frequency axis. Can be @code{vertical} or
24925 @code{horizontal}. Default is @code{vertical}.
24928 Set scale gain for calculating intensity color values.
24929 Default value is @code{1}.
24932 Draw time and frequency axes and legends. Default is enabled.
24935 Set color rotation, must be in [-1.0, 1.0] range.
24936 Default value is @code{0}.
24939 Set start frequency from which to display spectrogram. Default is @code{0}.
24942 Set stop frequency to which to display spectrogram. Default is @code{0}.
24945 @subsection Examples
24949 Extract an audio spectrogram of a whole audio track
24950 in a 1024x1024 picture using @command{ffmpeg}:
24952 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
24956 @section showvolume
24958 Convert input audio volume to a video output.
24960 The filter accepts the following options:
24967 Set border width, allowed range is [0, 5]. Default is 1.
24970 Set channel width, allowed range is [80, 8192]. Default is 400.
24973 Set channel height, allowed range is [1, 900]. Default is 20.
24976 Set fade, allowed range is [0, 1]. Default is 0.95.
24979 Set volume color expression.
24981 The expression can use the following variables:
24985 Current max volume of channel in dB.
24991 Current channel number, starting from 0.
24995 If set, displays channel names. Default is enabled.
24998 If set, displays volume values. Default is enabled.
25001 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
25002 default is @code{h}.
25005 Set step size, allowed range is [0, 5]. Default is 0, which means
25009 Set background opacity, allowed range is [0, 1]. Default is 0.
25012 Set metering mode, can be peak: @code{p} or rms: @code{r},
25013 default is @code{p}.
25016 Set display scale, can be linear: @code{lin} or log: @code{log},
25017 default is @code{lin}.
25021 If set to > 0., display a line for the max level
25022 in the previous seconds.
25023 default is disabled: @code{0.}
25026 The color of the max line. Use when @code{dm} option is set to > 0.
25027 default is: @code{orange}
25032 Convert input audio to a video output, representing the samples waves.
25034 The filter accepts the following options:
25038 Specify the video size for the output. For the syntax of this option, check the
25039 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25040 Default value is @code{600x240}.
25045 Available values are:
25048 Draw a point for each sample.
25051 Draw a vertical line for each sample.
25054 Draw a point for each sample and a line between them.
25057 Draw a centered vertical line for each sample.
25060 Default value is @code{point}.
25063 Set the number of samples which are printed on the same column. A
25064 larger value will decrease the frame rate. Must be a positive
25065 integer. This option can be set only if the value for @var{rate}
25066 is not explicitly specified.
25069 Set the (approximate) output frame rate. This is done by setting the
25070 option @var{n}. Default value is "25".
25072 @item split_channels
25073 Set if channels should be drawn separately or overlap. Default value is 0.
25076 Set colors separated by '|' which are going to be used for drawing of each channel.
25079 Set amplitude scale.
25081 Available values are:
25099 Set the draw mode. This is mostly useful to set for high @var{n}.
25101 Available values are:
25104 Scale pixel values for each drawn sample.
25107 Draw every sample directly.
25110 Default value is @code{scale}.
25113 @subsection Examples
25117 Output the input file audio and the corresponding video representation
25120 amovie=a.mp3,asplit[out0],showwaves[out1]
25124 Create a synthetic signal and show it with showwaves, forcing a
25125 frame rate of 30 frames per second:
25127 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
25131 @section showwavespic
25133 Convert input audio to a single video frame, representing the samples waves.
25135 The filter accepts the following options:
25139 Specify the video size for the output. For the syntax of this option, check the
25140 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25141 Default value is @code{600x240}.
25143 @item split_channels
25144 Set if channels should be drawn separately or overlap. Default value is 0.
25147 Set colors separated by '|' which are going to be used for drawing of each channel.
25150 Set amplitude scale.
25152 Available values are:
25172 Available values are:
25175 Scale pixel values for each drawn sample.
25178 Draw every sample directly.
25181 Default value is @code{scale}.
25184 @subsection Examples
25188 Extract a channel split representation of the wave form of a whole audio track
25189 in a 1024x800 picture using @command{ffmpeg}:
25191 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
25195 @section sidedata, asidedata
25197 Delete frame side data, or select frames based on it.
25199 This filter accepts the following options:
25203 Set mode of operation of the filter.
25205 Can be one of the following:
25209 Select every frame with side data of @code{type}.
25212 Delete side data of @code{type}. If @code{type} is not set, delete all side
25218 Set side data type used with all modes. Must be set for @code{select} mode. For
25219 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
25220 in @file{libavutil/frame.h}. For example, to choose
25221 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
25225 @section spectrumsynth
25227 Synthesize audio from 2 input video spectrums, first input stream represents
25228 magnitude across time and second represents phase across time.
25229 The filter will transform from frequency domain as displayed in videos back
25230 to time domain as presented in audio output.
25232 This filter is primarily created for reversing processed @ref{showspectrum}
25233 filter outputs, but can synthesize sound from other spectrograms too.
25234 But in such case results are going to be poor if the phase data is not
25235 available, because in such cases phase data need to be recreated, usually
25236 it's just recreated from random noise.
25237 For best results use gray only output (@code{channel} color mode in
25238 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
25239 @code{lin} scale for phase video. To produce phase, for 2nd video, use
25240 @code{data} option. Inputs videos should generally use @code{fullframe}
25241 slide mode as that saves resources needed for decoding video.
25243 The filter accepts the following options:
25247 Specify sample rate of output audio, the sample rate of audio from which
25248 spectrum was generated may differ.
25251 Set number of channels represented in input video spectrums.
25254 Set scale which was used when generating magnitude input spectrum.
25255 Can be @code{lin} or @code{log}. Default is @code{log}.
25258 Set slide which was used when generating inputs spectrums.
25259 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
25260 Default is @code{fullframe}.
25263 Set window function used for resynthesis.
25266 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25267 which means optimal overlap for selected window function will be picked.
25270 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
25271 Default is @code{vertical}.
25274 @subsection Examples
25278 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
25279 then resynthesize videos back to audio with spectrumsynth:
25281 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
25282 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
25283 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
25287 @section split, asplit
25289 Split input into several identical outputs.
25291 @code{asplit} works with audio input, @code{split} with video.
25293 The filter accepts a single parameter which specifies the number of outputs. If
25294 unspecified, it defaults to 2.
25296 @subsection Examples
25300 Create two separate outputs from the same input:
25302 [in] split [out0][out1]
25306 To create 3 or more outputs, you need to specify the number of
25309 [in] asplit=3 [out0][out1][out2]
25313 Create two separate outputs from the same input, one cropped and
25316 [in] split [splitout1][splitout2];
25317 [splitout1] crop=100:100:0:0 [cropout];
25318 [splitout2] pad=200:200:100:100 [padout];
25322 Create 5 copies of the input audio with @command{ffmpeg}:
25324 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
25330 Receive commands sent through a libzmq client, and forward them to
25331 filters in the filtergraph.
25333 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
25334 must be inserted between two video filters, @code{azmq} between two
25335 audio filters. Both are capable to send messages to any filter type.
25337 To enable these filters you need to install the libzmq library and
25338 headers and configure FFmpeg with @code{--enable-libzmq}.
25340 For more information about libzmq see:
25341 @url{http://www.zeromq.org/}
25343 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
25344 receives messages sent through a network interface defined by the
25345 @option{bind_address} (or the abbreviation "@option{b}") option.
25346 Default value of this option is @file{tcp://localhost:5555}. You may
25347 want to alter this value to your needs, but do not forget to escape any
25348 ':' signs (see @ref{filtergraph escaping}).
25350 The received message must be in the form:
25352 @var{TARGET} @var{COMMAND} [@var{ARG}]
25355 @var{TARGET} specifies the target of the command, usually the name of
25356 the filter class or a specific filter instance name. The default
25357 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
25358 but you can override this by using the @samp{filter_name@@id} syntax
25359 (see @ref{Filtergraph syntax}).
25361 @var{COMMAND} specifies the name of the command for the target filter.
25363 @var{ARG} is optional and specifies the optional argument list for the
25364 given @var{COMMAND}.
25366 Upon reception, the message is processed and the corresponding command
25367 is injected into the filtergraph. Depending on the result, the filter
25368 will send a reply to the client, adopting the format:
25370 @var{ERROR_CODE} @var{ERROR_REASON}
25374 @var{MESSAGE} is optional.
25376 @subsection Examples
25378 Look at @file{tools/zmqsend} for an example of a zmq client which can
25379 be used to send commands processed by these filters.
25381 Consider the following filtergraph generated by @command{ffplay}.
25382 In this example the last overlay filter has an instance name. All other
25383 filters will have default instance names.
25386 ffplay -dumpgraph 1 -f lavfi "
25387 color=s=100x100:c=red [l];
25388 color=s=100x100:c=blue [r];
25389 nullsrc=s=200x100, zmq [bg];
25390 [bg][l] overlay [bg+l];
25391 [bg+l][r] overlay@@my=x=100 "
25394 To change the color of the left side of the video, the following
25395 command can be used:
25397 echo Parsed_color_0 c yellow | tools/zmqsend
25400 To change the right side:
25402 echo Parsed_color_1 c pink | tools/zmqsend
25405 To change the position of the right side:
25407 echo overlay@@my x 150 | tools/zmqsend
25411 @c man end MULTIMEDIA FILTERS
25413 @chapter Multimedia Sources
25414 @c man begin MULTIMEDIA SOURCES
25416 Below is a description of the currently available multimedia sources.
25420 This is the same as @ref{movie} source, except it selects an audio
25426 Read audio and/or video stream(s) from a movie container.
25428 It accepts the following parameters:
25432 The name of the resource to read (not necessarily a file; it can also be a
25433 device or a stream accessed through some protocol).
25435 @item format_name, f
25436 Specifies the format assumed for the movie to read, and can be either
25437 the name of a container or an input device. If not specified, the
25438 format is guessed from @var{movie_name} or by probing.
25440 @item seek_point, sp
25441 Specifies the seek point in seconds. The frames will be output
25442 starting from this seek point. The parameter is evaluated with
25443 @code{av_strtod}, so the numerical value may be suffixed by an IS
25444 postfix. The default value is "0".
25447 Specifies the streams to read. Several streams can be specified,
25448 separated by "+". The source will then have as many outputs, in the
25449 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
25450 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
25451 respectively the default (best suited) video and audio stream. Default
25452 is "dv", or "da" if the filter is called as "amovie".
25454 @item stream_index, si
25455 Specifies the index of the video stream to read. If the value is -1,
25456 the most suitable video stream will be automatically selected. The default
25457 value is "-1". Deprecated. If the filter is called "amovie", it will select
25458 audio instead of video.
25461 Specifies how many times to read the stream in sequence.
25462 If the value is 0, the stream will be looped infinitely.
25463 Default value is "1".
25465 Note that when the movie is looped the source timestamps are not
25466 changed, so it will generate non monotonically increasing timestamps.
25468 @item discontinuity
25469 Specifies the time difference between frames above which the point is
25470 considered a timestamp discontinuity which is removed by adjusting the later
25474 It allows overlaying a second video on top of the main input of
25475 a filtergraph, as shown in this graph:
25477 input -----------> deltapts0 --> overlay --> output
25480 movie --> scale--> deltapts1 -------+
25482 @subsection Examples
25486 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
25487 on top of the input labelled "in":
25489 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
25490 [in] setpts=PTS-STARTPTS [main];
25491 [main][over] overlay=16:16 [out]
25495 Read from a video4linux2 device, and overlay it on top of the input
25498 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
25499 [in] setpts=PTS-STARTPTS [main];
25500 [main][over] overlay=16:16 [out]
25504 Read the first video stream and the audio stream with id 0x81 from
25505 dvd.vob; the video is connected to the pad named "video" and the audio is
25506 connected to the pad named "audio":
25508 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
25512 @subsection Commands
25514 Both movie and amovie support the following commands:
25517 Perform seek using "av_seek_frame".
25518 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
25521 @var{stream_index}: If stream_index is -1, a default
25522 stream is selected, and @var{timestamp} is automatically converted
25523 from AV_TIME_BASE units to the stream specific time_base.
25525 @var{timestamp}: Timestamp in AVStream.time_base units
25526 or, if no stream is specified, in AV_TIME_BASE units.
25528 @var{flags}: Flags which select direction and seeking mode.
25532 Get movie duration in AV_TIME_BASE units.
25536 @c man end MULTIMEDIA SOURCES