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:
2360 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2361 this @code{lavfi.astats.Overall.Peak_count}.
2363 For description what each key means read below.
2366 Set number of frame after which stats are going to be recalculated.
2367 Default is disabled.
2369 @item measure_perchannel
2370 Select the entries which need to be measured per channel. The metadata keys can
2371 be used as flags, default is @option{all} which measures everything.
2372 @option{none} disables all per channel measurement.
2374 @item measure_overall
2375 Select the entries which need to be measured overall. The metadata keys can
2376 be used as flags, default is @option{all} which measures everything.
2377 @option{none} disables all overall measurement.
2381 A description of each shown parameter follows:
2385 Mean amplitude displacement from zero.
2388 Minimal sample level.
2391 Maximal sample level.
2393 @item Min difference
2394 Minimal difference between two consecutive samples.
2396 @item Max difference
2397 Maximal difference between two consecutive samples.
2399 @item Mean difference
2400 Mean difference between two consecutive samples.
2401 The average of each difference between two consecutive samples.
2403 @item RMS difference
2404 Root Mean Square difference between two consecutive samples.
2408 Standard peak and RMS level measured in dBFS.
2412 Peak and trough values for RMS level measured over a short window.
2415 Standard ratio of peak to RMS level (note: not in dB).
2418 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2419 (i.e. either @var{Min level} or @var{Max level}).
2422 Number of occasions (not the number of samples) that the signal attained either
2423 @var{Min level} or @var{Max level}.
2426 Overall bit depth of audio. Number of bits used for each sample.
2429 Measured dynamic range of audio in dB.
2431 @item Zero crossings
2432 Number of points where the waveform crosses the zero level axis.
2434 @item Zero crossings rate
2435 Rate of Zero crossings and number of audio samples.
2442 The filter accepts exactly one parameter, the audio tempo. If not
2443 specified then the filter will assume nominal 1.0 tempo. Tempo must
2444 be in the [0.5, 100.0] range.
2446 Note that tempo greater than 2 will skip some samples rather than
2447 blend them in. If for any reason this is a concern it is always
2448 possible to daisy-chain several instances of atempo to achieve the
2449 desired product tempo.
2451 @subsection Examples
2455 Slow down audio to 80% tempo:
2461 To speed up audio to 300% tempo:
2467 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2469 atempo=sqrt(3),atempo=sqrt(3)
2473 @subsection Commands
2475 This filter supports the following commands:
2478 Change filter tempo scale factor.
2479 Syntax for the command is : "@var{tempo}"
2484 Trim the input so that the output contains one continuous subpart of the input.
2486 It accepts the following parameters:
2489 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2490 sample with the timestamp @var{start} will be the first sample in the output.
2493 Specify time of the first audio sample that will be dropped, i.e. the
2494 audio sample immediately preceding the one with the timestamp @var{end} will be
2495 the last sample in the output.
2498 Same as @var{start}, except this option sets the start timestamp in samples
2502 Same as @var{end}, except this option sets the end timestamp in samples instead
2506 The maximum duration of the output in seconds.
2509 The number of the first sample that should be output.
2512 The number of the first sample that should be dropped.
2515 @option{start}, @option{end}, and @option{duration} are expressed as time
2516 duration specifications; see
2517 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2519 Note that the first two sets of the start/end options and the @option{duration}
2520 option look at the frame timestamp, while the _sample options simply count the
2521 samples that pass through the filter. So start/end_pts and start/end_sample will
2522 give different results when the timestamps are wrong, inexact or do not start at
2523 zero. Also note that this filter does not modify the timestamps. If you wish
2524 to have the output timestamps start at zero, insert the asetpts filter after the
2527 If multiple start or end options are set, this filter tries to be greedy and
2528 keep all samples that match at least one of the specified constraints. To keep
2529 only the part that matches all the constraints at once, chain multiple atrim
2532 The defaults are such that all the input is kept. So it is possible to set e.g.
2533 just the end values to keep everything before the specified time.
2538 Drop everything except the second minute of input:
2540 ffmpeg -i INPUT -af atrim=60:120
2544 Keep only the first 1000 samples:
2546 ffmpeg -i INPUT -af atrim=end_sample=1000
2551 @section axcorrelate
2552 Calculate normalized cross-correlation between two input audio streams.
2554 Resulted samples are always between -1 and 1 inclusive.
2555 If result is 1 it means two input samples are highly correlated in that selected segment.
2556 Result 0 means they are not correlated at all.
2557 If result is -1 it means two input samples are out of phase, which means they cancel each
2560 The filter accepts the following options:
2564 Set size of segment over which cross-correlation is calculated.
2565 Default is 256. Allowed range is from 2 to 131072.
2568 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2569 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2570 are always zero and thus need much less calculations to make.
2571 This is generally not true, but is valid for typical audio streams.
2574 @subsection Examples
2578 Calculate correlation between channels in stereo audio stream:
2580 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2586 Apply a two-pole Butterworth band-pass filter with central
2587 frequency @var{frequency}, and (3dB-point) band-width width.
2588 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2589 instead of the default: constant 0dB peak gain.
2590 The filter roll off at 6dB per octave (20dB per decade).
2592 The filter accepts the following options:
2596 Set the filter's central frequency. Default is @code{3000}.
2599 Constant skirt gain if set to 1. Defaults to 0.
2602 Set method to specify band-width of filter.
2617 Specify the band-width of a filter in width_type units.
2620 How much to use filtered signal in output. Default is 1.
2621 Range is between 0 and 1.
2624 Specify which channels to filter, by default all available are filtered.
2627 Normalize biquad coefficients, by default is disabled.
2628 Enabling it will normalize magnitude response at DC to 0dB.
2631 @subsection Commands
2633 This filter supports the following commands:
2636 Change bandpass frequency.
2637 Syntax for the command is : "@var{frequency}"
2640 Change bandpass width_type.
2641 Syntax for the command is : "@var{width_type}"
2644 Change bandpass width.
2645 Syntax for the command is : "@var{width}"
2648 Change bandpass mix.
2649 Syntax for the command is : "@var{mix}"
2654 Apply a two-pole Butterworth band-reject filter with central
2655 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2656 The filter roll off at 6dB per octave (20dB per decade).
2658 The filter accepts the following options:
2662 Set the filter's central frequency. Default is @code{3000}.
2665 Set method to specify band-width of filter.
2680 Specify the band-width of a filter in width_type units.
2683 How much to use filtered signal in output. Default is 1.
2684 Range is between 0 and 1.
2687 Specify which channels to filter, by default all available are filtered.
2690 Normalize biquad coefficients, by default is disabled.
2691 Enabling it will normalize magnitude response at DC to 0dB.
2694 @subsection Commands
2696 This filter supports the following commands:
2699 Change bandreject frequency.
2700 Syntax for the command is : "@var{frequency}"
2703 Change bandreject width_type.
2704 Syntax for the command is : "@var{width_type}"
2707 Change bandreject width.
2708 Syntax for the command is : "@var{width}"
2711 Change bandreject mix.
2712 Syntax for the command is : "@var{mix}"
2715 @section bass, lowshelf
2717 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2718 shelving filter with a response similar to that of a standard
2719 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2721 The filter accepts the following options:
2725 Give the gain at 0 Hz. Its useful range is about -20
2726 (for a large cut) to +20 (for a large boost).
2727 Beware of clipping when using a positive gain.
2730 Set the filter's central frequency and so can be used
2731 to extend or reduce the frequency range to be boosted or cut.
2732 The default value is @code{100} Hz.
2735 Set method to specify band-width of filter.
2750 Determine how steep is the filter's shelf transition.
2753 How much to use filtered signal in output. Default is 1.
2754 Range is between 0 and 1.
2757 Specify which channels to filter, by default all available are filtered.
2760 Normalize biquad coefficients, by default is disabled.
2761 Enabling it will normalize magnitude response at DC to 0dB.
2764 @subsection Commands
2766 This filter supports the following commands:
2769 Change bass frequency.
2770 Syntax for the command is : "@var{frequency}"
2773 Change bass width_type.
2774 Syntax for the command is : "@var{width_type}"
2778 Syntax for the command is : "@var{width}"
2782 Syntax for the command is : "@var{gain}"
2786 Syntax for the command is : "@var{mix}"
2791 Apply a biquad IIR filter with the given coefficients.
2792 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2793 are the numerator and denominator coefficients respectively.
2794 and @var{channels}, @var{c} specify which channels to filter, by default all
2795 available are filtered.
2797 @subsection Commands
2799 This filter supports the following commands:
2807 Change biquad parameter.
2808 Syntax for the command is : "@var{value}"
2811 How much to use filtered signal in output. Default is 1.
2812 Range is between 0 and 1.
2815 Specify which channels to filter, by default all available are filtered.
2818 Normalize biquad coefficients, by default is disabled.
2819 Enabling it will normalize magnitude response at DC to 0dB.
2823 Bauer stereo to binaural transformation, which improves headphone listening of
2824 stereo audio records.
2826 To enable compilation of this filter you need to configure FFmpeg with
2827 @code{--enable-libbs2b}.
2829 It accepts the following parameters:
2833 Pre-defined crossfeed level.
2837 Default level (fcut=700, feed=50).
2840 Chu Moy circuit (fcut=700, feed=60).
2843 Jan Meier circuit (fcut=650, feed=95).
2848 Cut frequency (in Hz).
2857 Remap input channels to new locations.
2859 It accepts the following parameters:
2862 Map channels from input to output. The argument is a '|'-separated list of
2863 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2864 @var{in_channel} form. @var{in_channel} can be either the name of the input
2865 channel (e.g. FL for front left) or its index in the input channel layout.
2866 @var{out_channel} is the name of the output channel or its index in the output
2867 channel layout. If @var{out_channel} is not given then it is implicitly an
2868 index, starting with zero and increasing by one for each mapping.
2870 @item channel_layout
2871 The channel layout of the output stream.
2874 If no mapping is present, the filter will implicitly map input channels to
2875 output channels, preserving indices.
2877 @subsection Examples
2881 For example, assuming a 5.1+downmix input MOV file,
2883 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2885 will create an output WAV file tagged as stereo from the downmix channels of
2889 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2891 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2895 @section channelsplit
2897 Split each channel from an input audio stream into a separate output stream.
2899 It accepts the following parameters:
2901 @item channel_layout
2902 The channel layout of the input stream. The default is "stereo".
2904 A channel layout describing the channels to be extracted as separate output streams
2905 or "all" to extract each input channel as a separate stream. The default is "all".
2907 Choosing channels not present in channel layout in the input will result in an error.
2910 @subsection Examples
2914 For example, assuming a stereo input MP3 file,
2916 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2918 will create an output Matroska file with two audio streams, one containing only
2919 the left channel and the other the right channel.
2922 Split a 5.1 WAV file into per-channel files:
2924 ffmpeg -i in.wav -filter_complex
2925 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2926 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2927 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2932 Extract only LFE from a 5.1 WAV file:
2934 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
2935 -map '[LFE]' lfe.wav
2940 Add a chorus effect to the audio.
2942 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2944 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2945 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2946 The modulation depth defines the range the modulated delay is played before or after
2947 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2948 sound tuned around the original one, like in a chorus where some vocals are slightly
2951 It accepts the following parameters:
2954 Set input gain. Default is 0.4.
2957 Set output gain. Default is 0.4.
2960 Set delays. A typical delay is around 40ms to 60ms.
2972 @subsection Examples
2978 chorus=0.7:0.9:55:0.4:0.25:2
2984 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2988 Fuller sounding chorus with three delays:
2990 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
2995 Compress or expand the audio's dynamic range.
2997 It accepts the following parameters:
3003 A list of times in seconds for each channel over which the instantaneous level
3004 of the input signal is averaged to determine its volume. @var{attacks} refers to
3005 increase of volume and @var{decays} refers to decrease of volume. For most
3006 situations, the attack time (response to the audio getting louder) should be
3007 shorter than the decay time, because the human ear is more sensitive to sudden
3008 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3009 a typical value for decay is 0.8 seconds.
3010 If specified number of attacks & decays is lower than number of channels, the last
3011 set attack/decay will be used for all remaining channels.
3014 A list of points for the transfer function, specified in dB relative to the
3015 maximum possible signal amplitude. Each key points list must be defined using
3016 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3017 @code{x0/y0 x1/y1 x2/y2 ....}
3019 The input values must be in strictly increasing order but the transfer function
3020 does not have to be monotonically rising. The point @code{0/0} is assumed but
3021 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3022 function are @code{-70/-70|-60/-20|1/0}.
3025 Set the curve radius in dB for all joints. It defaults to 0.01.
3028 Set the additional gain in dB to be applied at all points on the transfer
3029 function. This allows for easy adjustment of the overall gain.
3033 Set an initial volume, in dB, to be assumed for each channel when filtering
3034 starts. This permits the user to supply a nominal level initially, so that, for
3035 example, a very large gain is not applied to initial signal levels before the
3036 companding has begun to operate. A typical value for audio which is initially
3037 quiet is -90 dB. It defaults to 0.
3040 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3041 delayed before being fed to the volume adjuster. Specifying a delay
3042 approximately equal to the attack/decay times allows the filter to effectively
3043 operate in predictive rather than reactive mode. It defaults to 0.
3047 @subsection Examples
3051 Make music with both quiet and loud passages suitable for listening to in a
3054 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3057 Another example for audio with whisper and explosion parts:
3059 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3063 A noise gate for when the noise is at a lower level than the signal:
3065 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3069 Here is another noise gate, this time for when the noise is at a higher level
3070 than the signal (making it, in some ways, similar to squelch):
3072 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3076 2:1 compression starting at -6dB:
3078 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3082 2:1 compression starting at -9dB:
3084 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3088 2:1 compression starting at -12dB:
3090 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3094 2:1 compression starting at -18dB:
3096 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3100 3:1 compression starting at -15dB:
3102 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3108 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3114 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
3118 Hard limiter at -6dB:
3120 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3124 Hard limiter at -12dB:
3126 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3130 Hard noise gate at -35 dB:
3132 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3138 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3142 @section compensationdelay
3144 Compensation Delay Line is a metric based delay to compensate differing
3145 positions of microphones or speakers.
3147 For example, you have recorded guitar with two microphones placed in
3148 different locations. Because the front of sound wave has fixed speed in
3149 normal conditions, the phasing of microphones can vary and depends on
3150 their location and interposition. The best sound mix can be achieved when
3151 these microphones are in phase (synchronized). Note that a distance of
3152 ~30 cm between microphones makes one microphone capture the signal in
3153 antiphase to the other microphone. That makes the final mix sound moody.
3154 This filter helps to solve phasing problems by adding different delays
3155 to each microphone track and make them synchronized.
3157 The best result can be reached when you take one track as base and
3158 synchronize other tracks one by one with it.
3159 Remember that synchronization/delay tolerance depends on sample rate, too.
3160 Higher sample rates will give more tolerance.
3162 The filter accepts the following parameters:
3166 Set millimeters distance. This is compensation distance for fine tuning.
3170 Set cm distance. This is compensation distance for tightening distance setup.
3174 Set meters distance. This is compensation distance for hard distance setup.
3178 Set dry amount. Amount of unprocessed (dry) signal.
3182 Set wet amount. Amount of processed (wet) signal.
3186 Set temperature in degrees Celsius. This is the temperature of the environment.
3191 Apply headphone crossfeed filter.
3193 Crossfeed is the process of blending the left and right channels of stereo
3195 It is mainly used to reduce extreme stereo separation of low frequencies.
3197 The intent is to produce more speaker like sound to the listener.
3199 The filter accepts the following options:
3203 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3204 This sets gain of low shelf filter for side part of stereo image.
3205 Default is -6dB. Max allowed is -30db when strength is set to 1.
3208 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3209 This sets cut off frequency of low shelf filter. Default is cut off near
3210 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3213 Set input gain. Default is 0.9.
3216 Set output gain. Default is 1.
3219 @section crystalizer
3220 Simple algorithm to expand audio dynamic range.
3222 The filter accepts the following options:
3226 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3227 (unchanged sound) to 10.0 (maximum effect).
3230 Enable clipping. By default is enabled.
3233 @subsection Commands
3235 This filter supports the all above options as @ref{commands}.
3238 Apply a DC shift to the audio.
3240 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3241 in the recording chain) from the audio. The effect of a DC offset is reduced
3242 headroom and hence volume. The @ref{astats} filter can be used to determine if
3243 a signal has a DC offset.
3247 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3251 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3252 used to prevent clipping.
3257 Apply de-essing to the audio samples.
3261 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3265 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3269 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3273 Set the output mode.
3275 It accepts the following values:
3278 Pass input unchanged.
3281 Pass ess filtered out.
3286 Default value is @var{o}.
3292 Measure audio dynamic range.
3294 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3295 is found in transition material. And anything less that 8 have very poor dynamics
3296 and is very compressed.
3298 The filter accepts the following options:
3302 Set window length in seconds used to split audio into segments of equal length.
3303 Default is 3 seconds.
3307 Dynamic Audio Normalizer.
3309 This filter applies a certain amount of gain to the input audio in order
3310 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3311 contrast to more "simple" normalization algorithms, the Dynamic Audio
3312 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3313 This allows for applying extra gain to the "quiet" sections of the audio
3314 while avoiding distortions or clipping the "loud" sections. In other words:
3315 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3316 sections, in the sense that the volume of each section is brought to the
3317 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3318 this goal *without* applying "dynamic range compressing". It will retain 100%
3319 of the dynamic range *within* each section of the audio file.
3323 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3324 Default is 500 milliseconds.
3325 The Dynamic Audio Normalizer processes the input audio in small chunks,
3326 referred to as frames. This is required, because a peak magnitude has no
3327 meaning for just a single sample value. Instead, we need to determine the
3328 peak magnitude for a contiguous sequence of sample values. While a "standard"
3329 normalizer would simply use the peak magnitude of the complete file, the
3330 Dynamic Audio Normalizer determines the peak magnitude individually for each
3331 frame. The length of a frame is specified in milliseconds. By default, the
3332 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3333 been found to give good results with most files.
3334 Note that the exact frame length, in number of samples, will be determined
3335 automatically, based on the sampling rate of the individual input audio file.
3338 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3339 number. Default is 31.
3340 Probably the most important parameter of the Dynamic Audio Normalizer is the
3341 @code{window size} of the Gaussian smoothing filter. The filter's window size
3342 is specified in frames, centered around the current frame. For the sake of
3343 simplicity, this must be an odd number. Consequently, the default value of 31
3344 takes into account the current frame, as well as the 15 preceding frames and
3345 the 15 subsequent frames. Using a larger window results in a stronger
3346 smoothing effect and thus in less gain variation, i.e. slower gain
3347 adaptation. Conversely, using a smaller window results in a weaker smoothing
3348 effect and thus in more gain variation, i.e. faster gain adaptation.
3349 In other words, the more you increase this value, the more the Dynamic Audio
3350 Normalizer will behave like a "traditional" normalization filter. On the
3351 contrary, the more you decrease this value, the more the Dynamic Audio
3352 Normalizer will behave like a dynamic range compressor.
3355 Set the target peak value. This specifies the highest permissible magnitude
3356 level for the normalized audio input. This filter will try to approach the
3357 target peak magnitude as closely as possible, but at the same time it also
3358 makes sure that the normalized signal will never exceed the peak magnitude.
3359 A frame's maximum local gain factor is imposed directly by the target peak
3360 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3361 It is not recommended to go above this value.
3364 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3365 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3366 factor for each input frame, i.e. the maximum gain factor that does not
3367 result in clipping or distortion. The maximum gain factor is determined by
3368 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3369 additionally bounds the frame's maximum gain factor by a predetermined
3370 (global) maximum gain factor. This is done in order to avoid excessive gain
3371 factors in "silent" or almost silent frames. By default, the maximum gain
3372 factor is 10.0, For most inputs the default value should be sufficient and
3373 it usually is not recommended to increase this value. Though, for input
3374 with an extremely low overall volume level, it may be necessary to allow even
3375 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3376 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3377 Instead, a "sigmoid" threshold function will be applied. This way, the
3378 gain factors will smoothly approach the threshold value, but never exceed that
3382 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3383 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3384 This means that the maximum local gain factor for each frame is defined
3385 (only) by the frame's highest magnitude sample. This way, the samples can
3386 be amplified as much as possible without exceeding the maximum signal
3387 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3388 Normalizer can also take into account the frame's root mean square,
3389 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3390 determine the power of a time-varying signal. It is therefore considered
3391 that the RMS is a better approximation of the "perceived loudness" than
3392 just looking at the signal's peak magnitude. Consequently, by adjusting all
3393 frames to a constant RMS value, a uniform "perceived loudness" can be
3394 established. If a target RMS value has been specified, a frame's local gain
3395 factor is defined as the factor that would result in exactly that RMS value.
3396 Note, however, that the maximum local gain factor is still restricted by the
3397 frame's highest magnitude sample, in order to prevent clipping.
3400 Enable channels coupling. By default is enabled.
3401 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3402 amount. This means the same gain factor will be applied to all channels, i.e.
3403 the maximum possible gain factor is determined by the "loudest" channel.
3404 However, in some recordings, it may happen that the volume of the different
3405 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3406 In this case, this option can be used to disable the channel coupling. This way,
3407 the gain factor will be determined independently for each channel, depending
3408 only on the individual channel's highest magnitude sample. This allows for
3409 harmonizing the volume of the different channels.
3412 Enable DC bias correction. By default is disabled.
3413 An audio signal (in the time domain) is a sequence of sample values.
3414 In the Dynamic Audio Normalizer these sample values are represented in the
3415 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3416 audio signal, or "waveform", should be centered around the zero point.
3417 That means if we calculate the mean value of all samples in a file, or in a
3418 single frame, then the result should be 0.0 or at least very close to that
3419 value. If, however, there is a significant deviation of the mean value from
3420 0.0, in either positive or negative direction, this is referred to as a
3421 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3422 Audio Normalizer provides optional DC bias correction.
3423 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3424 the mean value, or "DC correction" offset, of each input frame and subtract
3425 that value from all of the frame's sample values which ensures those samples
3426 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3427 boundaries, the DC correction offset values will be interpolated smoothly
3428 between neighbouring frames.
3430 @item altboundary, b
3431 Enable alternative boundary mode. By default is disabled.
3432 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3433 around each frame. This includes the preceding frames as well as the
3434 subsequent frames. However, for the "boundary" frames, located at the very
3435 beginning and at the very end of the audio file, not all neighbouring
3436 frames are available. In particular, for the first few frames in the audio
3437 file, the preceding frames are not known. And, similarly, for the last few
3438 frames in the audio file, the subsequent frames are not known. Thus, the
3439 question arises which gain factors should be assumed for the missing frames
3440 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3441 to deal with this situation. The default boundary mode assumes a gain factor
3442 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3443 "fade out" at the beginning and at the end of the input, respectively.
3446 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3447 By default, the Dynamic Audio Normalizer does not apply "traditional"
3448 compression. This means that signal peaks will not be pruned and thus the
3449 full dynamic range will be retained within each local neighbourhood. However,
3450 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3451 normalization algorithm with a more "traditional" compression.
3452 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3453 (thresholding) function. If (and only if) the compression feature is enabled,
3454 all input frames will be processed by a soft knee thresholding function prior
3455 to the actual normalization process. Put simply, the thresholding function is
3456 going to prune all samples whose magnitude exceeds a certain threshold value.
3457 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3458 value. Instead, the threshold value will be adjusted for each individual
3460 In general, smaller parameters result in stronger compression, and vice versa.
3461 Values below 3.0 are not recommended, because audible distortion may appear.
3464 Set the target threshold value. This specifies the lowest permissible
3465 magnitude level for the audio input which will be normalized.
3466 If input frame volume is above this value frame will be normalized.
3467 Otherwise frame may not be normalized at all. The default value is set
3468 to 0, which means all input frames will be normalized.
3469 This option is mostly useful if digital noise is not wanted to be amplified.
3472 @subsection Commands
3474 This filter supports the all above options as @ref{commands}.
3478 Make audio easier to listen to on headphones.
3480 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3481 so that when listened to on headphones the stereo image is moved from
3482 inside your head (standard for headphones) to outside and in front of
3483 the listener (standard for speakers).
3489 Apply a two-pole peaking equalisation (EQ) filter. With this
3490 filter, the signal-level at and around a selected frequency can
3491 be increased or decreased, whilst (unlike bandpass and bandreject
3492 filters) that at all other frequencies is unchanged.
3494 In order to produce complex equalisation curves, this filter can
3495 be given several times, each with a different central frequency.
3497 The filter accepts the following options:
3501 Set the filter's central frequency in Hz.
3504 Set method to specify band-width of filter.
3519 Specify the band-width of a filter in width_type units.
3522 Set the required gain or attenuation in dB.
3523 Beware of clipping when using a positive gain.
3526 How much to use filtered signal in output. Default is 1.
3527 Range is between 0 and 1.
3530 Specify which channels to filter, by default all available are filtered.
3533 Normalize biquad coefficients, by default is disabled.
3534 Enabling it will normalize magnitude response at DC to 0dB.
3537 @subsection Examples
3540 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3542 equalizer=f=1000:t=h:width=200:g=-10
3546 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3548 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3552 @subsection Commands
3554 This filter supports the following commands:
3557 Change equalizer frequency.
3558 Syntax for the command is : "@var{frequency}"
3561 Change equalizer width_type.
3562 Syntax for the command is : "@var{width_type}"
3565 Change equalizer width.
3566 Syntax for the command is : "@var{width}"
3569 Change equalizer gain.
3570 Syntax for the command is : "@var{gain}"
3573 Change equalizer mix.
3574 Syntax for the command is : "@var{mix}"
3577 @section extrastereo
3579 Linearly increases the difference between left and right channels which
3580 adds some sort of "live" effect to playback.
3582 The filter accepts the following options:
3586 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3587 (average of both channels), with 1.0 sound will be unchanged, with
3588 -1.0 left and right channels will be swapped.
3591 Enable clipping. By default is enabled.
3594 @subsection Commands
3596 This filter supports the all above options as @ref{commands}.
3598 @section firequalizer
3599 Apply FIR Equalization using arbitrary frequency response.
3601 The filter accepts the following option:
3605 Set gain curve equation (in dB). The expression can contain variables:
3608 the evaluated frequency
3612 channel number, set to 0 when multichannels evaluation is disabled
3614 channel id, see libavutil/channel_layout.h, set to the first channel id when
3615 multichannels evaluation is disabled
3619 channel_layout, see libavutil/channel_layout.h
3624 @item gain_interpolate(f)
3625 interpolate gain on frequency f based on gain_entry
3626 @item cubic_interpolate(f)
3627 same as gain_interpolate, but smoother
3629 This option is also available as command. Default is @code{gain_interpolate(f)}.
3632 Set gain entry for gain_interpolate function. The expression can
3636 store gain entry at frequency f with value g
3638 This option is also available as command.
3641 Set filter delay in seconds. Higher value means more accurate.
3642 Default is @code{0.01}.
3645 Set filter accuracy in Hz. Lower value means more accurate.
3646 Default is @code{5}.
3649 Set window function. Acceptable values are:
3652 rectangular window, useful when gain curve is already smooth
3654 hann window (default)
3660 3-terms continuous 1st derivative nuttall window
3662 minimum 3-terms discontinuous nuttall window
3664 4-terms continuous 1st derivative nuttall window
3666 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3668 blackman-harris window
3674 If enabled, use fixed number of audio samples. This improves speed when
3675 filtering with large delay. Default is disabled.
3678 Enable multichannels evaluation on gain. Default is disabled.
3681 Enable zero phase mode by subtracting timestamp to compensate delay.
3682 Default is disabled.
3685 Set scale used by gain. Acceptable values are:
3688 linear frequency, linear gain
3690 linear frequency, logarithmic (in dB) gain (default)
3692 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3694 logarithmic frequency, logarithmic gain
3698 Set file for dumping, suitable for gnuplot.
3701 Set scale for dumpfile. Acceptable values are same with scale option.
3705 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3706 Default is disabled.
3709 Enable minimum phase impulse response. Default is disabled.
3712 @subsection Examples
3717 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3720 lowpass at 1000 Hz with gain_entry:
3722 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3725 custom equalization:
3727 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3730 higher delay with zero phase to compensate delay:
3732 firequalizer=delay=0.1:fixed=on:zero_phase=on
3735 lowpass on left channel, highpass on right channel:
3737 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3738 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3743 Apply a flanging effect to the audio.
3745 The filter accepts the following options:
3749 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3752 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3755 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3759 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3760 Default value is 71.
3763 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3766 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3767 Default value is @var{sinusoidal}.
3770 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3771 Default value is 25.
3774 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3775 Default is @var{linear}.
3779 Apply Haas effect to audio.
3781 Note that this makes most sense to apply on mono signals.
3782 With this filter applied to mono signals it give some directionality and
3783 stretches its stereo image.
3785 The filter accepts the following options:
3789 Set input level. By default is @var{1}, or 0dB
3792 Set output level. By default is @var{1}, or 0dB.
3795 Set gain applied to side part of signal. By default is @var{1}.
3798 Set kind of middle source. Can be one of the following:
3808 Pick middle part signal of stereo image.
3811 Pick side part signal of stereo image.
3815 Change middle phase. By default is disabled.
3818 Set left channel delay. By default is @var{2.05} milliseconds.
3821 Set left channel balance. By default is @var{-1}.
3824 Set left channel gain. By default is @var{1}.
3827 Change left phase. By default is disabled.
3830 Set right channel delay. By defaults is @var{2.12} milliseconds.
3833 Set right channel balance. By default is @var{1}.
3836 Set right channel gain. By default is @var{1}.
3839 Change right phase. By default is enabled.
3844 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3845 embedded HDCD codes is expanded into a 20-bit PCM stream.
3847 The filter supports the Peak Extend and Low-level Gain Adjustment features
3848 of HDCD, and detects the Transient Filter flag.
3851 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3854 When using the filter with wav, note the default encoding for wav is 16-bit,
3855 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3856 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3858 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3859 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3862 The filter accepts the following options:
3865 @item disable_autoconvert
3866 Disable any automatic format conversion or resampling in the filter graph.
3868 @item process_stereo
3869 Process the stereo channels together. If target_gain does not match between
3870 channels, consider it invalid and use the last valid target_gain.
3873 Set the code detect timer period in ms.
3876 Always extend peaks above -3dBFS even if PE isn't signaled.
3879 Replace audio with a solid tone and adjust the amplitude to signal some
3880 specific aspect of the decoding process. The output file can be loaded in
3881 an audio editor alongside the original to aid analysis.
3883 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3890 Gain adjustment level at each sample
3892 Samples where peak extend occurs
3894 Samples where the code detect timer is active
3896 Samples where the target gain does not match between channels
3902 Apply head-related transfer functions (HRTFs) to create virtual
3903 loudspeakers around the user for binaural listening via headphones.
3904 The HRIRs are provided via additional streams, for each channel
3905 one stereo input stream is needed.
3907 The filter accepts the following options:
3911 Set mapping of input streams for convolution.
3912 The argument is a '|'-separated list of channel names in order as they
3913 are given as additional stream inputs for filter.
3914 This also specify number of input streams. Number of input streams
3915 must be not less than number of channels in first stream plus one.
3918 Set gain applied to audio. Value is in dB. Default is 0.
3921 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3922 processing audio in time domain which is slow.
3923 @var{freq} is processing audio in frequency domain which is fast.
3924 Default is @var{freq}.
3927 Set custom gain for LFE channels. Value is in dB. Default is 0.
3930 Set size of frame in number of samples which will be processed at once.
3931 Default value is @var{1024}. Allowed range is from 1024 to 96000.
3934 Set format of hrir stream.
3935 Default value is @var{stereo}. Alternative value is @var{multich}.
3936 If value is set to @var{stereo}, number of additional streams should
3937 be greater or equal to number of input channels in first input stream.
3938 Also each additional stream should have stereo number of channels.
3939 If value is set to @var{multich}, number of additional streams should
3940 be exactly one. Also number of input channels of additional stream
3941 should be equal or greater than twice number of channels of first input
3945 @subsection Examples
3949 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3950 each amovie filter use stereo file with IR coefficients as input.
3951 The files give coefficients for each position of virtual loudspeaker:
3954 -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"
3959 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3960 but now in @var{multich} @var{hrir} format.
3962 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"
3969 Apply a high-pass filter with 3dB point frequency.
3970 The filter can be either single-pole, or double-pole (the default).
3971 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3973 The filter accepts the following options:
3977 Set frequency in Hz. Default is 3000.
3980 Set number of poles. Default is 2.
3983 Set method to specify band-width of filter.
3998 Specify the band-width of a filter in width_type units.
3999 Applies only to double-pole filter.
4000 The default is 0.707q and gives a Butterworth response.
4003 How much to use filtered signal in output. Default is 1.
4004 Range is between 0 and 1.
4007 Specify which channels to filter, by default all available are filtered.
4010 Normalize biquad coefficients, by default is disabled.
4011 Enabling it will normalize magnitude response at DC to 0dB.
4014 @subsection Commands
4016 This filter supports the following commands:
4019 Change highpass frequency.
4020 Syntax for the command is : "@var{frequency}"
4023 Change highpass width_type.
4024 Syntax for the command is : "@var{width_type}"
4027 Change highpass width.
4028 Syntax for the command is : "@var{width}"
4031 Change highpass mix.
4032 Syntax for the command is : "@var{mix}"
4037 Join multiple input streams into one multi-channel stream.
4039 It accepts the following parameters:
4043 The number of input streams. It defaults to 2.
4045 @item channel_layout
4046 The desired output channel layout. It defaults to stereo.
4049 Map channels from inputs to output. The argument is a '|'-separated list of
4050 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4051 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4052 can be either the name of the input channel (e.g. FL for front left) or its
4053 index in the specified input stream. @var{out_channel} is the name of the output
4057 The filter will attempt to guess the mappings when they are not specified
4058 explicitly. It does so by first trying to find an unused matching input channel
4059 and if that fails it picks the first unused input channel.
4061 Join 3 inputs (with properly set channel layouts):
4063 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4066 Build a 5.1 output from 6 single-channel streams:
4068 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4069 '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'
4075 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4077 To enable compilation of this filter you need to configure FFmpeg with
4078 @code{--enable-ladspa}.
4082 Specifies the name of LADSPA plugin library to load. If the environment
4083 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4084 each one of the directories specified by the colon separated list in
4085 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4086 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4087 @file{/usr/lib/ladspa/}.
4090 Specifies the plugin within the library. Some libraries contain only
4091 one plugin, but others contain many of them. If this is not set filter
4092 will list all available plugins within the specified library.
4095 Set the '|' separated list of controls which are zero or more floating point
4096 values that determine the behavior of the loaded plugin (for example delay,
4098 Controls need to be defined using the following syntax:
4099 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4100 @var{valuei} is the value set on the @var{i}-th control.
4101 Alternatively they can be also defined using the following syntax:
4102 @var{value0}|@var{value1}|@var{value2}|..., where
4103 @var{valuei} is the value set on the @var{i}-th control.
4104 If @option{controls} is set to @code{help}, all available controls and
4105 their valid ranges are printed.
4107 @item sample_rate, s
4108 Specify the sample rate, default to 44100. Only used if plugin have
4112 Set the number of samples per channel per each output frame, default
4113 is 1024. Only used if plugin have zero inputs.
4116 Set the minimum duration of the sourced audio. See
4117 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4118 for the accepted syntax.
4119 Note that the resulting duration may be greater than the specified duration,
4120 as the generated audio is always cut at the end of a complete frame.
4121 If not specified, or the expressed duration is negative, the audio is
4122 supposed to be generated forever.
4123 Only used if plugin have zero inputs.
4127 @subsection Examples
4131 List all available plugins within amp (LADSPA example plugin) library:
4137 List all available controls and their valid ranges for @code{vcf_notch}
4138 plugin from @code{VCF} library:
4140 ladspa=f=vcf:p=vcf_notch:c=help
4144 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4147 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4151 Add reverberation to the audio using TAP-plugins
4152 (Tom's Audio Processing plugins):
4154 ladspa=file=tap_reverb:tap_reverb
4158 Generate white noise, with 0.2 amplitude:
4160 ladspa=file=cmt:noise_source_white:c=c0=.2
4164 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4165 @code{C* Audio Plugin Suite} (CAPS) library:
4167 ladspa=file=caps:Click:c=c1=20'
4171 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4173 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4177 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4178 @code{SWH Plugins} collection:
4180 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4184 Attenuate low frequencies using Multiband EQ from Steve Harris
4185 @code{SWH Plugins} collection:
4187 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4191 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4194 ladspa=caps:Narrower
4198 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4200 ladspa=caps:White:.2
4204 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4206 ladspa=caps:Fractal:c=c1=1
4210 Dynamic volume normalization using @code{VLevel} plugin:
4212 ladspa=vlevel-ladspa:vlevel_mono
4216 @subsection Commands
4218 This filter supports the following commands:
4221 Modify the @var{N}-th control value.
4223 If the specified value is not valid, it is ignored and prior one is kept.
4228 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4229 Support for both single pass (livestreams, files) and double pass (files) modes.
4230 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4231 detect true peaks, the audio stream will be upsampled to 192 kHz.
4232 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4234 The filter accepts the following options:
4238 Set integrated loudness target.
4239 Range is -70.0 - -5.0. Default value is -24.0.
4242 Set loudness range target.
4243 Range is 1.0 - 20.0. Default value is 7.0.
4246 Set maximum true peak.
4247 Range is -9.0 - +0.0. Default value is -2.0.
4249 @item measured_I, measured_i
4250 Measured IL of input file.
4251 Range is -99.0 - +0.0.
4253 @item measured_LRA, measured_lra
4254 Measured LRA of input file.
4255 Range is 0.0 - 99.0.
4257 @item measured_TP, measured_tp
4258 Measured true peak of input file.
4259 Range is -99.0 - +99.0.
4261 @item measured_thresh
4262 Measured threshold of input file.
4263 Range is -99.0 - +0.0.
4266 Set offset gain. Gain is applied before the true-peak limiter.
4267 Range is -99.0 - +99.0. Default is +0.0.
4270 Normalize by linearly scaling the source audio.
4271 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4272 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4273 be lower than source LRA and the change in integrated loudness shouldn't
4274 result in a true peak which exceeds the target TP. If any of these
4275 conditions aren't met, normalization mode will revert to @var{dynamic}.
4276 Options are @code{true} or @code{false}. Default is @code{true}.
4279 Treat mono input files as "dual-mono". If a mono file is intended for playback
4280 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4281 If set to @code{true}, this option will compensate for this effect.
4282 Multi-channel input files are not affected by this option.
4283 Options are true or false. Default is false.
4286 Set print format for stats. Options are summary, json, or none.
4287 Default value is none.
4292 Apply a low-pass filter with 3dB point frequency.
4293 The filter can be either single-pole or double-pole (the default).
4294 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4296 The filter accepts the following options:
4300 Set frequency in Hz. Default is 500.
4303 Set number of poles. Default is 2.
4306 Set method to specify band-width of filter.
4321 Specify the band-width of a filter in width_type units.
4322 Applies only to double-pole filter.
4323 The default is 0.707q and gives a Butterworth response.
4326 How much to use filtered signal in output. Default is 1.
4327 Range is between 0 and 1.
4330 Specify which channels to filter, by default all available are filtered.
4333 Normalize biquad coefficients, by default is disabled.
4334 Enabling it will normalize magnitude response at DC to 0dB.
4337 @subsection Examples
4340 Lowpass only LFE channel, it LFE is not present it does nothing:
4346 @subsection Commands
4348 This filter supports the following commands:
4351 Change lowpass frequency.
4352 Syntax for the command is : "@var{frequency}"
4355 Change lowpass width_type.
4356 Syntax for the command is : "@var{width_type}"
4359 Change lowpass width.
4360 Syntax for the command is : "@var{width}"
4364 Syntax for the command is : "@var{mix}"
4369 Load a LV2 (LADSPA Version 2) plugin.
4371 To enable compilation of this filter you need to configure FFmpeg with
4372 @code{--enable-lv2}.
4376 Specifies the plugin URI. You may need to escape ':'.
4379 Set the '|' separated list of controls which are zero or more floating point
4380 values that determine the behavior of the loaded plugin (for example delay,
4382 If @option{controls} is set to @code{help}, all available controls and
4383 their valid ranges are printed.
4385 @item sample_rate, s
4386 Specify the sample rate, default to 44100. Only used if plugin have
4390 Set the number of samples per channel per each output frame, default
4391 is 1024. Only used if plugin have zero inputs.
4394 Set the minimum duration of the sourced audio. See
4395 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4396 for the accepted syntax.
4397 Note that the resulting duration may be greater than the specified duration,
4398 as the generated audio is always cut at the end of a complete frame.
4399 If not specified, or the expressed duration is negative, the audio is
4400 supposed to be generated forever.
4401 Only used if plugin have zero inputs.
4404 @subsection Examples
4408 Apply bass enhancer plugin from Calf:
4410 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4414 Apply vinyl plugin from Calf:
4416 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4420 Apply bit crusher plugin from ArtyFX:
4422 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4427 Multiband Compress or expand the audio's dynamic range.
4429 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4430 This is akin to the crossover of a loudspeaker, and results in flat frequency
4431 response when absent compander action.
4433 It accepts the following parameters:
4437 This option syntax is:
4438 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4439 For explanation of each item refer to compand filter documentation.
4445 Mix channels with specific gain levels. The filter accepts the output
4446 channel layout followed by a set of channels definitions.
4448 This filter is also designed to efficiently remap the channels of an audio
4451 The filter accepts parameters of the form:
4452 "@var{l}|@var{outdef}|@var{outdef}|..."
4456 output channel layout or number of channels
4459 output channel specification, of the form:
4460 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4463 output channel to define, either a channel name (FL, FR, etc.) or a channel
4464 number (c0, c1, etc.)
4467 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4470 input channel to use, see out_name for details; it is not possible to mix
4471 named and numbered input channels
4474 If the `=' in a channel specification is replaced by `<', then the gains for
4475 that specification will be renormalized so that the total is 1, thus
4476 avoiding clipping noise.
4478 @subsection Mixing examples
4480 For example, if you want to down-mix from stereo to mono, but with a bigger
4481 factor for the left channel:
4483 pan=1c|c0=0.9*c0+0.1*c1
4486 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4487 7-channels surround:
4489 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4492 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4493 that should be preferred (see "-ac" option) unless you have very specific
4496 @subsection Remapping examples
4498 The channel remapping will be effective if, and only if:
4501 @item gain coefficients are zeroes or ones,
4502 @item only one input per channel output,
4505 If all these conditions are satisfied, the filter will notify the user ("Pure
4506 channel mapping detected"), and use an optimized and lossless method to do the
4509 For example, if you have a 5.1 source and want a stereo audio stream by
4510 dropping the extra channels:
4512 pan="stereo| c0=FL | c1=FR"
4515 Given the same source, you can also switch front left and front right channels
4516 and keep the input channel layout:
4518 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4521 If the input is a stereo audio stream, you can mute the front left channel (and
4522 still keep the stereo channel layout) with:
4527 Still with a stereo audio stream input, you can copy the right channel in both
4528 front left and right:
4530 pan="stereo| c0=FR | c1=FR"
4535 ReplayGain scanner filter. This filter takes an audio stream as an input and
4536 outputs it unchanged.
4537 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4541 Convert the audio sample format, sample rate and channel layout. It is
4542 not meant to be used directly.
4545 Apply time-stretching and pitch-shifting with librubberband.
4547 To enable compilation of this filter, you need to configure FFmpeg with
4548 @code{--enable-librubberband}.
4550 The filter accepts the following options:
4554 Set tempo scale factor.
4557 Set pitch scale factor.
4560 Set transients detector.
4561 Possible values are:
4570 Possible values are:
4579 Possible values are:
4586 Set processing window size.
4587 Possible values are:
4596 Possible values are:
4603 Enable formant preservation when shift pitching.
4604 Possible values are:
4612 Possible values are:
4621 Possible values are:
4628 @subsection Commands
4630 This filter supports the following commands:
4633 Change filter tempo scale factor.
4634 Syntax for the command is : "@var{tempo}"
4637 Change filter pitch scale factor.
4638 Syntax for the command is : "@var{pitch}"
4641 @section sidechaincompress
4643 This filter acts like normal compressor but has the ability to compress
4644 detected signal using second input signal.
4645 It needs two input streams and returns one output stream.
4646 First input stream will be processed depending on second stream signal.
4647 The filtered signal then can be filtered with other filters in later stages of
4648 processing. See @ref{pan} and @ref{amerge} filter.
4650 The filter accepts the following options:
4654 Set input gain. Default is 1. Range is between 0.015625 and 64.
4657 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4658 Default is @code{downward}.
4661 If a signal of second stream raises above this level it will affect the gain
4662 reduction of first stream.
4663 By default is 0.125. Range is between 0.00097563 and 1.
4666 Set a ratio about which the signal is reduced. 1:2 means that if the level
4667 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4668 Default is 2. Range is between 1 and 20.
4671 Amount of milliseconds the signal has to rise above the threshold before gain
4672 reduction starts. Default is 20. Range is between 0.01 and 2000.
4675 Amount of milliseconds the signal has to fall below the threshold before
4676 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4679 Set the amount by how much signal will be amplified after processing.
4680 Default is 1. Range is from 1 to 64.
4683 Curve the sharp knee around the threshold to enter gain reduction more softly.
4684 Default is 2.82843. Range is between 1 and 8.
4687 Choose if the @code{average} level between all channels of side-chain stream
4688 or the louder(@code{maximum}) channel of side-chain stream affects the
4689 reduction. Default is @code{average}.
4692 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4693 of @code{rms}. Default is @code{rms} which is mainly smoother.
4696 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4699 How much to use compressed signal in output. Default is 1.
4700 Range is between 0 and 1.
4703 @subsection Commands
4705 This filter supports the all above options as @ref{commands}.
4707 @subsection Examples
4711 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4712 depending on the signal of 2nd input and later compressed signal to be
4713 merged with 2nd input:
4715 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4719 @section sidechaingate
4721 A sidechain gate acts like a normal (wideband) gate but has the ability to
4722 filter the detected signal before sending it to the gain reduction stage.
4723 Normally a gate uses the full range signal to detect a level above the
4725 For example: If you cut all lower frequencies from your sidechain signal
4726 the gate will decrease the volume of your track only if not enough highs
4727 appear. With this technique you are able to reduce the resonation of a
4728 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4730 It needs two input streams and returns one output stream.
4731 First input stream will be processed depending on second stream signal.
4733 The filter accepts the following options:
4737 Set input level before filtering.
4738 Default is 1. Allowed range is from 0.015625 to 64.
4741 Set the mode of operation. Can be @code{upward} or @code{downward}.
4742 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
4743 will be amplified, expanding dynamic range in upward direction.
4744 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
4747 Set the level of gain reduction when the signal is below the threshold.
4748 Default is 0.06125. Allowed range is from 0 to 1.
4749 Setting this to 0 disables reduction and then filter behaves like expander.
4752 If a signal rises above this level the gain reduction is released.
4753 Default is 0.125. Allowed range is from 0 to 1.
4756 Set a ratio about which the signal is reduced.
4757 Default is 2. Allowed range is from 1 to 9000.
4760 Amount of milliseconds the signal has to rise above the threshold before gain
4762 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4765 Amount of milliseconds the signal has to fall below the threshold before the
4766 reduction is increased again. Default is 250 milliseconds.
4767 Allowed range is from 0.01 to 9000.
4770 Set amount of amplification of signal after processing.
4771 Default is 1. Allowed range is from 1 to 64.
4774 Curve the sharp knee around the threshold to enter gain reduction more softly.
4775 Default is 2.828427125. Allowed range is from 1 to 8.
4778 Choose if exact signal should be taken for detection or an RMS like one.
4779 Default is rms. Can be peak or rms.
4782 Choose if the average level between all channels or the louder channel affects
4784 Default is average. Can be average or maximum.
4787 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4790 @section silencedetect
4792 Detect silence in an audio stream.
4794 This filter logs a message when it detects that the input audio volume is less
4795 or equal to a noise tolerance value for a duration greater or equal to the
4796 minimum detected noise duration.
4798 The printed times and duration are expressed in seconds. The
4799 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
4800 is set on the first frame whose timestamp equals or exceeds the detection
4801 duration and it contains the timestamp of the first frame of the silence.
4803 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
4804 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
4805 keys are set on the first frame after the silence. If @option{mono} is
4806 enabled, and each channel is evaluated separately, the @code{.X}
4807 suffixed keys are used, and @code{X} corresponds to the channel number.
4809 The filter accepts the following options:
4813 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4814 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4817 Set silence duration until notification (default is 2 seconds). See
4818 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4819 for the accepted syntax.
4822 Process each channel separately, instead of combined. By default is disabled.
4825 @subsection Examples
4829 Detect 5 seconds of silence with -50dB noise tolerance:
4831 silencedetect=n=-50dB:d=5
4835 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4836 tolerance in @file{silence.mp3}:
4838 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4842 @section silenceremove
4844 Remove silence from the beginning, middle or end of the audio.
4846 The filter accepts the following options:
4850 This value is used to indicate if audio should be trimmed at beginning of
4851 the audio. A value of zero indicates no silence should be trimmed from the
4852 beginning. When specifying a non-zero value, it trims audio up until it
4853 finds non-silence. Normally, when trimming silence from beginning of audio
4854 the @var{start_periods} will be @code{1} but it can be increased to higher
4855 values to trim all audio up to specific count of non-silence periods.
4856 Default value is @code{0}.
4858 @item start_duration
4859 Specify the amount of time that non-silence must be detected before it stops
4860 trimming audio. By increasing the duration, bursts of noises can be treated
4861 as silence and trimmed off. Default value is @code{0}.
4863 @item start_threshold
4864 This indicates what sample value should be treated as silence. For digital
4865 audio, a value of @code{0} may be fine but for audio recorded from analog,
4866 you may wish to increase the value to account for background noise.
4867 Can be specified in dB (in case "dB" is appended to the specified value)
4868 or amplitude ratio. Default value is @code{0}.
4871 Specify max duration of silence at beginning that will be kept after
4872 trimming. Default is 0, which is equal to trimming all samples detected
4876 Specify mode of detection of silence end in start of multi-channel audio.
4877 Can be @var{any} or @var{all}. Default is @var{any}.
4878 With @var{any}, any sample that is detected as non-silence will cause
4879 stopped trimming of silence.
4880 With @var{all}, only if all channels are detected as non-silence will cause
4881 stopped trimming of silence.
4884 Set the count for trimming silence from the end of audio.
4885 To remove silence from the middle of a file, specify a @var{stop_periods}
4886 that is negative. This value is then treated as a positive value and is
4887 used to indicate the effect should restart processing as specified by
4888 @var{start_periods}, making it suitable for removing periods of silence
4889 in the middle of the audio.
4890 Default value is @code{0}.
4893 Specify a duration of silence that must exist before audio is not copied any
4894 more. By specifying a higher duration, silence that is wanted can be left in
4896 Default value is @code{0}.
4898 @item stop_threshold
4899 This is the same as @option{start_threshold} but for trimming silence from
4901 Can be specified in dB (in case "dB" is appended to the specified value)
4902 or amplitude ratio. Default value is @code{0}.
4905 Specify max duration of silence at end that will be kept after
4906 trimming. Default is 0, which is equal to trimming all samples detected
4910 Specify mode of detection of silence start in end of multi-channel audio.
4911 Can be @var{any} or @var{all}. Default is @var{any}.
4912 With @var{any}, any sample that is detected as non-silence will cause
4913 stopped trimming of silence.
4914 With @var{all}, only if all channels are detected as non-silence will cause
4915 stopped trimming of silence.
4918 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4919 and works better with digital silence which is exactly 0.
4920 Default value is @code{rms}.
4923 Set duration in number of seconds used to calculate size of window in number
4924 of samples for detecting silence.
4925 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4928 @subsection Examples
4932 The following example shows how this filter can be used to start a recording
4933 that does not contain the delay at the start which usually occurs between
4934 pressing the record button and the start of the performance:
4936 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
4940 Trim all silence encountered from beginning to end where there is more than 1
4941 second of silence in audio:
4943 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
4947 Trim all digital silence samples, using peak detection, from beginning to end
4948 where there is more than 0 samples of digital silence in audio and digital
4949 silence is detected in all channels at same positions in stream:
4951 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
4957 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4958 loudspeakers around the user for binaural listening via headphones (audio
4959 formats up to 9 channels supported).
4960 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4961 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4962 Austrian Academy of Sciences.
4964 To enable compilation of this filter you need to configure FFmpeg with
4965 @code{--enable-libmysofa}.
4967 The filter accepts the following options:
4971 Set the SOFA file used for rendering.
4974 Set gain applied to audio. Value is in dB. Default is 0.
4977 Set rotation of virtual loudspeakers in deg. Default is 0.
4980 Set elevation of virtual speakers in deg. Default is 0.
4983 Set distance in meters between loudspeakers and the listener with near-field
4984 HRTFs. Default is 1.
4987 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4988 processing audio in time domain which is slow.
4989 @var{freq} is processing audio in frequency domain which is fast.
4990 Default is @var{freq}.
4993 Set custom positions of virtual loudspeakers. Syntax for this option is:
4994 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4995 Each virtual loudspeaker is described with short channel name following with
4996 azimuth and elevation in degrees.
4997 Each virtual loudspeaker description is separated by '|'.
4998 For example to override front left and front right channel positions use:
4999 'speakers=FL 45 15|FR 345 15'.
5000 Descriptions with unrecognised channel names are ignored.
5003 Set custom gain for LFE channels. Value is in dB. Default is 0.
5006 Set custom frame size in number of samples. Default is 1024.
5007 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5008 is set to @var{freq}.
5011 Should all IRs be normalized upon importing SOFA file.
5012 By default is enabled.
5015 Should nearest IRs be interpolated with neighbor IRs if exact position
5016 does not match. By default is disabled.
5019 Minphase all IRs upon loading of SOFA file. By default is disabled.
5022 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5025 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5028 @subsection Examples
5032 Using ClubFritz6 sofa file:
5034 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5038 Using ClubFritz12 sofa file and bigger radius with small rotation:
5040 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5044 Similar as above but with custom speaker positions for front left, front right, back left and back right
5045 and also with custom gain:
5047 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5051 @section stereotools
5053 This filter has some handy utilities to manage stereo signals, for converting
5054 M/S stereo recordings to L/R signal while having control over the parameters
5055 or spreading the stereo image of master track.
5057 The filter accepts the following options:
5061 Set input level before filtering for both channels. Defaults is 1.
5062 Allowed range is from 0.015625 to 64.
5065 Set output level after filtering for both channels. Defaults is 1.
5066 Allowed range is from 0.015625 to 64.
5069 Set input balance between both channels. Default is 0.
5070 Allowed range is from -1 to 1.
5073 Set output balance between both channels. Default is 0.
5074 Allowed range is from -1 to 1.
5077 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5078 clipping. Disabled by default.
5081 Mute the left channel. Disabled by default.
5084 Mute the right channel. Disabled by default.
5087 Change the phase of the left channel. Disabled by default.
5090 Change the phase of the right channel. Disabled by default.
5093 Set stereo mode. Available values are:
5097 Left/Right to Left/Right, this is default.
5100 Left/Right to Mid/Side.
5103 Mid/Side to Left/Right.
5106 Left/Right to Left/Left.
5109 Left/Right to Right/Right.
5112 Left/Right to Left + Right.
5115 Left/Right to Right/Left.
5118 Mid/Side to Left/Left.
5121 Mid/Side to Right/Right.
5125 Set level of side signal. Default is 1.
5126 Allowed range is from 0.015625 to 64.
5129 Set balance of side signal. Default is 0.
5130 Allowed range is from -1 to 1.
5133 Set level of the middle signal. Default is 1.
5134 Allowed range is from 0.015625 to 64.
5137 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5140 Set stereo base between mono and inversed channels. Default is 0.
5141 Allowed range is from -1 to 1.
5144 Set delay in milliseconds how much to delay left from right channel and
5145 vice versa. Default is 0. Allowed range is from -20 to 20.
5148 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5151 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5153 @item bmode_in, bmode_out
5154 Set balance mode for balance_in/balance_out option.
5156 Can be one of the following:
5160 Classic balance mode. Attenuate one channel at time.
5161 Gain is raised up to 1.
5164 Similar as classic mode above but gain is raised up to 2.
5167 Equal power distribution, from -6dB to +6dB range.
5171 @subsection Examples
5175 Apply karaoke like effect:
5177 stereotools=mlev=0.015625
5181 Convert M/S signal to L/R:
5183 "stereotools=mode=ms>lr"
5187 @section stereowiden
5189 This filter enhance the stereo effect by suppressing signal common to both
5190 channels and by delaying the signal of left into right and vice versa,
5191 thereby widening the stereo effect.
5193 The filter accepts the following options:
5197 Time in milliseconds of the delay of left signal into right and vice versa.
5198 Default is 20 milliseconds.
5201 Amount of gain in delayed signal into right and vice versa. Gives a delay
5202 effect of left signal in right output and vice versa which gives widening
5203 effect. Default is 0.3.
5206 Cross feed of left into right with inverted phase. This helps in suppressing
5207 the mono. If the value is 1 it will cancel all the signal common to both
5208 channels. Default is 0.3.
5211 Set level of input signal of original channel. Default is 0.8.
5214 @subsection Commands
5216 This filter supports the all above options except @code{delay} as @ref{commands}.
5218 @section superequalizer
5219 Apply 18 band equalizer.
5221 The filter accepts the following options:
5228 Set 131Hz band gain.
5230 Set 185Hz band gain.
5232 Set 262Hz band gain.
5234 Set 370Hz band gain.
5236 Set 523Hz band gain.
5238 Set 740Hz band gain.
5240 Set 1047Hz band gain.
5242 Set 1480Hz band gain.
5244 Set 2093Hz band gain.
5246 Set 2960Hz band gain.
5248 Set 4186Hz band gain.
5250 Set 5920Hz band gain.
5252 Set 8372Hz band gain.
5254 Set 11840Hz band gain.
5256 Set 16744Hz band gain.
5258 Set 20000Hz band gain.
5262 Apply audio surround upmix filter.
5264 This filter allows to produce multichannel output from audio stream.
5266 The filter accepts the following options:
5270 Set output channel layout. By default, this is @var{5.1}.
5272 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5273 for the required syntax.
5276 Set input channel layout. By default, this is @var{stereo}.
5278 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5279 for the required syntax.
5282 Set input volume level. By default, this is @var{1}.
5285 Set output volume level. By default, this is @var{1}.
5288 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5291 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5294 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5297 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5298 In @var{add} mode, LFE channel is created from input audio and added to output.
5299 In @var{sub} mode, LFE channel is created from input audio and added to output but
5300 also all non-LFE output channels are subtracted with output LFE channel.
5303 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5304 Default is @var{90}.
5307 Set front center input volume. By default, this is @var{1}.
5310 Set front center output volume. By default, this is @var{1}.
5313 Set front left input volume. By default, this is @var{1}.
5316 Set front left output volume. By default, this is @var{1}.
5319 Set front right input volume. By default, this is @var{1}.
5322 Set front right output volume. By default, this is @var{1}.
5325 Set side left input volume. By default, this is @var{1}.
5328 Set side left output volume. By default, this is @var{1}.
5331 Set side right input volume. By default, this is @var{1}.
5334 Set side right output volume. By default, this is @var{1}.
5337 Set back left input volume. By default, this is @var{1}.
5340 Set back left output volume. By default, this is @var{1}.
5343 Set back right input volume. By default, this is @var{1}.
5346 Set back right output volume. By default, this is @var{1}.
5349 Set back center input volume. By default, this is @var{1}.
5352 Set back center output volume. By default, this is @var{1}.
5355 Set LFE input volume. By default, this is @var{1}.
5358 Set LFE output volume. By default, this is @var{1}.
5361 Set spread usage of stereo image across X axis for all channels.
5364 Set spread usage of stereo image across Y axis for all channels.
5366 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5367 Set spread usage of stereo image across X axis for each channel.
5369 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5370 Set spread usage of stereo image across Y axis for each channel.
5373 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5376 Set window function.
5378 It accepts the following values:
5401 Default is @code{hann}.
5404 Set window overlap. If set to 1, the recommended overlap for selected
5405 window function will be picked. Default is @code{0.5}.
5408 @section treble, highshelf
5410 Boost or cut treble (upper) frequencies of the audio using a two-pole
5411 shelving filter with a response similar to that of a standard
5412 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5414 The filter accepts the following options:
5418 Give the gain at whichever is the lower of ~22 kHz and the
5419 Nyquist frequency. Its useful range is about -20 (for a large cut)
5420 to +20 (for a large boost). Beware of clipping when using a positive gain.
5423 Set the filter's central frequency and so can be used
5424 to extend or reduce the frequency range to be boosted or cut.
5425 The default value is @code{3000} Hz.
5428 Set method to specify band-width of filter.
5443 Determine how steep is the filter's shelf transition.
5446 How much to use filtered signal in output. Default is 1.
5447 Range is between 0 and 1.
5450 Specify which channels to filter, by default all available are filtered.
5453 Normalize biquad coefficients, by default is disabled.
5454 Enabling it will normalize magnitude response at DC to 0dB.
5457 @subsection Commands
5459 This filter supports the following commands:
5462 Change treble frequency.
5463 Syntax for the command is : "@var{frequency}"
5466 Change treble width_type.
5467 Syntax for the command is : "@var{width_type}"
5470 Change treble width.
5471 Syntax for the command is : "@var{width}"
5475 Syntax for the command is : "@var{gain}"
5479 Syntax for the command is : "@var{mix}"
5484 Sinusoidal amplitude modulation.
5486 The filter accepts the following options:
5490 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5491 (20 Hz or lower) will result in a tremolo effect.
5492 This filter may also be used as a ring modulator by specifying
5493 a modulation frequency higher than 20 Hz.
5494 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5497 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5498 Default value is 0.5.
5503 Sinusoidal phase modulation.
5505 The filter accepts the following options:
5509 Modulation frequency in Hertz.
5510 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5513 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5514 Default value is 0.5.
5519 Adjust the input audio volume.
5521 It accepts the following parameters:
5525 Set audio volume expression.
5527 Output values are clipped to the maximum value.
5529 The output audio volume is given by the relation:
5531 @var{output_volume} = @var{volume} * @var{input_volume}
5534 The default value for @var{volume} is "1.0".
5537 This parameter represents the mathematical precision.
5539 It determines which input sample formats will be allowed, which affects the
5540 precision of the volume scaling.
5544 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5546 32-bit floating-point; this limits input sample format to FLT. (default)
5548 64-bit floating-point; this limits input sample format to DBL.
5552 Choose the behaviour on encountering ReplayGain side data in input frames.
5556 Remove ReplayGain side data, ignoring its contents (the default).
5559 Ignore ReplayGain side data, but leave it in the frame.
5562 Prefer the track gain, if present.
5565 Prefer the album gain, if present.
5568 @item replaygain_preamp
5569 Pre-amplification gain in dB to apply to the selected replaygain gain.
5571 Default value for @var{replaygain_preamp} is 0.0.
5573 @item replaygain_noclip
5574 Prevent clipping by limiting the gain applied.
5576 Default value for @var{replaygain_noclip} is 1.
5579 Set when the volume expression is evaluated.
5581 It accepts the following values:
5584 only evaluate expression once during the filter initialization, or
5585 when the @samp{volume} command is sent
5588 evaluate expression for each incoming frame
5591 Default value is @samp{once}.
5594 The volume expression can contain the following parameters.
5598 frame number (starting at zero)
5601 @item nb_consumed_samples
5602 number of samples consumed by the filter
5604 number of samples in the current frame
5606 original frame position in the file
5612 PTS at start of stream
5614 time at start of stream
5620 last set volume value
5623 Note that when @option{eval} is set to @samp{once} only the
5624 @var{sample_rate} and @var{tb} variables are available, all other
5625 variables will evaluate to NAN.
5627 @subsection Commands
5629 This filter supports the following commands:
5632 Modify the volume expression.
5633 The command accepts the same syntax of the corresponding option.
5635 If the specified expression is not valid, it is kept at its current
5639 @subsection Examples
5643 Halve the input audio volume:
5647 volume=volume=-6.0206dB
5650 In all the above example the named key for @option{volume} can be
5651 omitted, for example like in:
5657 Increase input audio power by 6 decibels using fixed-point precision:
5659 volume=volume=6dB:precision=fixed
5663 Fade volume after time 10 with an annihilation period of 5 seconds:
5665 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5669 @section volumedetect
5671 Detect the volume of the input video.
5673 The filter has no parameters. The input is not modified. Statistics about
5674 the volume will be printed in the log when the input stream end is reached.
5676 In particular it will show the mean volume (root mean square), maximum
5677 volume (on a per-sample basis), and the beginning of a histogram of the
5678 registered volume values (from the maximum value to a cumulated 1/1000 of
5681 All volumes are in decibels relative to the maximum PCM value.
5683 @subsection Examples
5685 Here is an excerpt of the output:
5687 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5688 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5689 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5690 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5691 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5692 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5693 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5694 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5695 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5701 The mean square energy is approximately -27 dB, or 10^-2.7.
5703 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5705 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5708 In other words, raising the volume by +4 dB does not cause any clipping,
5709 raising it by +5 dB causes clipping for 6 samples, etc.
5711 @c man end AUDIO FILTERS
5713 @chapter Audio Sources
5714 @c man begin AUDIO SOURCES
5716 Below is a description of the currently available audio sources.
5720 Buffer audio frames, and make them available to the filter chain.
5722 This source is mainly intended for a programmatic use, in particular
5723 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5725 It accepts the following parameters:
5729 The timebase which will be used for timestamps of submitted frames. It must be
5730 either a floating-point number or in @var{numerator}/@var{denominator} form.
5733 The sample rate of the incoming audio buffers.
5736 The sample format of the incoming audio buffers.
5737 Either a sample format name or its corresponding integer representation from
5738 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5740 @item channel_layout
5741 The channel layout of the incoming audio buffers.
5742 Either a channel layout name from channel_layout_map in
5743 @file{libavutil/channel_layout.c} or its corresponding integer representation
5744 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5747 The number of channels of the incoming audio buffers.
5748 If both @var{channels} and @var{channel_layout} are specified, then they
5753 @subsection Examples
5756 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5759 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5760 Since the sample format with name "s16p" corresponds to the number
5761 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5764 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5769 Generate an audio signal specified by an expression.
5771 This source accepts in input one or more expressions (one for each
5772 channel), which are evaluated and used to generate a corresponding
5775 This source accepts the following options:
5779 Set the '|'-separated expressions list for each separate channel. In case the
5780 @option{channel_layout} option is not specified, the selected channel layout
5781 depends on the number of provided expressions. Otherwise the last
5782 specified expression is applied to the remaining output channels.
5784 @item channel_layout, c
5785 Set the channel layout. The number of channels in the specified layout
5786 must be equal to the number of specified expressions.
5789 Set the minimum duration of the sourced audio. See
5790 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5791 for the accepted syntax.
5792 Note that the resulting duration may be greater than the specified
5793 duration, as the generated audio is always cut at the end of a
5796 If not specified, or the expressed duration is negative, the audio is
5797 supposed to be generated forever.
5800 Set the number of samples per channel per each output frame,
5803 @item sample_rate, s
5804 Specify the sample rate, default to 44100.
5807 Each expression in @var{exprs} can contain the following constants:
5811 number of the evaluated sample, starting from 0
5814 time of the evaluated sample expressed in seconds, starting from 0
5821 @subsection Examples
5831 Generate a sin signal with frequency of 440 Hz, set sample rate to
5834 aevalsrc="sin(440*2*PI*t):s=8000"
5838 Generate a two channels signal, specify the channel layout (Front
5839 Center + Back Center) explicitly:
5841 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5845 Generate white noise:
5847 aevalsrc="-2+random(0)"
5851 Generate an amplitude modulated signal:
5853 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5857 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5859 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5866 Generate a FIR coefficients using frequency sampling method.
5868 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
5870 The filter accepts the following options:
5874 Set number of filter coefficents in output audio stream.
5875 Default value is 1025.
5878 Set frequency points from where magnitude and phase are set.
5879 This must be in non decreasing order, and first element must be 0, while last element
5880 must be 1. Elements are separated by white spaces.
5883 Set magnitude value for every frequency point set by @option{frequency}.
5884 Number of values must be same as number of frequency points.
5885 Values are separated by white spaces.
5888 Set phase value for every frequency point set by @option{frequency}.
5889 Number of values must be same as number of frequency points.
5890 Values are separated by white spaces.
5892 @item sample_rate, r
5893 Set sample rate, default is 44100.
5896 Set number of samples per each frame. Default is 1024.
5899 Set window function. Default is blackman.
5904 The null audio source, return unprocessed audio frames. It is mainly useful
5905 as a template and to be employed in analysis / debugging tools, or as
5906 the source for filters which ignore the input data (for example the sox
5909 This source accepts the following options:
5913 @item channel_layout, cl
5915 Specifies the channel layout, and can be either an integer or a string
5916 representing a channel layout. The default value of @var{channel_layout}
5919 Check the channel_layout_map definition in
5920 @file{libavutil/channel_layout.c} for the mapping between strings and
5921 channel layout values.
5923 @item sample_rate, r
5924 Specifies the sample rate, and defaults to 44100.
5927 Set the number of samples per requested frames.
5931 @subsection Examples
5935 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5937 anullsrc=r=48000:cl=4
5941 Do the same operation with a more obvious syntax:
5943 anullsrc=r=48000:cl=mono
5947 All the parameters need to be explicitly defined.
5951 Synthesize a voice utterance using the libflite library.
5953 To enable compilation of this filter you need to configure FFmpeg with
5954 @code{--enable-libflite}.
5956 Note that versions of the flite library prior to 2.0 are not thread-safe.
5958 The filter accepts the following options:
5963 If set to 1, list the names of the available voices and exit
5964 immediately. Default value is 0.
5967 Set the maximum number of samples per frame. Default value is 512.
5970 Set the filename containing the text to speak.
5973 Set the text to speak.
5976 Set the voice to use for the speech synthesis. Default value is
5977 @code{kal}. See also the @var{list_voices} option.
5980 @subsection Examples
5984 Read from file @file{speech.txt}, and synthesize the text using the
5985 standard flite voice:
5987 flite=textfile=speech.txt
5991 Read the specified text selecting the @code{slt} voice:
5993 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5997 Input text to ffmpeg:
5999 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6003 Make @file{ffplay} speak the specified text, using @code{flite} and
6004 the @code{lavfi} device:
6006 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6010 For more information about libflite, check:
6011 @url{http://www.festvox.org/flite/}
6015 Generate a noise audio signal.
6017 The filter accepts the following options:
6020 @item sample_rate, r
6021 Specify the sample rate. Default value is 48000 Hz.
6024 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6028 Specify the duration of the generated audio stream. Not specifying this option
6029 results in noise with an infinite length.
6031 @item color, colour, c
6032 Specify the color of noise. Available noise colors are white, pink, brown,
6033 blue, violet and velvet. Default color is white.
6036 Specify a value used to seed the PRNG.
6039 Set the number of samples per each output frame, default is 1024.
6042 @subsection Examples
6047 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6049 anoisesrc=d=60:c=pink:r=44100:a=0.5
6055 Generate odd-tap Hilbert transform FIR coefficients.
6057 The resulting stream can be used with @ref{afir} filter for phase-shifting
6058 the signal by 90 degrees.
6060 This is used in many matrix coding schemes and for analytic signal generation.
6061 The process is often written as a multiplication by i (or j), the imaginary unit.
6063 The filter accepts the following options:
6067 @item sample_rate, s
6068 Set sample rate, default is 44100.
6071 Set length of FIR filter, default is 22051.
6074 Set number of samples per each frame.
6077 Set window function to be used when generating FIR coefficients.
6082 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6084 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6086 The filter accepts the following options:
6089 @item sample_rate, r
6090 Set sample rate, default is 44100.
6093 Set number of samples per each frame. Default is 1024.
6096 Set high-pass frequency. Default is 0.
6099 Set low-pass frequency. Default is 0.
6100 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6101 is higher than 0 then filter will create band-pass filter coefficients,
6102 otherwise band-reject filter coefficients.
6105 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6108 Set Kaiser window beta.
6111 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6114 Enable rounding, by default is disabled.
6117 Set number of taps for high-pass filter.
6120 Set number of taps for low-pass filter.
6125 Generate an audio signal made of a sine wave with amplitude 1/8.
6127 The audio signal is bit-exact.
6129 The filter accepts the following options:
6134 Set the carrier frequency. Default is 440 Hz.
6136 @item beep_factor, b
6137 Enable a periodic beep every second with frequency @var{beep_factor} times
6138 the carrier frequency. Default is 0, meaning the beep is disabled.
6140 @item sample_rate, r
6141 Specify the sample rate, default is 44100.
6144 Specify the duration of the generated audio stream.
6146 @item samples_per_frame
6147 Set the number of samples per output frame.
6149 The expression can contain the following constants:
6153 The (sequential) number of the output audio frame, starting from 0.
6156 The PTS (Presentation TimeStamp) of the output audio frame,
6157 expressed in @var{TB} units.
6160 The PTS of the output audio frame, expressed in seconds.
6163 The timebase of the output audio frames.
6166 Default is @code{1024}.
6169 @subsection Examples
6174 Generate a simple 440 Hz sine wave:
6180 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6184 sine=frequency=220:beep_factor=4:duration=5
6188 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6191 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6195 @c man end AUDIO SOURCES
6197 @chapter Audio Sinks
6198 @c man begin AUDIO SINKS
6200 Below is a description of the currently available audio sinks.
6202 @section abuffersink
6204 Buffer audio frames, and make them available to the end of filter chain.
6206 This sink is mainly intended for programmatic use, in particular
6207 through the interface defined in @file{libavfilter/buffersink.h}
6208 or the options system.
6210 It accepts a pointer to an AVABufferSinkContext structure, which
6211 defines the incoming buffers' formats, to be passed as the opaque
6212 parameter to @code{avfilter_init_filter} for initialization.
6215 Null audio sink; do absolutely nothing with the input audio. It is
6216 mainly useful as a template and for use in analysis / debugging
6219 @c man end AUDIO SINKS
6221 @chapter Video Filters
6222 @c man begin VIDEO FILTERS
6224 When you configure your FFmpeg build, you can disable any of the
6225 existing filters using @code{--disable-filters}.
6226 The configure output will show the video filters included in your
6229 Below is a description of the currently available video filters.
6233 Mark a region of interest in a video frame.
6235 The frame data is passed through unchanged, but metadata is attached
6236 to the frame indicating regions of interest which can affect the
6237 behaviour of later encoding. Multiple regions can be marked by
6238 applying the filter multiple times.
6242 Region distance in pixels from the left edge of the frame.
6244 Region distance in pixels from the top edge of the frame.
6246 Region width in pixels.
6248 Region height in pixels.
6250 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6251 and may contain the following variables:
6254 Width of the input frame.
6256 Height of the input frame.
6260 Quantisation offset to apply within the region.
6262 This must be a real value in the range -1 to +1. A value of zero
6263 indicates no quality change. A negative value asks for better quality
6264 (less quantisation), while a positive value asks for worse quality
6265 (greater quantisation).
6267 The range is calibrated so that the extreme values indicate the
6268 largest possible offset - if the rest of the frame is encoded with the
6269 worst possible quality, an offset of -1 indicates that this region
6270 should be encoded with the best possible quality anyway. Intermediate
6271 values are then interpolated in some codec-dependent way.
6273 For example, in 10-bit H.264 the quantisation parameter varies between
6274 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6275 this region should be encoded with a QP around one-tenth of the full
6276 range better than the rest of the frame. So, if most of the frame
6277 were to be encoded with a QP of around 30, this region would get a QP
6278 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6279 An extreme value of -1 would indicate that this region should be
6280 encoded with the best possible quality regardless of the treatment of
6281 the rest of the frame - that is, should be encoded at a QP of -12.
6283 If set to true, remove any existing regions of interest marked on the
6284 frame before adding the new one.
6287 @subsection Examples
6291 Mark the centre quarter of the frame as interesting.
6293 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6296 Mark the 100-pixel-wide region on the left edge of the frame as very
6297 uninteresting (to be encoded at much lower quality than the rest of
6300 addroi=0:0:100:ih:+1/5
6304 @section alphaextract
6306 Extract the alpha component from the input as a grayscale video. This
6307 is especially useful with the @var{alphamerge} filter.
6311 Add or replace the alpha component of the primary input with the
6312 grayscale value of a second input. This is intended for use with
6313 @var{alphaextract} to allow the transmission or storage of frame
6314 sequences that have alpha in a format that doesn't support an alpha
6317 For example, to reconstruct full frames from a normal YUV-encoded video
6318 and a separate video created with @var{alphaextract}, you might use:
6320 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6323 Since this filter is designed for reconstruction, it operates on frame
6324 sequences without considering timestamps, and terminates when either
6325 input reaches end of stream. This will cause problems if your encoding
6326 pipeline drops frames. If you're trying to apply an image as an
6327 overlay to a video stream, consider the @var{overlay} filter instead.
6331 Amplify differences between current pixel and pixels of adjacent frames in
6332 same pixel location.
6334 This filter accepts the following options:
6338 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6339 For example radius of 3 will instruct filter to calculate average of 7 frames.
6342 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6345 Set threshold for difference amplification. Any difference greater or equal to
6346 this value will not alter source pixel. Default is 10.
6347 Allowed range is from 0 to 65535.
6350 Set tolerance for difference amplification. Any difference lower to
6351 this value will not alter source pixel. Default is 0.
6352 Allowed range is from 0 to 65535.
6355 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6356 This option controls maximum possible value that will decrease source pixel value.
6359 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6360 This option controls maximum possible value that will increase source pixel value.
6363 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6366 @subsection Commands
6368 This filter supports the following @ref{commands} that corresponds to option of same name:
6380 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6381 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6382 Substation Alpha) subtitles files.
6384 This filter accepts the following option in addition to the common options from
6385 the @ref{subtitles} filter:
6389 Set the shaping engine
6391 Available values are:
6394 The default libass shaping engine, which is the best available.
6396 Fast, font-agnostic shaper that can do only substitutions
6398 Slower shaper using OpenType for substitutions and positioning
6401 The default is @code{auto}.
6405 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6407 The filter accepts the following options:
6411 Set threshold A for 1st plane. Default is 0.02.
6412 Valid range is 0 to 0.3.
6415 Set threshold B for 1st plane. Default is 0.04.
6416 Valid range is 0 to 5.
6419 Set threshold A for 2nd plane. Default is 0.02.
6420 Valid range is 0 to 0.3.
6423 Set threshold B for 2nd plane. Default is 0.04.
6424 Valid range is 0 to 5.
6427 Set threshold A for 3rd plane. Default is 0.02.
6428 Valid range is 0 to 0.3.
6431 Set threshold B for 3rd plane. Default is 0.04.
6432 Valid range is 0 to 5.
6434 Threshold A is designed to react on abrupt changes in the input signal and
6435 threshold B is designed to react on continuous changes in the input signal.
6438 Set number of frames filter will use for averaging. Default is 9. Must be odd
6439 number in range [5, 129].
6442 Set what planes of frame filter will use for averaging. Default is all.
6445 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
6446 Alternatively can be set to @code{s} serial.
6448 Parallel can be faster then serial, while other way around is never true.
6449 Parallel will abort early on first change being greater then thresholds, while serial
6450 will continue processing other side of frames if they are equal or bellow thresholds.
6453 @subsection Commands
6454 This filter supports same @ref{commands} as options except option @code{s}.
6455 The command accepts the same syntax of the corresponding option.
6459 Apply average blur filter.
6461 The filter accepts the following options:
6465 Set horizontal radius size.
6468 Set which planes to filter. By default all planes are filtered.
6471 Set vertical radius size, if zero it will be same as @code{sizeX}.
6472 Default is @code{0}.
6475 @subsection Commands
6476 This filter supports same commands as options.
6477 The command accepts the same syntax of the corresponding option.
6479 If the specified expression is not valid, it is kept at its current
6484 Compute the bounding box for the non-black pixels in the input frame
6487 This filter computes the bounding box containing all the pixels with a
6488 luminance value greater than the minimum allowed value.
6489 The parameters describing the bounding box are printed on the filter
6492 The filter accepts the following option:
6496 Set the minimal luminance value. Default is @code{16}.
6500 Apply bilateral filter, spatial smoothing while preserving edges.
6502 The filter accepts the following options:
6505 Set sigma of gaussian function to calculate spatial weight.
6506 Allowed range is 0 to 10. Default is 0.1.
6509 Set sigma of gaussian function to calculate range weight.
6510 Allowed range is 0 to 1. Default is 0.1.
6513 Set planes to filter. Default is first only.
6516 @section bitplanenoise
6518 Show and measure bit plane noise.
6520 The filter accepts the following options:
6524 Set which plane to analyze. Default is @code{1}.
6527 Filter out noisy pixels from @code{bitplane} set above.
6528 Default is disabled.
6531 @section blackdetect
6533 Detect video intervals that are (almost) completely black. Can be
6534 useful to detect chapter transitions, commercials, or invalid
6535 recordings. Output lines contains the time for the start, end and
6536 duration of the detected black interval expressed in seconds.
6538 In order to display the output lines, you need to set the loglevel at
6539 least to the AV_LOG_INFO value.
6541 The filter accepts the following options:
6544 @item black_min_duration, d
6545 Set the minimum detected black duration expressed in seconds. It must
6546 be a non-negative floating point number.
6548 Default value is 2.0.
6550 @item picture_black_ratio_th, pic_th
6551 Set the threshold for considering a picture "black".
6552 Express the minimum value for the ratio:
6554 @var{nb_black_pixels} / @var{nb_pixels}
6557 for which a picture is considered black.
6558 Default value is 0.98.
6560 @item pixel_black_th, pix_th
6561 Set the threshold for considering a pixel "black".
6563 The threshold expresses the maximum pixel luminance value for which a
6564 pixel is considered "black". The provided value is scaled according to
6565 the following equation:
6567 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6570 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6571 the input video format, the range is [0-255] for YUV full-range
6572 formats and [16-235] for YUV non full-range formats.
6574 Default value is 0.10.
6577 The following example sets the maximum pixel threshold to the minimum
6578 value, and detects only black intervals of 2 or more seconds:
6580 blackdetect=d=2:pix_th=0.00
6585 Detect frames that are (almost) completely black. Can be useful to
6586 detect chapter transitions or commercials. Output lines consist of
6587 the frame number of the detected frame, the percentage of blackness,
6588 the position in the file if known or -1 and the timestamp in seconds.
6590 In order to display the output lines, you need to set the loglevel at
6591 least to the AV_LOG_INFO value.
6593 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6594 The value represents the percentage of pixels in the picture that
6595 are below the threshold value.
6597 It accepts the following parameters:
6602 The percentage of the pixels that have to be below the threshold; it defaults to
6605 @item threshold, thresh
6606 The threshold below which a pixel value is considered black; it defaults to
6611 @section blend, tblend
6613 Blend two video frames into each other.
6615 The @code{blend} filter takes two input streams and outputs one
6616 stream, the first input is the "top" layer and second input is
6617 "bottom" layer. By default, the output terminates when the longest input terminates.
6619 The @code{tblend} (time blend) filter takes two consecutive frames
6620 from one single stream, and outputs the result obtained by blending
6621 the new frame on top of the old frame.
6623 A description of the accepted options follows.
6631 Set blend mode for specific pixel component or all pixel components in case
6632 of @var{all_mode}. Default value is @code{normal}.
6634 Available values for component modes are:
6676 Set blend opacity for specific pixel component or all pixel components in case
6677 of @var{all_opacity}. Only used in combination with pixel component blend modes.
6684 Set blend expression for specific pixel component or all pixel components in case
6685 of @var{all_expr}. Note that related mode options will be ignored if those are set.
6687 The expressions can use the following variables:
6691 The sequential number of the filtered frame, starting from @code{0}.
6695 the coordinates of the current sample
6699 the width and height of currently filtered plane
6703 Width and height scale for the plane being filtered. It is the
6704 ratio between the dimensions of the current plane to the luma plane,
6705 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
6706 the luma plane and @code{0.5,0.5} for the chroma planes.
6709 Time of the current frame, expressed in seconds.
6712 Value of pixel component at current location for first video frame (top layer).
6715 Value of pixel component at current location for second video frame (bottom layer).
6719 The @code{blend} filter also supports the @ref{framesync} options.
6721 @subsection Examples
6725 Apply transition from bottom layer to top layer in first 10 seconds:
6727 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
6731 Apply linear horizontal transition from top layer to bottom layer:
6733 blend=all_expr='A*(X/W)+B*(1-X/W)'
6737 Apply 1x1 checkerboard effect:
6739 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
6743 Apply uncover left effect:
6745 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6749 Apply uncover down effect:
6751 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6755 Apply uncover up-left effect:
6757 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6761 Split diagonally video and shows top and bottom layer on each side:
6763 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6767 Display differences between the current and the previous frame:
6769 tblend=all_mode=grainextract
6775 Denoise frames using Block-Matching 3D algorithm.
6777 The filter accepts the following options.
6781 Set denoising strength. Default value is 1.
6782 Allowed range is from 0 to 999.9.
6783 The denoising algorithm is very sensitive to sigma, so adjust it
6784 according to the source.
6787 Set local patch size. This sets dimensions in 2D.
6790 Set sliding step for processing blocks. Default value is 4.
6791 Allowed range is from 1 to 64.
6792 Smaller values allows processing more reference blocks and is slower.
6795 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6796 When set to 1, no block matching is done. Larger values allows more blocks
6798 Allowed range is from 1 to 256.
6801 Set radius for search block matching. Default is 9.
6802 Allowed range is from 1 to INT32_MAX.
6805 Set step between two search locations for block matching. Default is 1.
6806 Allowed range is from 1 to 64. Smaller is slower.
6809 Set threshold of mean square error for block matching. Valid range is 0 to
6813 Set thresholding parameter for hard thresholding in 3D transformed domain.
6814 Larger values results in stronger hard-thresholding filtering in frequency
6818 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6819 Default is @code{basic}.
6822 If enabled, filter will use 2nd stream for block matching.
6823 Default is disabled for @code{basic} value of @var{estim} option,
6824 and always enabled if value of @var{estim} is @code{final}.
6827 Set planes to filter. Default is all available except alpha.
6830 @subsection Examples
6834 Basic filtering with bm3d:
6836 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
6840 Same as above, but filtering only luma:
6842 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
6846 Same as above, but with both estimation modes:
6848 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
6852 Same as above, but prefilter with @ref{nlmeans} filter instead:
6854 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
6860 Apply a boxblur algorithm to the input video.
6862 It accepts the following parameters:
6866 @item luma_radius, lr
6867 @item luma_power, lp
6868 @item chroma_radius, cr
6869 @item chroma_power, cp
6870 @item alpha_radius, ar
6871 @item alpha_power, ap
6875 A description of the accepted options follows.
6878 @item luma_radius, lr
6879 @item chroma_radius, cr
6880 @item alpha_radius, ar
6881 Set an expression for the box radius in pixels used for blurring the
6882 corresponding input plane.
6884 The radius value must be a non-negative number, and must not be
6885 greater than the value of the expression @code{min(w,h)/2} for the
6886 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
6889 Default value for @option{luma_radius} is "2". If not specified,
6890 @option{chroma_radius} and @option{alpha_radius} default to the
6891 corresponding value set for @option{luma_radius}.
6893 The expressions can contain the following constants:
6897 The input width and height in pixels.
6901 The input chroma image width and height in pixels.
6905 The horizontal and vertical chroma subsample values. For example, for the
6906 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6909 @item luma_power, lp
6910 @item chroma_power, cp
6911 @item alpha_power, ap
6912 Specify how many times the boxblur filter is applied to the
6913 corresponding plane.
6915 Default value for @option{luma_power} is 2. If not specified,
6916 @option{chroma_power} and @option{alpha_power} default to the
6917 corresponding value set for @option{luma_power}.
6919 A value of 0 will disable the effect.
6922 @subsection Examples
6926 Apply a boxblur filter with the luma, chroma, and alpha radii
6929 boxblur=luma_radius=2:luma_power=1
6934 Set the luma radius to 2, and alpha and chroma radius to 0:
6936 boxblur=2:1:cr=0:ar=0
6940 Set the luma and chroma radii to a fraction of the video dimension:
6942 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6948 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6949 Deinterlacing Filter").
6951 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6952 interpolation algorithms.
6953 It accepts the following parameters:
6957 The interlacing mode to adopt. It accepts one of the following values:
6961 Output one frame for each frame.
6963 Output one frame for each field.
6966 The default value is @code{send_field}.
6969 The picture field parity assumed for the input interlaced video. It accepts one
6970 of the following values:
6974 Assume the top field is first.
6976 Assume the bottom field is first.
6978 Enable automatic detection of field parity.
6981 The default value is @code{auto}.
6982 If the interlacing is unknown or the decoder does not export this information,
6983 top field first will be assumed.
6986 Specify which frames to deinterlace. Accepts one of the following
6991 Deinterlace all frames.
6993 Only deinterlace frames marked as interlaced.
6996 The default value is @code{all}.
7001 Apply Contrast Adaptive Sharpen filter to video stream.
7003 The filter accepts the following options:
7007 Set the sharpening strength. Default value is 0.
7010 Set planes to filter. Default value is to filter all
7011 planes except alpha plane.
7015 Remove all color information for all colors except for certain one.
7017 The filter accepts the following options:
7021 The color which will not be replaced with neutral chroma.
7024 Similarity percentage with the above color.
7025 0.01 matches only the exact key color, while 1.0 matches everything.
7029 0.0 makes pixels either fully gray, or not gray at all.
7030 Higher values result in more preserved color.
7033 Signals that the color passed is already in YUV instead of RGB.
7035 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7036 This can be used to pass exact YUV values as hexadecimal numbers.
7039 @subsection Commands
7040 This filter supports same @ref{commands} as options.
7041 The command accepts the same syntax of the corresponding option.
7043 If the specified expression is not valid, it is kept at its current
7047 YUV colorspace color/chroma keying.
7049 The filter accepts the following options:
7053 The color which will be replaced with transparency.
7056 Similarity percentage with the key color.
7058 0.01 matches only the exact key color, while 1.0 matches everything.
7063 0.0 makes pixels either fully transparent, or not transparent at all.
7065 Higher values result in semi-transparent pixels, with a higher transparency
7066 the more similar the pixels color is to the key color.
7069 Signals that the color passed is already in YUV instead of RGB.
7071 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7072 This can be used to pass exact YUV values as hexadecimal numbers.
7075 @subsection Commands
7076 This filter supports same @ref{commands} as options.
7077 The command accepts the same syntax of the corresponding option.
7079 If the specified expression is not valid, it is kept at its current
7082 @subsection Examples
7086 Make every green pixel in the input image transparent:
7088 ffmpeg -i input.png -vf chromakey=green out.png
7092 Overlay a greenscreen-video on top of a static black background.
7094 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
7098 @section chromashift
7099 Shift chroma pixels horizontally and/or vertically.
7101 The filter accepts the following options:
7104 Set amount to shift chroma-blue horizontally.
7106 Set amount to shift chroma-blue vertically.
7108 Set amount to shift chroma-red horizontally.
7110 Set amount to shift chroma-red vertically.
7112 Set edge mode, can be @var{smear}, default, or @var{warp}.
7115 @subsection Commands
7117 This filter supports the all above options as @ref{commands}.
7121 Display CIE color diagram with pixels overlaid onto it.
7123 The filter accepts the following options:
7138 @item uhdtv, rec2020
7152 Set what gamuts to draw.
7154 See @code{system} option for available values.
7157 Set ciescope size, by default set to 512.
7160 Set intensity used to map input pixel values to CIE diagram.
7163 Set contrast used to draw tongue colors that are out of active color system gamut.
7166 Correct gamma displayed on scope, by default enabled.
7169 Show white point on CIE diagram, by default disabled.
7172 Set input gamma. Used only with XYZ input color space.
7177 Visualize information exported by some codecs.
7179 Some codecs can export information through frames using side-data or other
7180 means. For example, some MPEG based codecs export motion vectors through the
7181 @var{export_mvs} flag in the codec @option{flags2} option.
7183 The filter accepts the following option:
7187 Set motion vectors to visualize.
7189 Available flags for @var{mv} are:
7193 forward predicted MVs of P-frames
7195 forward predicted MVs of B-frames
7197 backward predicted MVs of B-frames
7201 Display quantization parameters using the chroma planes.
7204 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7206 Available flags for @var{mv_type} are:
7210 forward predicted MVs
7212 backward predicted MVs
7215 @item frame_type, ft
7216 Set frame type to visualize motion vectors of.
7218 Available flags for @var{frame_type} are:
7222 intra-coded frames (I-frames)
7224 predicted frames (P-frames)
7226 bi-directionally predicted frames (B-frames)
7230 @subsection Examples
7234 Visualize forward predicted MVs of all frames using @command{ffplay}:
7236 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7240 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7242 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7246 @section colorbalance
7247 Modify intensity of primary colors (red, green and blue) of input frames.
7249 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7250 regions for the red-cyan, green-magenta or blue-yellow balance.
7252 A positive adjustment value shifts the balance towards the primary color, a negative
7253 value towards the complementary color.
7255 The filter accepts the following options:
7261 Adjust red, green and blue shadows (darkest pixels).
7266 Adjust red, green and blue midtones (medium pixels).
7271 Adjust red, green and blue highlights (brightest pixels).
7273 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7276 Preserve lightness when changing color balance. Default is disabled.
7279 @subsection Examples
7283 Add red color cast to shadows:
7289 @subsection Commands
7291 This filter supports the all above options as @ref{commands}.
7293 @section colorchannelmixer
7295 Adjust video input frames by re-mixing color channels.
7297 This filter modifies a color channel by adding the values associated to
7298 the other channels of the same pixels. For example if the value to
7299 modify is red, the output value will be:
7301 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7304 The filter accepts the following options:
7311 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7312 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7318 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7319 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7325 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7326 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7332 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7333 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7335 Allowed ranges for options are @code{[-2.0, 2.0]}.
7338 @subsection Examples
7342 Convert source to grayscale:
7344 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7347 Simulate sepia tones:
7349 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7353 @subsection Commands
7355 This filter supports the all above options as @ref{commands}.
7358 RGB colorspace color keying.
7360 The filter accepts the following options:
7364 The color which will be replaced with transparency.
7367 Similarity percentage with the key color.
7369 0.01 matches only the exact key color, while 1.0 matches everything.
7374 0.0 makes pixels either fully transparent, or not transparent at all.
7376 Higher values result in semi-transparent pixels, with a higher transparency
7377 the more similar the pixels color is to the key color.
7380 @subsection Examples
7384 Make every green pixel in the input image transparent:
7386 ffmpeg -i input.png -vf colorkey=green out.png
7390 Overlay a greenscreen-video on top of a static background image.
7392 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
7396 @subsection Commands
7397 This filter supports same @ref{commands} as options.
7398 The command accepts the same syntax of the corresponding option.
7400 If the specified expression is not valid, it is kept at its current
7404 Remove all color information for all RGB colors except for certain one.
7406 The filter accepts the following options:
7410 The color which will not be replaced with neutral gray.
7413 Similarity percentage with the above color.
7414 0.01 matches only the exact key color, while 1.0 matches everything.
7417 Blend percentage. 0.0 makes pixels fully gray.
7418 Higher values result in more preserved color.
7421 @subsection Commands
7422 This filter supports same @ref{commands} as options.
7423 The command accepts the same syntax of the corresponding option.
7425 If the specified expression is not valid, it is kept at its current
7428 @section colorlevels
7430 Adjust video input frames using levels.
7432 The filter accepts the following options:
7439 Adjust red, green, blue and alpha input black point.
7440 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7446 Adjust red, green, blue and alpha input white point.
7447 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7449 Input levels are used to lighten highlights (bright tones), darken shadows
7450 (dark tones), change the balance of bright and dark tones.
7456 Adjust red, green, blue and alpha output black point.
7457 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7463 Adjust red, green, blue and alpha output white point.
7464 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7466 Output levels allows manual selection of a constrained output level range.
7469 @subsection Examples
7473 Make video output darker:
7475 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7481 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7485 Make video output lighter:
7487 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7491 Increase brightness:
7493 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7497 @subsection Commands
7499 This filter supports the all above options as @ref{commands}.
7501 @section colormatrix
7503 Convert color matrix.
7505 The filter accepts the following options:
7510 Specify the source and destination color matrix. Both values must be
7513 The accepted values are:
7541 For example to convert from BT.601 to SMPTE-240M, use the command:
7543 colormatrix=bt601:smpte240m
7548 Convert colorspace, transfer characteristics or color primaries.
7549 Input video needs to have an even size.
7551 The filter accepts the following options:
7556 Specify all color properties at once.
7558 The accepted values are:
7588 Specify output colorspace.
7590 The accepted values are:
7599 BT.470BG or BT.601-6 625
7602 SMPTE-170M or BT.601-6 525
7611 BT.2020 with non-constant luminance
7617 Specify output transfer characteristics.
7619 The accepted values are:
7631 Constant gamma of 2.2
7634 Constant gamma of 2.8
7637 SMPTE-170M, BT.601-6 625 or BT.601-6 525
7655 BT.2020 for 10-bits content
7658 BT.2020 for 12-bits content
7664 Specify output color primaries.
7666 The accepted values are:
7675 BT.470BG or BT.601-6 625
7678 SMPTE-170M or BT.601-6 525
7702 Specify output color range.
7704 The accepted values are:
7707 TV (restricted) range
7710 MPEG (restricted) range
7721 Specify output color format.
7723 The accepted values are:
7726 YUV 4:2:0 planar 8-bits
7729 YUV 4:2:0 planar 10-bits
7732 YUV 4:2:0 planar 12-bits
7735 YUV 4:2:2 planar 8-bits
7738 YUV 4:2:2 planar 10-bits
7741 YUV 4:2:2 planar 12-bits
7744 YUV 4:4:4 planar 8-bits
7747 YUV 4:4:4 planar 10-bits
7750 YUV 4:4:4 planar 12-bits
7755 Do a fast conversion, which skips gamma/primary correction. This will take
7756 significantly less CPU, but will be mathematically incorrect. To get output
7757 compatible with that produced by the colormatrix filter, use fast=1.
7760 Specify dithering mode.
7762 The accepted values are:
7768 Floyd-Steinberg dithering
7772 Whitepoint adaptation mode.
7774 The accepted values are:
7777 Bradford whitepoint adaptation
7780 von Kries whitepoint adaptation
7783 identity whitepoint adaptation (i.e. no whitepoint adaptation)
7787 Override all input properties at once. Same accepted values as @ref{all}.
7790 Override input colorspace. Same accepted values as @ref{space}.
7793 Override input color primaries. Same accepted values as @ref{primaries}.
7796 Override input transfer characteristics. Same accepted values as @ref{trc}.
7799 Override input color range. Same accepted values as @ref{range}.
7803 The filter converts the transfer characteristics, color space and color
7804 primaries to the specified user values. The output value, if not specified,
7805 is set to a default value based on the "all" property. If that property is
7806 also not specified, the filter will log an error. The output color range and
7807 format default to the same value as the input color range and format. The
7808 input transfer characteristics, color space, color primaries and color range
7809 should be set on the input data. If any of these are missing, the filter will
7810 log an error and no conversion will take place.
7812 For example to convert the input to SMPTE-240M, use the command:
7814 colorspace=smpte240m
7817 @section convolution
7819 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
7821 The filter accepts the following options:
7828 Set matrix for each plane.
7829 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
7830 and from 1 to 49 odd number of signed integers in @var{row} mode.
7836 Set multiplier for calculated value for each plane.
7837 If unset or 0, it will be sum of all matrix elements.
7843 Set bias for each plane. This value is added to the result of the multiplication.
7844 Useful for making the overall image brighter or darker. Default is 0.0.
7850 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
7851 Default is @var{square}.
7854 @subsection Examples
7860 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"
7866 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"
7872 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"
7878 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"
7882 Apply laplacian edge detector which includes diagonals:
7884 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"
7890 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"
7896 Apply 2D convolution of video stream in frequency domain using second stream
7899 The filter accepts the following options:
7903 Set which planes to process.
7906 Set which impulse video frames will be processed, can be @var{first}
7907 or @var{all}. Default is @var{all}.
7910 The @code{convolve} filter also supports the @ref{framesync} options.
7914 Copy the input video source unchanged to the output. This is mainly useful for
7919 Video filtering on GPU using Apple's CoreImage API on OSX.
7921 Hardware acceleration is based on an OpenGL context. Usually, this means it is
7922 processed by video hardware. However, software-based OpenGL implementations
7923 exist which means there is no guarantee for hardware processing. It depends on
7926 There are many filters and image generators provided by Apple that come with a
7927 large variety of options. The filter has to be referenced by its name along
7930 The coreimage filter accepts the following options:
7933 List all available filters and generators along with all their respective
7934 options as well as possible minimum and maximum values along with the default
7941 Specify all filters by their respective name and options.
7942 Use @var{list_filters} to determine all valid filter names and options.
7943 Numerical options are specified by a float value and are automatically clamped
7944 to their respective value range. Vector and color options have to be specified
7945 by a list of space separated float values. Character escaping has to be done.
7946 A special option name @code{default} is available to use default options for a
7949 It is required to specify either @code{default} or at least one of the filter options.
7950 All omitted options are used with their default values.
7951 The syntax of the filter string is as follows:
7953 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
7957 Specify a rectangle where the output of the filter chain is copied into the
7958 input image. It is given by a list of space separated float values:
7960 output_rect=x\ y\ width\ height
7962 If not given, the output rectangle equals the dimensions of the input image.
7963 The output rectangle is automatically cropped at the borders of the input
7964 image. Negative values are valid for each component.
7966 output_rect=25\ 25\ 100\ 100
7970 Several filters can be chained for successive processing without GPU-HOST
7971 transfers allowing for fast processing of complex filter chains.
7972 Currently, only filters with zero (generators) or exactly one (filters) input
7973 image and one output image are supported. Also, transition filters are not yet
7976 Some filters generate output images with additional padding depending on the
7977 respective filter kernel. The padding is automatically removed to ensure the
7978 filter output has the same size as the input image.
7980 For image generators, the size of the output image is determined by the
7981 previous output image of the filter chain or the input image of the whole
7982 filterchain, respectively. The generators do not use the pixel information of
7983 this image to generate their output. However, the generated output is
7984 blended onto this image, resulting in partial or complete coverage of the
7987 The @ref{coreimagesrc} video source can be used for generating input images
7988 which are directly fed into the filter chain. By using it, providing input
7989 images by another video source or an input video is not required.
7991 @subsection Examples
7996 List all filters available:
7998 coreimage=list_filters=true
8002 Use the CIBoxBlur filter with default options to blur an image:
8004 coreimage=filter=CIBoxBlur@@default
8008 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8009 its center at 100x100 and a radius of 50 pixels:
8011 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8015 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8016 given as complete and escaped command-line for Apple's standard bash shell:
8018 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8024 Cover a rectangular object
8026 It accepts the following options:
8030 Filepath of the optional cover image, needs to be in yuv420.
8035 It accepts the following values:
8038 cover it by the supplied image
8040 cover it by interpolating the surrounding pixels
8043 Default value is @var{blur}.
8046 @subsection Examples
8050 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8052 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8058 Crop the input video to given dimensions.
8060 It accepts the following parameters:
8064 The width of the output video. It defaults to @code{iw}.
8065 This expression is evaluated only once during the filter
8066 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8069 The height of the output video. It defaults to @code{ih}.
8070 This expression is evaluated only once during the filter
8071 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8074 The horizontal position, in the input video, of the left edge of the output
8075 video. It defaults to @code{(in_w-out_w)/2}.
8076 This expression is evaluated per-frame.
8079 The vertical position, in the input video, of the top edge of the output video.
8080 It defaults to @code{(in_h-out_h)/2}.
8081 This expression is evaluated per-frame.
8084 If set to 1 will force the output display aspect ratio
8085 to be the same of the input, by changing the output sample aspect
8086 ratio. It defaults to 0.
8089 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8090 width/height/x/y as specified and will not be rounded to nearest smaller value.
8094 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8095 expressions containing the following constants:
8100 The computed values for @var{x} and @var{y}. They are evaluated for
8105 The input width and height.
8109 These are the same as @var{in_w} and @var{in_h}.
8113 The output (cropped) width and height.
8117 These are the same as @var{out_w} and @var{out_h}.
8120 same as @var{iw} / @var{ih}
8123 input sample aspect ratio
8126 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8130 horizontal and vertical chroma subsample values. For example for the
8131 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8134 The number of the input frame, starting from 0.
8137 the position in the file of the input frame, NAN if unknown
8140 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8144 The expression for @var{out_w} may depend on the value of @var{out_h},
8145 and the expression for @var{out_h} may depend on @var{out_w}, but they
8146 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8147 evaluated after @var{out_w} and @var{out_h}.
8149 The @var{x} and @var{y} parameters specify the expressions for the
8150 position of the top-left corner of the output (non-cropped) area. They
8151 are evaluated for each frame. If the evaluated value is not valid, it
8152 is approximated to the nearest valid value.
8154 The expression for @var{x} may depend on @var{y}, and the expression
8155 for @var{y} may depend on @var{x}.
8157 @subsection Examples
8161 Crop area with size 100x100 at position (12,34).
8166 Using named options, the example above becomes:
8168 crop=w=100:h=100:x=12:y=34
8172 Crop the central input area with size 100x100:
8178 Crop the central input area with size 2/3 of the input video:
8180 crop=2/3*in_w:2/3*in_h
8184 Crop the input video central square:
8191 Delimit the rectangle with the top-left corner placed at position
8192 100:100 and the right-bottom corner corresponding to the right-bottom
8193 corner of the input image.
8195 crop=in_w-100:in_h-100:100:100
8199 Crop 10 pixels from the left and right borders, and 20 pixels from
8200 the top and bottom borders
8202 crop=in_w-2*10:in_h-2*20
8206 Keep only the bottom right quarter of the input image:
8208 crop=in_w/2:in_h/2:in_w/2:in_h/2
8212 Crop height for getting Greek harmony:
8214 crop=in_w:1/PHI*in_w
8218 Apply trembling effect:
8220 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)
8224 Apply erratic camera effect depending on timestamp:
8226 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)"
8230 Set x depending on the value of y:
8232 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8236 @subsection Commands
8238 This filter supports the following commands:
8244 Set width/height of the output video and the horizontal/vertical position
8246 The command accepts the same syntax of the corresponding option.
8248 If the specified expression is not valid, it is kept at its current
8254 Auto-detect the crop size.
8256 It calculates the necessary cropping parameters and prints the
8257 recommended parameters via the logging system. The detected dimensions
8258 correspond to the non-black area of the input video.
8260 It accepts the following parameters:
8265 Set higher black value threshold, which can be optionally specified
8266 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8267 value greater to the set value is considered non-black. It defaults to 24.
8268 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8269 on the bitdepth of the pixel format.
8272 The value which the width/height should be divisible by. It defaults to
8273 16. The offset is automatically adjusted to center the video. Use 2 to
8274 get only even dimensions (needed for 4:2:2 video). 16 is best when
8275 encoding to most video codecs.
8277 @item reset_count, reset
8278 Set the counter that determines after how many frames cropdetect will
8279 reset the previously detected largest video area and start over to
8280 detect the current optimal crop area. Default value is 0.
8282 This can be useful when channel logos distort the video area. 0
8283 indicates 'never reset', and returns the largest area encountered during
8290 Delay video filtering until a given wallclock timestamp. The filter first
8291 passes on @option{preroll} amount of frames, then it buffers at most
8292 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8293 it forwards the buffered frames and also any subsequent frames coming in its
8296 The filter can be used synchronize the output of multiple ffmpeg processes for
8297 realtime output devices like decklink. By putting the delay in the filtering
8298 chain and pre-buffering frames the process can pass on data to output almost
8299 immediately after the target wallclock timestamp is reached.
8301 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8307 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8310 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8313 The maximum duration of content to buffer before waiting for the cue expressed
8314 in seconds. Default is 0.
8321 Apply color adjustments using curves.
8323 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8324 component (red, green and blue) has its values defined by @var{N} key points
8325 tied from each other using a smooth curve. The x-axis represents the pixel
8326 values from the input frame, and the y-axis the new pixel values to be set for
8329 By default, a component curve is defined by the two points @var{(0;0)} and
8330 @var{(1;1)}. This creates a straight line where each original pixel value is
8331 "adjusted" to its own value, which means no change to the image.
8333 The filter allows you to redefine these two points and add some more. A new
8334 curve (using a natural cubic spline interpolation) will be define to pass
8335 smoothly through all these new coordinates. The new defined points needs to be
8336 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8337 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8338 the vector spaces, the values will be clipped accordingly.
8340 The filter accepts the following options:
8344 Select one of the available color presets. This option can be used in addition
8345 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8346 options takes priority on the preset values.
8347 Available presets are:
8350 @item color_negative
8353 @item increase_contrast
8355 @item linear_contrast
8356 @item medium_contrast
8358 @item strong_contrast
8361 Default is @code{none}.
8363 Set the master key points. These points will define a second pass mapping. It
8364 is sometimes called a "luminance" or "value" mapping. It can be used with
8365 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8366 post-processing LUT.
8368 Set the key points for the red component.
8370 Set the key points for the green component.
8372 Set the key points for the blue component.
8374 Set the key points for all components (not including master).
8375 Can be used in addition to the other key points component
8376 options. In this case, the unset component(s) will fallback on this
8377 @option{all} setting.
8379 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
8381 Save Gnuplot script of the curves in specified file.
8384 To avoid some filtergraph syntax conflicts, each key points list need to be
8385 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
8387 @subsection Examples
8391 Increase slightly the middle level of blue:
8393 curves=blue='0/0 0.5/0.58 1/1'
8399 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'
8401 Here we obtain the following coordinates for each components:
8404 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8406 @code{(0;0) (0.50;0.48) (1;1)}
8408 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8412 The previous example can also be achieved with the associated built-in preset:
8414 curves=preset=vintage
8424 Use a Photoshop preset and redefine the points of the green component:
8426 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8430 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8431 and @command{gnuplot}:
8433 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8434 gnuplot -p /tmp/curves.plt
8440 Video data analysis filter.
8442 This filter shows hexadecimal pixel values of part of video.
8444 The filter accepts the following options:
8448 Set output video size.
8451 Set x offset from where to pick pixels.
8454 Set y offset from where to pick pixels.
8457 Set scope mode, can be one of the following:
8460 Draw hexadecimal pixel values with white color on black background.
8463 Draw hexadecimal pixel values with input video pixel color on black
8467 Draw hexadecimal pixel values on color background picked from input video,
8468 the text color is picked in such way so its always visible.
8472 Draw rows and columns numbers on left and top of video.
8475 Set background opacity.
8478 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
8483 Denoise frames using 2D DCT (frequency domain filtering).
8485 This filter is not designed for real time.
8487 The filter accepts the following options:
8491 Set the noise sigma constant.
8493 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8494 coefficient (absolute value) below this threshold with be dropped.
8496 If you need a more advanced filtering, see @option{expr}.
8498 Default is @code{0}.
8501 Set number overlapping pixels for each block. Since the filter can be slow, you
8502 may want to reduce this value, at the cost of a less effective filter and the
8503 risk of various artefacts.
8505 If the overlapping value doesn't permit processing the whole input width or
8506 height, a warning will be displayed and according borders won't be denoised.
8508 Default value is @var{blocksize}-1, which is the best possible setting.
8511 Set the coefficient factor expression.
8513 For each coefficient of a DCT block, this expression will be evaluated as a
8514 multiplier value for the coefficient.
8516 If this is option is set, the @option{sigma} option will be ignored.
8518 The absolute value of the coefficient can be accessed through the @var{c}
8522 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8523 @var{blocksize}, which is the width and height of the processed blocks.
8525 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8526 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8527 on the speed processing. Also, a larger block size does not necessarily means a
8531 @subsection Examples
8533 Apply a denoise with a @option{sigma} of @code{4.5}:
8538 The same operation can be achieved using the expression system:
8540 dctdnoiz=e='gte(c, 4.5*3)'
8543 Violent denoise using a block size of @code{16x16}:
8550 Remove banding artifacts from input video.
8551 It works by replacing banded pixels with average value of referenced pixels.
8553 The filter accepts the following options:
8560 Set banding detection threshold for each plane. Default is 0.02.
8561 Valid range is 0.00003 to 0.5.
8562 If difference between current pixel and reference pixel is less than threshold,
8563 it will be considered as banded.
8566 Banding detection range in pixels. Default is 16. If positive, random number
8567 in range 0 to set value will be used. If negative, exact absolute value
8569 The range defines square of four pixels around current pixel.
8572 Set direction in radians from which four pixel will be compared. If positive,
8573 random direction from 0 to set direction will be picked. If negative, exact of
8574 absolute value will be picked. For example direction 0, -PI or -2*PI radians
8575 will pick only pixels on same row and -PI/2 will pick only pixels on same
8579 If enabled, current pixel is compared with average value of all four
8580 surrounding pixels. The default is enabled. If disabled current pixel is
8581 compared with all four surrounding pixels. The pixel is considered banded
8582 if only all four differences with surrounding pixels are less than threshold.
8585 If enabled, current pixel is changed if and only if all pixel components are banded,
8586 e.g. banding detection threshold is triggered for all color components.
8587 The default is disabled.
8592 Remove blocking artifacts from input video.
8594 The filter accepts the following options:
8598 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
8599 This controls what kind of deblocking is applied.
8602 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
8608 Set blocking detection thresholds. Allowed range is 0 to 1.
8609 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
8610 Using higher threshold gives more deblocking strength.
8611 Setting @var{alpha} controls threshold detection at exact edge of block.
8612 Remaining options controls threshold detection near the edge. Each one for
8613 below/above or left/right. Setting any of those to @var{0} disables
8617 Set planes to filter. Default is to filter all available planes.
8620 @subsection Examples
8624 Deblock using weak filter and block size of 4 pixels.
8626 deblock=filter=weak:block=4
8630 Deblock using strong filter, block size of 4 pixels and custom thresholds for
8631 deblocking more edges.
8633 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
8637 Similar as above, but filter only first plane.
8639 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
8643 Similar as above, but filter only second and third plane.
8645 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
8652 Drop duplicated frames at regular intervals.
8654 The filter accepts the following options:
8658 Set the number of frames from which one will be dropped. Setting this to
8659 @var{N} means one frame in every batch of @var{N} frames will be dropped.
8660 Default is @code{5}.
8663 Set the threshold for duplicate detection. If the difference metric for a frame
8664 is less than or equal to this value, then it is declared as duplicate. Default
8668 Set scene change threshold. Default is @code{15}.
8672 Set the size of the x and y-axis blocks used during metric calculations.
8673 Larger blocks give better noise suppression, but also give worse detection of
8674 small movements. Must be a power of two. Default is @code{32}.
8677 Mark main input as a pre-processed input and activate clean source input
8678 stream. This allows the input to be pre-processed with various filters to help
8679 the metrics calculation while keeping the frame selection lossless. When set to
8680 @code{1}, the first stream is for the pre-processed input, and the second
8681 stream is the clean source from where the kept frames are chosen. Default is
8685 Set whether or not chroma is considered in the metric calculations. Default is
8691 Apply 2D deconvolution of video stream in frequency domain using second stream
8694 The filter accepts the following options:
8698 Set which planes to process.
8701 Set which impulse video frames will be processed, can be @var{first}
8702 or @var{all}. Default is @var{all}.
8705 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
8706 and height are not same and not power of 2 or if stream prior to convolving
8710 The @code{deconvolve} filter also supports the @ref{framesync} options.
8714 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
8716 It accepts the following options:
8720 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
8721 @var{rainbows} for cross-color reduction.
8724 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
8727 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
8730 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
8733 Set temporal chroma threshold. Lower values increases reduction of cross-color.
8738 Apply deflate effect to the video.
8740 This filter replaces the pixel by the local(3x3) average by taking into account
8741 only values lower than the pixel.
8743 It accepts the following options:
8750 Limit the maximum change for each plane, default is 65535.
8751 If 0, plane will remain unchanged.
8754 @subsection Commands
8756 This filter supports the all above options as @ref{commands}.
8760 Remove temporal frame luminance variations.
8762 It accepts the following options:
8766 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
8769 Set averaging mode to smooth temporal luminance variations.
8771 Available values are:
8796 Do not actually modify frame. Useful when one only wants metadata.
8801 Remove judder produced by partially interlaced telecined content.
8803 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
8804 source was partially telecined content then the output of @code{pullup,dejudder}
8805 will have a variable frame rate. May change the recorded frame rate of the
8806 container. Aside from that change, this filter will not affect constant frame
8809 The option available in this filter is:
8813 Specify the length of the window over which the judder repeats.
8815 Accepts any integer greater than 1. Useful values are:
8819 If the original was telecined from 24 to 30 fps (Film to NTSC).
8822 If the original was telecined from 25 to 30 fps (PAL to NTSC).
8825 If a mixture of the two.
8828 The default is @samp{4}.
8833 Suppress a TV station logo by a simple interpolation of the surrounding
8834 pixels. Just set a rectangle covering the logo and watch it disappear
8835 (and sometimes something even uglier appear - your mileage may vary).
8837 It accepts the following parameters:
8842 Specify the top left corner coordinates of the logo. They must be
8847 Specify the width and height of the logo to clear. They must be
8851 Specify the thickness of the fuzzy edge of the rectangle (added to
8852 @var{w} and @var{h}). The default value is 1. This option is
8853 deprecated, setting higher values should no longer be necessary and
8857 When set to 1, a green rectangle is drawn on the screen to simplify
8858 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
8859 The default value is 0.
8861 The rectangle is drawn on the outermost pixels which will be (partly)
8862 replaced with interpolated values. The values of the next pixels
8863 immediately outside this rectangle in each direction will be used to
8864 compute the interpolated pixel values inside the rectangle.
8868 @subsection Examples
8872 Set a rectangle covering the area with top left corner coordinates 0,0
8873 and size 100x77, and a band of size 10:
8875 delogo=x=0:y=0:w=100:h=77:band=10
8882 Remove the rain in the input image/video by applying the derain methods based on
8883 convolutional neural networks. Supported models:
8887 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
8888 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
8891 Training as well as model generation scripts are provided in
8892 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
8894 Native model files (.model) can be generated from TensorFlow model
8895 files (.pb) by using tools/python/convert.py
8897 The filter accepts the following options:
8901 Specify which filter to use. This option accepts the following values:
8905 Derain filter. To conduct derain filter, you need to use a derain model.
8908 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
8910 Default value is @samp{derain}.
8913 Specify which DNN backend to use for model loading and execution. This option accepts
8914 the following values:
8918 Native implementation of DNN loading and execution.
8921 TensorFlow backend. To enable this backend you
8922 need to install the TensorFlow for C library (see
8923 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
8924 @code{--enable-libtensorflow}
8926 Default value is @samp{native}.
8929 Set path to model file specifying network architecture and its parameters.
8930 Note that different backends use different file formats. TensorFlow and native
8931 backend can load files for only its format.
8936 Attempt to fix small changes in horizontal and/or vertical shift. This
8937 filter helps remove camera shake from hand-holding a camera, bumping a
8938 tripod, moving on a vehicle, etc.
8940 The filter accepts the following options:
8948 Specify a rectangular area where to limit the search for motion
8950 If desired the search for motion vectors can be limited to a
8951 rectangular area of the frame defined by its top left corner, width
8952 and height. These parameters have the same meaning as the drawbox
8953 filter which can be used to visualise the position of the bounding
8956 This is useful when simultaneous movement of subjects within the frame
8957 might be confused for camera motion by the motion vector search.
8959 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
8960 then the full frame is used. This allows later options to be set
8961 without specifying the bounding box for the motion vector search.
8963 Default - search the whole frame.
8967 Specify the maximum extent of movement in x and y directions in the
8968 range 0-64 pixels. Default 16.
8971 Specify how to generate pixels to fill blanks at the edge of the
8972 frame. Available values are:
8975 Fill zeroes at blank locations
8977 Original image at blank locations
8979 Extruded edge value at blank locations
8981 Mirrored edge at blank locations
8983 Default value is @samp{mirror}.
8986 Specify the blocksize to use for motion search. Range 4-128 pixels,
8990 Specify the contrast threshold for blocks. Only blocks with more than
8991 the specified contrast (difference between darkest and lightest
8992 pixels) will be considered. Range 1-255, default 125.
8995 Specify the search strategy. Available values are:
8998 Set exhaustive search
9000 Set less exhaustive search.
9002 Default value is @samp{exhaustive}.
9005 If set then a detailed log of the motion search is written to the
9012 Remove unwanted contamination of foreground colors, caused by reflected color of
9013 greenscreen or bluescreen.
9015 This filter accepts the following options:
9019 Set what type of despill to use.
9022 Set how spillmap will be generated.
9025 Set how much to get rid of still remaining spill.
9028 Controls amount of red in spill area.
9031 Controls amount of green in spill area.
9032 Should be -1 for greenscreen.
9035 Controls amount of blue in spill area.
9036 Should be -1 for bluescreen.
9039 Controls brightness of spill area, preserving colors.
9042 Modify alpha from generated spillmap.
9047 Apply an exact inverse of the telecine operation. It requires a predefined
9048 pattern specified using the pattern option which must be the same as that passed
9049 to the telecine filter.
9051 This filter accepts the following options:
9060 The default value is @code{top}.
9064 A string of numbers representing the pulldown pattern you wish to apply.
9065 The default value is @code{23}.
9068 A number representing position of the first frame with respect to the telecine
9069 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9074 Apply dilation effect to the video.
9076 This filter replaces the pixel by the local(3x3) maximum.
9078 It accepts the following options:
9085 Limit the maximum change for each plane, default is 65535.
9086 If 0, plane will remain unchanged.
9089 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9092 Flags to local 3x3 coordinates maps like this:
9099 @subsection Commands
9101 This filter supports the all above options as @ref{commands}.
9105 Displace pixels as indicated by second and third input stream.
9107 It takes three input streams and outputs one stream, the first input is the
9108 source, and second and third input are displacement maps.
9110 The second input specifies how much to displace pixels along the
9111 x-axis, while the third input specifies how much to displace pixels
9113 If one of displacement map streams terminates, last frame from that
9114 displacement map will be used.
9116 Note that once generated, displacements maps can be reused over and over again.
9118 A description of the accepted options follows.
9122 Set displace behavior for pixels that are out of range.
9124 Available values are:
9127 Missing pixels are replaced by black pixels.
9130 Adjacent pixels will spread out to replace missing pixels.
9133 Out of range pixels are wrapped so they point to pixels of other side.
9136 Out of range pixels will be replaced with mirrored pixels.
9138 Default is @samp{smear}.
9142 @subsection Examples
9146 Add ripple effect to rgb input of video size hd720:
9148 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
9152 Add wave effect to rgb input of video size hd720:
9154 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
9158 @section dnn_processing
9160 Do image processing with deep neural networks. It works together with another filter
9161 which converts the pixel format of the Frame to what the dnn network requires.
9163 The filter accepts the following options:
9167 Specify which DNN backend to use for model loading and execution. This option accepts
9168 the following values:
9172 Native implementation of DNN loading and execution.
9175 TensorFlow backend. To enable this backend you
9176 need to install the TensorFlow for C library (see
9177 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9178 @code{--enable-libtensorflow}
9181 Default value is @samp{native}.
9184 Set path to model file specifying network architecture and its parameters.
9185 Note that different backends use different file formats. TensorFlow and native
9186 backend can load files for only its format.
9188 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9191 Set the input name of the dnn network.
9194 Set the output name of the dnn network.
9200 Halve the red channle of the frame with format rgb24:
9202 ffmpeg -i input.jpg -vf format=rgb24,dnn_processing=model=halve_first_channel.model:input=dnn_in:output=dnn_out:dnn_backend=native out.native.png
9206 Halve the pixel value of the frame with format gray32f:
9208 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
9215 Draw a colored box on the input image.
9217 It accepts the following parameters:
9222 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9226 The expressions which specify the width and height of the box; if 0 they are interpreted as
9227 the input width and height. It defaults to 0.
9230 Specify the color of the box to write. For the general syntax of this option,
9231 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9232 value @code{invert} is used, the box edge color is the same as the
9233 video with inverted luma.
9236 The expression which sets the thickness of the box edge.
9237 A value of @code{fill} will create a filled box. Default value is @code{3}.
9239 See below for the list of accepted constants.
9242 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
9243 will overwrite the video's color and alpha pixels.
9244 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
9247 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9248 following constants:
9252 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9256 horizontal and vertical chroma subsample values. For example for the
9257 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9261 The input width and height.
9264 The input sample aspect ratio.
9268 The x and y offset coordinates where the box is drawn.
9272 The width and height of the drawn box.
9275 The thickness of the drawn box.
9277 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9278 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9282 @subsection Examples
9286 Draw a black box around the edge of the input image:
9292 Draw a box with color red and an opacity of 50%:
9294 drawbox=10:20:200:60:red@@0.5
9297 The previous example can be specified as:
9299 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
9303 Fill the box with pink color:
9305 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
9309 Draw a 2-pixel red 2.40:1 mask:
9311 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
9315 @subsection Commands
9316 This filter supports same commands as options.
9317 The command accepts the same syntax of the corresponding option.
9319 If the specified expression is not valid, it is kept at its current
9324 Draw a graph using input video metadata.
9326 It accepts the following parameters:
9330 Set 1st frame metadata key from which metadata values will be used to draw a graph.
9333 Set 1st foreground color expression.
9336 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
9339 Set 2nd foreground color expression.
9342 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
9345 Set 3rd foreground color expression.
9348 Set 4th frame metadata key from which metadata values will be used to draw a graph.
9351 Set 4th foreground color expression.
9354 Set minimal value of metadata value.
9357 Set maximal value of metadata value.
9360 Set graph background color. Default is white.
9365 Available values for mode is:
9372 Default is @code{line}.
9377 Available values for slide is:
9380 Draw new frame when right border is reached.
9383 Replace old columns with new ones.
9386 Scroll from right to left.
9389 Scroll from left to right.
9392 Draw single picture.
9395 Default is @code{frame}.
9398 Set size of graph video. For the syntax of this option, check the
9399 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
9400 The default value is @code{900x256}.
9403 Set the output frame rate. Default value is @code{25}.
9405 The foreground color expressions can use the following variables:
9408 Minimal value of metadata value.
9411 Maximal value of metadata value.
9414 Current metadata key value.
9417 The color is defined as 0xAABBGGRR.
9420 Example using metadata from @ref{signalstats} filter:
9422 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
9425 Example using metadata from @ref{ebur128} filter:
9427 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
9432 Draw a grid on the input image.
9434 It accepts the following parameters:
9439 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
9443 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
9444 input width and height, respectively, minus @code{thickness}, so image gets
9445 framed. Default to 0.
9448 Specify the color of the grid. For the general syntax of this option,
9449 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9450 value @code{invert} is used, the grid color is the same as the
9451 video with inverted luma.
9454 The expression which sets the thickness of the grid line. Default value is @code{1}.
9456 See below for the list of accepted constants.
9459 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
9460 will overwrite the video's color and alpha pixels.
9461 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
9464 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9465 following constants:
9469 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9473 horizontal and vertical chroma subsample values. For example for the
9474 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9478 The input grid cell width and height.
9481 The input sample aspect ratio.
9485 The x and y coordinates of some point of grid intersection (meant to configure offset).
9489 The width and height of the drawn cell.
9492 The thickness of the drawn cell.
9494 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9495 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9499 @subsection Examples
9503 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
9505 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
9509 Draw a white 3x3 grid with an opacity of 50%:
9511 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
9515 @subsection Commands
9516 This filter supports same commands as options.
9517 The command accepts the same syntax of the corresponding option.
9519 If the specified expression is not valid, it is kept at its current
9525 Draw a text string or text from a specified file on top of a video, using the
9526 libfreetype library.
9528 To enable compilation of this filter, you need to configure FFmpeg with
9529 @code{--enable-libfreetype}.
9530 To enable default font fallback and the @var{font} option you need to
9531 configure FFmpeg with @code{--enable-libfontconfig}.
9532 To enable the @var{text_shaping} option, you need to configure FFmpeg with
9533 @code{--enable-libfribidi}.
9537 It accepts the following parameters:
9542 Used to draw a box around text using the background color.
9543 The value must be either 1 (enable) or 0 (disable).
9544 The default value of @var{box} is 0.
9547 Set the width of the border to be drawn around the box using @var{boxcolor}.
9548 The default value of @var{boxborderw} is 0.
9551 The color to be used for drawing box around text. For the syntax of this
9552 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9554 The default value of @var{boxcolor} is "white".
9557 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
9558 The default value of @var{line_spacing} is 0.
9561 Set the width of the border to be drawn around the text using @var{bordercolor}.
9562 The default value of @var{borderw} is 0.
9565 Set the color to be used for drawing border around text. For the syntax of this
9566 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9568 The default value of @var{bordercolor} is "black".
9571 Select how the @var{text} is expanded. Can be either @code{none},
9572 @code{strftime} (deprecated) or
9573 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
9577 Set a start time for the count. Value is in microseconds. Only applied
9578 in the deprecated strftime expansion mode. To emulate in normal expansion
9579 mode use the @code{pts} function, supplying the start time (in seconds)
9580 as the second argument.
9583 If true, check and fix text coords to avoid clipping.
9586 The color to be used for drawing fonts. For the syntax of this option, check
9587 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9589 The default value of @var{fontcolor} is "black".
9591 @item fontcolor_expr
9592 String which is expanded the same way as @var{text} to obtain dynamic
9593 @var{fontcolor} value. By default this option has empty value and is not
9594 processed. When this option is set, it overrides @var{fontcolor} option.
9597 The font family to be used for drawing text. By default Sans.
9600 The font file to be used for drawing text. The path must be included.
9601 This parameter is mandatory if the fontconfig support is disabled.
9604 Draw the text applying alpha blending. The value can
9605 be a number between 0.0 and 1.0.
9606 The expression accepts the same variables @var{x, y} as well.
9607 The default value is 1.
9608 Please see @var{fontcolor_expr}.
9611 The font size to be used for drawing text.
9612 The default value of @var{fontsize} is 16.
9615 If set to 1, attempt to shape the text (for example, reverse the order of
9616 right-to-left text and join Arabic characters) before drawing it.
9617 Otherwise, just draw the text exactly as given.
9618 By default 1 (if supported).
9621 The flags to be used for loading the fonts.
9623 The flags map the corresponding flags supported by libfreetype, and are
9624 a combination of the following values:
9631 @item vertical_layout
9632 @item force_autohint
9635 @item ignore_global_advance_width
9637 @item ignore_transform
9643 Default value is "default".
9645 For more information consult the documentation for the FT_LOAD_*
9649 The color to be used for drawing a shadow behind the drawn text. For the
9650 syntax of this option, check the @ref{color syntax,,"Color" section in the
9651 ffmpeg-utils manual,ffmpeg-utils}.
9653 The default value of @var{shadowcolor} is "black".
9657 The x and y offsets for the text shadow position with respect to the
9658 position of the text. They can be either positive or negative
9659 values. The default value for both is "0".
9662 The starting frame number for the n/frame_num variable. The default value
9666 The size in number of spaces to use for rendering the tab.
9670 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
9671 format. It can be used with or without text parameter. @var{timecode_rate}
9672 option must be specified.
9674 @item timecode_rate, rate, r
9675 Set the timecode frame rate (timecode only). Value will be rounded to nearest
9676 integer. Minimum value is "1".
9677 Drop-frame timecode is supported for frame rates 30 & 60.
9680 If set to 1, the output of the timecode option will wrap around at 24 hours.
9681 Default is 0 (disabled).
9684 The text string to be drawn. The text must be a sequence of UTF-8
9686 This parameter is mandatory if no file is specified with the parameter
9690 A text file containing text to be drawn. The text must be a sequence
9691 of UTF-8 encoded characters.
9693 This parameter is mandatory if no text string is specified with the
9694 parameter @var{text}.
9696 If both @var{text} and @var{textfile} are specified, an error is thrown.
9699 If set to 1, the @var{textfile} will be reloaded before each frame.
9700 Be sure to update it atomically, or it may be read partially, or even fail.
9704 The expressions which specify the offsets where text will be drawn
9705 within the video frame. They are relative to the top/left border of the
9708 The default value of @var{x} and @var{y} is "0".
9710 See below for the list of accepted constants and functions.
9713 The parameters for @var{x} and @var{y} are expressions containing the
9714 following constants and functions:
9718 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
9722 horizontal and vertical chroma subsample values. For example for the
9723 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9726 the height of each text line
9734 @item max_glyph_a, ascent
9735 the maximum distance from the baseline to the highest/upper grid
9736 coordinate used to place a glyph outline point, for all the rendered
9738 It is a positive value, due to the grid's orientation with the Y axis
9741 @item max_glyph_d, descent
9742 the maximum distance from the baseline to the lowest grid coordinate
9743 used to place a glyph outline point, for all the rendered glyphs.
9744 This is a negative value, due to the grid's orientation, with the Y axis
9748 maximum glyph height, that is the maximum height for all the glyphs
9749 contained in the rendered text, it is equivalent to @var{ascent} -
9753 maximum glyph width, that is the maximum width for all the glyphs
9754 contained in the rendered text
9757 the number of input frame, starting from 0
9759 @item rand(min, max)
9760 return a random number included between @var{min} and @var{max}
9763 The input sample aspect ratio.
9766 timestamp expressed in seconds, NAN if the input timestamp is unknown
9769 the height of the rendered text
9772 the width of the rendered text
9776 the x and y offset coordinates where the text is drawn.
9778 These parameters allow the @var{x} and @var{y} expressions to refer
9779 to each other, so you can for example specify @code{y=x/dar}.
9782 A one character description of the current frame's picture type.
9785 The current packet's position in the input file or stream
9786 (in bytes, from the start of the input). A value of -1 indicates
9787 this info is not available.
9790 The current packet's duration, in seconds.
9793 The current packet's size (in bytes).
9796 @anchor{drawtext_expansion}
9797 @subsection Text expansion
9799 If @option{expansion} is set to @code{strftime},
9800 the filter recognizes strftime() sequences in the provided text and
9801 expands them accordingly. Check the documentation of strftime(). This
9802 feature is deprecated.
9804 If @option{expansion} is set to @code{none}, the text is printed verbatim.
9806 If @option{expansion} is set to @code{normal} (which is the default),
9807 the following expansion mechanism is used.
9809 The backslash character @samp{\}, followed by any character, always expands to
9810 the second character.
9812 Sequences of the form @code{%@{...@}} are expanded. The text between the
9813 braces is a function name, possibly followed by arguments separated by ':'.
9814 If the arguments contain special characters or delimiters (':' or '@}'),
9815 they should be escaped.
9817 Note that they probably must also be escaped as the value for the
9818 @option{text} option in the filter argument string and as the filter
9819 argument in the filtergraph description, and possibly also for the shell,
9820 that makes up to four levels of escaping; using a text file avoids these
9823 The following functions are available:
9828 The expression evaluation result.
9830 It must take one argument specifying the expression to be evaluated,
9831 which accepts the same constants and functions as the @var{x} and
9832 @var{y} values. Note that not all constants should be used, for
9833 example the text size is not known when evaluating the expression, so
9834 the constants @var{text_w} and @var{text_h} will have an undefined
9837 @item expr_int_format, eif
9838 Evaluate the expression's value and output as formatted integer.
9840 The first argument is the expression to be evaluated, just as for the @var{expr} function.
9841 The second argument specifies the output format. Allowed values are @samp{x},
9842 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
9843 @code{printf} function.
9844 The third parameter is optional and sets the number of positions taken by the output.
9845 It can be used to add padding with zeros from the left.
9848 The time at which the filter is running, expressed in UTC.
9849 It can accept an argument: a strftime() format string.
9852 The time at which the filter is running, expressed in the local time zone.
9853 It can accept an argument: a strftime() format string.
9856 Frame metadata. Takes one or two arguments.
9858 The first argument is mandatory and specifies the metadata key.
9860 The second argument is optional and specifies a default value, used when the
9861 metadata key is not found or empty.
9863 Available metadata can be identified by inspecting entries
9864 starting with TAG included within each frame section
9865 printed by running @code{ffprobe -show_frames}.
9867 String metadata generated in filters leading to
9868 the drawtext filter are also available.
9871 The frame number, starting from 0.
9874 A one character description of the current picture type.
9877 The timestamp of the current frame.
9878 It can take up to three arguments.
9880 The first argument is the format of the timestamp; it defaults to @code{flt}
9881 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
9882 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
9883 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
9884 @code{localtime} stands for the timestamp of the frame formatted as
9885 local time zone time.
9887 The second argument is an offset added to the timestamp.
9889 If the format is set to @code{hms}, a third argument @code{24HH} may be
9890 supplied to present the hour part of the formatted timestamp in 24h format
9893 If the format is set to @code{localtime} or @code{gmtime},
9894 a third argument may be supplied: a strftime() format string.
9895 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
9898 @subsection Commands
9900 This filter supports altering parameters via commands:
9903 Alter existing filter parameters.
9905 Syntax for the argument is the same as for filter invocation, e.g.
9908 fontsize=56:fontcolor=green:text='Hello World'
9911 Full filter invocation with sendcmd would look like this:
9914 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
9918 If the entire argument can't be parsed or applied as valid values then the filter will
9919 continue with its existing parameters.
9921 @subsection Examples
9925 Draw "Test Text" with font FreeSerif, using the default values for the
9926 optional parameters.
9929 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
9933 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
9934 and y=50 (counting from the top-left corner of the screen), text is
9935 yellow with a red box around it. Both the text and the box have an
9939 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
9940 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
9943 Note that the double quotes are not necessary if spaces are not used
9944 within the parameter list.
9947 Show the text at the center of the video frame:
9949 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
9953 Show the text at a random position, switching to a new position every 30 seconds:
9955 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)"
9959 Show a text line sliding from right to left in the last row of the video
9960 frame. The file @file{LONG_LINE} is assumed to contain a single line
9963 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
9967 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
9969 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
9973 Draw a single green letter "g", at the center of the input video.
9974 The glyph baseline is placed at half screen height.
9976 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
9980 Show text for 1 second every 3 seconds:
9982 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
9986 Use fontconfig to set the font. Note that the colons need to be escaped.
9988 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
9992 Print the date of a real-time encoding (see strftime(3)):
9994 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
9998 Show text fading in and out (appearing/disappearing):
10001 DS=1.0 # display start
10002 DE=10.0 # display end
10003 FID=1.5 # fade in duration
10004 FOD=5 # fade out duration
10005 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 @}"
10009 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10010 and the @option{fontsize} value are included in the @option{y} offset.
10012 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10013 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10017 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10018 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10019 must have option @option{-export_path_metadata 1} for the special metadata fields
10020 to be available for filters.
10022 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10027 For more information about libfreetype, check:
10028 @url{http://www.freetype.org/}.
10030 For more information about fontconfig, check:
10031 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10033 For more information about libfribidi, check:
10034 @url{http://fribidi.org/}.
10036 @section edgedetect
10038 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10040 The filter accepts the following options:
10045 Set low and high threshold values used by the Canny thresholding
10048 The high threshold selects the "strong" edge pixels, which are then
10049 connected through 8-connectivity with the "weak" edge pixels selected
10050 by the low threshold.
10052 @var{low} and @var{high} threshold values must be chosen in the range
10053 [0,1], and @var{low} should be lesser or equal to @var{high}.
10055 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10059 Define the drawing mode.
10063 Draw white/gray wires on black background.
10066 Mix the colors to create a paint/cartoon effect.
10069 Apply Canny edge detector on all selected planes.
10071 Default value is @var{wires}.
10074 Select planes for filtering. By default all available planes are filtered.
10077 @subsection Examples
10081 Standard edge detection with custom values for the hysteresis thresholding:
10083 edgedetect=low=0.1:high=0.4
10087 Painting effect without thresholding:
10089 edgedetect=mode=colormix:high=0
10095 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10097 For each input image, the filter will compute the optimal mapping from
10098 the input to the output given the codebook length, that is the number
10099 of distinct output colors.
10101 This filter accepts the following options.
10104 @item codebook_length, l
10105 Set codebook length. The value must be a positive integer, and
10106 represents the number of distinct output colors. Default value is 256.
10109 Set the maximum number of iterations to apply for computing the optimal
10110 mapping. The higher the value the better the result and the higher the
10111 computation time. Default value is 1.
10114 Set a random seed, must be an integer included between 0 and
10115 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10116 will try to use a good random seed on a best effort basis.
10119 Set pal8 output pixel format. This option does not work with codebook
10120 length greater than 256.
10125 Measure graylevel entropy in histogram of color channels of video frames.
10127 It accepts the following parameters:
10131 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10133 @var{diff} mode measures entropy of histogram delta values, absolute differences
10134 between neighbour histogram values.
10138 Set brightness, contrast, saturation and approximate gamma adjustment.
10140 The filter accepts the following options:
10144 Set the contrast expression. The value must be a float value in range
10145 @code{-1000.0} to @code{1000.0}. The default value is "1".
10148 Set the brightness expression. The value must be a float value in
10149 range @code{-1.0} to @code{1.0}. The default value is "0".
10152 Set the saturation expression. The value must be a float in
10153 range @code{0.0} to @code{3.0}. The default value is "1".
10156 Set the gamma expression. The value must be a float in range
10157 @code{0.1} to @code{10.0}. The default value is "1".
10160 Set the gamma expression for red. The value must be a float in
10161 range @code{0.1} to @code{10.0}. The default value is "1".
10164 Set the gamma expression for green. The value must be a float in range
10165 @code{0.1} to @code{10.0}. The default value is "1".
10168 Set the gamma expression for blue. The value must be a float in range
10169 @code{0.1} to @code{10.0}. The default value is "1".
10172 Set the gamma weight expression. It can be used to reduce the effect
10173 of a high gamma value on bright image areas, e.g. keep them from
10174 getting overamplified and just plain white. The value must be a float
10175 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10176 gamma correction all the way down while @code{1.0} leaves it at its
10177 full strength. Default is "1".
10180 Set when the expressions for brightness, contrast, saturation and
10181 gamma expressions are evaluated.
10183 It accepts the following values:
10186 only evaluate expressions once during the filter initialization or
10187 when a command is processed
10190 evaluate expressions for each incoming frame
10193 Default value is @samp{init}.
10196 The expressions accept the following parameters:
10199 frame count of the input frame starting from 0
10202 byte position of the corresponding packet in the input file, NAN if
10206 frame rate of the input video, NAN if the input frame rate is unknown
10209 timestamp expressed in seconds, NAN if the input timestamp is unknown
10212 @subsection Commands
10213 The filter supports the following commands:
10217 Set the contrast expression.
10220 Set the brightness expression.
10223 Set the saturation expression.
10226 Set the gamma expression.
10229 Set the gamma_r expression.
10232 Set gamma_g expression.
10235 Set gamma_b expression.
10238 Set gamma_weight expression.
10240 The command accepts the same syntax of the corresponding option.
10242 If the specified expression is not valid, it is kept at its current
10249 Apply erosion effect to the video.
10251 This filter replaces the pixel by the local(3x3) minimum.
10253 It accepts the following options:
10260 Limit the maximum change for each plane, default is 65535.
10261 If 0, plane will remain unchanged.
10264 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
10267 Flags to local 3x3 coordinates maps like this:
10274 @subsection Commands
10276 This filter supports the all above options as @ref{commands}.
10278 @section extractplanes
10280 Extract color channel components from input video stream into
10281 separate grayscale video streams.
10283 The filter accepts the following option:
10287 Set plane(s) to extract.
10289 Available values for planes are:
10300 Choosing planes not available in the input will result in an error.
10301 That means you cannot select @code{r}, @code{g}, @code{b} planes
10302 with @code{y}, @code{u}, @code{v} planes at same time.
10305 @subsection Examples
10309 Extract luma, u and v color channel component from input video frame
10310 into 3 grayscale outputs:
10312 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
10318 Apply a fade-in/out effect to the input video.
10320 It accepts the following parameters:
10324 The effect type can be either "in" for a fade-in, or "out" for a fade-out
10326 Default is @code{in}.
10328 @item start_frame, s
10329 Specify the number of the frame to start applying the fade
10330 effect at. Default is 0.
10333 The number of frames that the fade effect lasts. At the end of the
10334 fade-in effect, the output video will have the same intensity as the input video.
10335 At the end of the fade-out transition, the output video will be filled with the
10336 selected @option{color}.
10340 If set to 1, fade only alpha channel, if one exists on the input.
10341 Default value is 0.
10343 @item start_time, st
10344 Specify the timestamp (in seconds) of the frame to start to apply the fade
10345 effect. If both start_frame and start_time are specified, the fade will start at
10346 whichever comes last. Default is 0.
10349 The number of seconds for which the fade effect has to last. At the end of the
10350 fade-in effect the output video will have the same intensity as the input video,
10351 at the end of the fade-out transition the output video will be filled with the
10352 selected @option{color}.
10353 If both duration and nb_frames are specified, duration is used. Default is 0
10354 (nb_frames is used by default).
10357 Specify the color of the fade. Default is "black".
10360 @subsection Examples
10364 Fade in the first 30 frames of video:
10369 The command above is equivalent to:
10375 Fade out the last 45 frames of a 200-frame video:
10378 fade=type=out:start_frame=155:nb_frames=45
10382 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
10384 fade=in:0:25, fade=out:975:25
10388 Make the first 5 frames yellow, then fade in from frame 5-24:
10390 fade=in:5:20:color=yellow
10394 Fade in alpha over first 25 frames of video:
10396 fade=in:0:25:alpha=1
10400 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
10402 fade=t=in:st=5.5:d=0.5
10408 Denoise frames using 3D FFT (frequency domain filtering).
10410 The filter accepts the following options:
10414 Set the noise sigma constant. This sets denoising strength.
10415 Default value is 1. Allowed range is from 0 to 30.
10416 Using very high sigma with low overlap may give blocking artifacts.
10419 Set amount of denoising. By default all detected noise is reduced.
10420 Default value is 1. Allowed range is from 0 to 1.
10423 Set size of block, Default is 4, can be 3, 4, 5 or 6.
10424 Actual size of block in pixels is 2 to power of @var{block}, so by default
10425 block size in pixels is 2^4 which is 16.
10428 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
10431 Set number of previous frames to use for denoising. By default is set to 0.
10434 Set number of next frames to to use for denoising. By default is set to 0.
10437 Set planes which will be filtered, by default are all available filtered
10442 Apply arbitrary expressions to samples in frequency domain
10446 Adjust the dc value (gain) of the luma plane of the image. The filter
10447 accepts an integer value in range @code{0} to @code{1000}. The default
10448 value is set to @code{0}.
10451 Adjust the dc value (gain) of the 1st chroma plane of the image. The
10452 filter accepts an integer value in range @code{0} to @code{1000}. The
10453 default value is set to @code{0}.
10456 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
10457 filter accepts an integer value in range @code{0} to @code{1000}. The
10458 default value is set to @code{0}.
10461 Set the frequency domain weight expression for the luma plane.
10464 Set the frequency domain weight expression for the 1st chroma plane.
10467 Set the frequency domain weight expression for the 2nd chroma plane.
10470 Set when the expressions are evaluated.
10472 It accepts the following values:
10475 Only evaluate expressions once during the filter initialization.
10478 Evaluate expressions for each incoming frame.
10481 Default value is @samp{init}.
10483 The filter accepts the following variables:
10486 The coordinates of the current sample.
10490 The width and height of the image.
10493 The number of input frame, starting from 0.
10496 @subsection Examples
10502 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
10508 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
10514 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
10520 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
10527 Extract a single field from an interlaced image using stride
10528 arithmetic to avoid wasting CPU time. The output frames are marked as
10531 The filter accepts the following options:
10535 Specify whether to extract the top (if the value is @code{0} or
10536 @code{top}) or the bottom field (if the value is @code{1} or
10542 Create new frames by copying the top and bottom fields from surrounding frames
10543 supplied as numbers by the hint file.
10547 Set file containing hints: absolute/relative frame numbers.
10549 There must be one line for each frame in a clip. Each line must contain two
10550 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
10551 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
10552 is current frame number for @code{absolute} mode or out of [-1, 1] range
10553 for @code{relative} mode. First number tells from which frame to pick up top
10554 field and second number tells from which frame to pick up bottom field.
10556 If optionally followed by @code{+} output frame will be marked as interlaced,
10557 else if followed by @code{-} output frame will be marked as progressive, else
10558 it will be marked same as input frame.
10559 If optionally followed by @code{t} output frame will use only top field, or in
10560 case of @code{b} it will use only bottom field.
10561 If line starts with @code{#} or @code{;} that line is skipped.
10564 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
10567 Example of first several lines of @code{hint} file for @code{relative} mode:
10569 0,0 - # first frame
10570 1,0 - # second frame, use third's frame top field and second's frame bottom field
10571 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
10586 @section fieldmatch
10588 Field matching filter for inverse telecine. It is meant to reconstruct the
10589 progressive frames from a telecined stream. The filter does not drop duplicated
10590 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
10591 followed by a decimation filter such as @ref{decimate} in the filtergraph.
10593 The separation of the field matching and the decimation is notably motivated by
10594 the possibility of inserting a de-interlacing filter fallback between the two.
10595 If the source has mixed telecined and real interlaced content,
10596 @code{fieldmatch} will not be able to match fields for the interlaced parts.
10597 But these remaining combed frames will be marked as interlaced, and thus can be
10598 de-interlaced by a later filter such as @ref{yadif} before decimation.
10600 In addition to the various configuration options, @code{fieldmatch} can take an
10601 optional second stream, activated through the @option{ppsrc} option. If
10602 enabled, the frames reconstruction will be based on the fields and frames from
10603 this second stream. This allows the first input to be pre-processed in order to
10604 help the various algorithms of the filter, while keeping the output lossless
10605 (assuming the fields are matched properly). Typically, a field-aware denoiser,
10606 or brightness/contrast adjustments can help.
10608 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
10609 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
10610 which @code{fieldmatch} is based on. While the semantic and usage are very
10611 close, some behaviour and options names can differ.
10613 The @ref{decimate} filter currently only works for constant frame rate input.
10614 If your input has mixed telecined (30fps) and progressive content with a lower
10615 framerate like 24fps use the following filterchain to produce the necessary cfr
10616 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
10618 The filter accepts the following options:
10622 Specify the assumed field order of the input stream. Available values are:
10626 Auto detect parity (use FFmpeg's internal parity value).
10628 Assume bottom field first.
10630 Assume top field first.
10633 Note that it is sometimes recommended not to trust the parity announced by the
10636 Default value is @var{auto}.
10639 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
10640 sense that it won't risk creating jerkiness due to duplicate frames when
10641 possible, but if there are bad edits or blended fields it will end up
10642 outputting combed frames when a good match might actually exist. On the other
10643 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
10644 but will almost always find a good frame if there is one. The other values are
10645 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
10646 jerkiness and creating duplicate frames versus finding good matches in sections
10647 with bad edits, orphaned fields, blended fields, etc.
10649 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
10651 Available values are:
10655 2-way matching (p/c)
10657 2-way matching, and trying 3rd match if still combed (p/c + n)
10659 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
10661 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
10662 still combed (p/c + n + u/b)
10664 3-way matching (p/c/n)
10666 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
10667 detected as combed (p/c/n + u/b)
10670 The parenthesis at the end indicate the matches that would be used for that
10671 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
10674 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
10677 Default value is @var{pc_n}.
10680 Mark the main input stream as a pre-processed input, and enable the secondary
10681 input stream as the clean source to pick the fields from. See the filter
10682 introduction for more details. It is similar to the @option{clip2} feature from
10685 Default value is @code{0} (disabled).
10688 Set the field to match from. It is recommended to set this to the same value as
10689 @option{order} unless you experience matching failures with that setting. In
10690 certain circumstances changing the field that is used to match from can have a
10691 large impact on matching performance. Available values are:
10695 Automatic (same value as @option{order}).
10697 Match from the bottom field.
10699 Match from the top field.
10702 Default value is @var{auto}.
10705 Set whether or not chroma is included during the match comparisons. In most
10706 cases it is recommended to leave this enabled. You should set this to @code{0}
10707 only if your clip has bad chroma problems such as heavy rainbowing or other
10708 artifacts. Setting this to @code{0} could also be used to speed things up at
10709 the cost of some accuracy.
10711 Default value is @code{1}.
10715 These define an exclusion band which excludes the lines between @option{y0} and
10716 @option{y1} from being included in the field matching decision. An exclusion
10717 band can be used to ignore subtitles, a logo, or other things that may
10718 interfere with the matching. @option{y0} sets the starting scan line and
10719 @option{y1} sets the ending line; all lines in between @option{y0} and
10720 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
10721 @option{y0} and @option{y1} to the same value will disable the feature.
10722 @option{y0} and @option{y1} defaults to @code{0}.
10725 Set the scene change detection threshold as a percentage of maximum change on
10726 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
10727 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
10728 @option{scthresh} is @code{[0.0, 100.0]}.
10730 Default value is @code{12.0}.
10733 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
10734 account the combed scores of matches when deciding what match to use as the
10735 final match. Available values are:
10739 No final matching based on combed scores.
10741 Combed scores are only used when a scene change is detected.
10743 Use combed scores all the time.
10746 Default is @var{sc}.
10749 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
10750 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
10751 Available values are:
10755 No forced calculation.
10757 Force p/c/n calculations.
10759 Force p/c/n/u/b calculations.
10762 Default value is @var{none}.
10765 This is the area combing threshold used for combed frame detection. This
10766 essentially controls how "strong" or "visible" combing must be to be detected.
10767 Larger values mean combing must be more visible and smaller values mean combing
10768 can be less visible or strong and still be detected. Valid settings are from
10769 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
10770 be detected as combed). This is basically a pixel difference value. A good
10771 range is @code{[8, 12]}.
10773 Default value is @code{9}.
10776 Sets whether or not chroma is considered in the combed frame decision. Only
10777 disable this if your source has chroma problems (rainbowing, etc.) that are
10778 causing problems for the combed frame detection with chroma enabled. Actually,
10779 using @option{chroma}=@var{0} is usually more reliable, except for the case
10780 where there is chroma only combing in the source.
10782 Default value is @code{0}.
10786 Respectively set the x-axis and y-axis size of the window used during combed
10787 frame detection. This has to do with the size of the area in which
10788 @option{combpel} pixels are required to be detected as combed for a frame to be
10789 declared combed. See the @option{combpel} parameter description for more info.
10790 Possible values are any number that is a power of 2 starting at 4 and going up
10793 Default value is @code{16}.
10796 The number of combed pixels inside any of the @option{blocky} by
10797 @option{blockx} size blocks on the frame for the frame to be detected as
10798 combed. While @option{cthresh} controls how "visible" the combing must be, this
10799 setting controls "how much" combing there must be in any localized area (a
10800 window defined by the @option{blockx} and @option{blocky} settings) on the
10801 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
10802 which point no frames will ever be detected as combed). This setting is known
10803 as @option{MI} in TFM/VFM vocabulary.
10805 Default value is @code{80}.
10808 @anchor{p/c/n/u/b meaning}
10809 @subsection p/c/n/u/b meaning
10811 @subsubsection p/c/n
10813 We assume the following telecined stream:
10816 Top fields: 1 2 2 3 4
10817 Bottom fields: 1 2 3 4 4
10820 The numbers correspond to the progressive frame the fields relate to. Here, the
10821 first two frames are progressive, the 3rd and 4th are combed, and so on.
10823 When @code{fieldmatch} is configured to run a matching from bottom
10824 (@option{field}=@var{bottom}) this is how this input stream get transformed:
10829 B 1 2 3 4 4 <-- matching reference
10838 As a result of the field matching, we can see that some frames get duplicated.
10839 To perform a complete inverse telecine, you need to rely on a decimation filter
10840 after this operation. See for instance the @ref{decimate} filter.
10842 The same operation now matching from top fields (@option{field}=@var{top})
10847 T 1 2 2 3 4 <-- matching reference
10857 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
10858 basically, they refer to the frame and field of the opposite parity:
10861 @item @var{p} matches the field of the opposite parity in the previous frame
10862 @item @var{c} matches the field of the opposite parity in the current frame
10863 @item @var{n} matches the field of the opposite parity in the next frame
10868 The @var{u} and @var{b} matching are a bit special in the sense that they match
10869 from the opposite parity flag. In the following examples, we assume that we are
10870 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
10871 'x' is placed above and below each matched fields.
10873 With bottom matching (@option{field}=@var{bottom}):
10878 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10879 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10887 With top matching (@option{field}=@var{top}):
10892 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10893 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10901 @subsection Examples
10903 Simple IVTC of a top field first telecined stream:
10905 fieldmatch=order=tff:combmatch=none, decimate
10908 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
10910 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
10913 @section fieldorder
10915 Transform the field order of the input video.
10917 It accepts the following parameters:
10922 The output field order. Valid values are @var{tff} for top field first or @var{bff}
10923 for bottom field first.
10926 The default value is @samp{tff}.
10928 The transformation is done by shifting the picture content up or down
10929 by one line, and filling the remaining line with appropriate picture content.
10930 This method is consistent with most broadcast field order converters.
10932 If the input video is not flagged as being interlaced, or it is already
10933 flagged as being of the required output field order, then this filter does
10934 not alter the incoming video.
10936 It is very useful when converting to or from PAL DV material,
10937 which is bottom field first.
10941 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
10944 @section fifo, afifo
10946 Buffer input images and send them when they are requested.
10948 It is mainly useful when auto-inserted by the libavfilter
10951 It does not take parameters.
10953 @section fillborders
10955 Fill borders of the input video, without changing video stream dimensions.
10956 Sometimes video can have garbage at the four edges and you may not want to
10957 crop video input to keep size multiple of some number.
10959 This filter accepts the following options:
10963 Number of pixels to fill from left border.
10966 Number of pixels to fill from right border.
10969 Number of pixels to fill from top border.
10972 Number of pixels to fill from bottom border.
10977 It accepts the following values:
10980 fill pixels using outermost pixels
10983 fill pixels using mirroring
10986 fill pixels with constant value
10989 Default is @var{smear}.
10992 Set color for pixels in fixed mode. Default is @var{black}.
10995 @subsection Commands
10996 This filter supports same @ref{commands} as options.
10997 The command accepts the same syntax of the corresponding option.
10999 If the specified expression is not valid, it is kept at its current
11004 Find a rectangular object
11006 It accepts the following options:
11010 Filepath of the object image, needs to be in gray8.
11013 Detection threshold, default is 0.5.
11016 Number of mipmaps, default is 3.
11018 @item xmin, ymin, xmax, ymax
11019 Specifies the rectangle in which to search.
11022 @subsection Examples
11026 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11028 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11034 Flood area with values of same pixel components with another values.
11036 It accepts the following options:
11039 Set pixel x coordinate.
11042 Set pixel y coordinate.
11045 Set source #0 component value.
11048 Set source #1 component value.
11051 Set source #2 component value.
11054 Set source #3 component value.
11057 Set destination #0 component value.
11060 Set destination #1 component value.
11063 Set destination #2 component value.
11066 Set destination #3 component value.
11072 Convert the input video to one of the specified pixel formats.
11073 Libavfilter will try to pick one that is suitable as input to
11076 It accepts the following parameters:
11080 A '|'-separated list of pixel format names, such as
11081 "pix_fmts=yuv420p|monow|rgb24".
11085 @subsection Examples
11089 Convert the input video to the @var{yuv420p} format
11091 format=pix_fmts=yuv420p
11094 Convert the input video to any of the formats in the list
11096 format=pix_fmts=yuv420p|yuv444p|yuv410p
11103 Convert the video to specified constant frame rate by duplicating or dropping
11104 frames as necessary.
11106 It accepts the following parameters:
11110 The desired output frame rate. The default is @code{25}.
11113 Assume the first PTS should be the given value, in seconds. This allows for
11114 padding/trimming at the start of stream. By default, no assumption is made
11115 about the first frame's expected PTS, so no padding or trimming is done.
11116 For example, this could be set to 0 to pad the beginning with duplicates of
11117 the first frame if a video stream starts after the audio stream or to trim any
11118 frames with a negative PTS.
11121 Timestamp (PTS) rounding method.
11123 Possible values are:
11130 round towards -infinity
11132 round towards +infinity
11136 The default is @code{near}.
11139 Action performed when reading the last frame.
11141 Possible values are:
11144 Use same timestamp rounding method as used for other frames.
11146 Pass through last frame if input duration has not been reached yet.
11148 The default is @code{round}.
11152 Alternatively, the options can be specified as a flat string:
11153 @var{fps}[:@var{start_time}[:@var{round}]].
11155 See also the @ref{setpts} filter.
11157 @subsection Examples
11161 A typical usage in order to set the fps to 25:
11167 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11169 fps=fps=film:round=near
11175 Pack two different video streams into a stereoscopic video, setting proper
11176 metadata on supported codecs. The two views should have the same size and
11177 framerate and processing will stop when the shorter video ends. Please note
11178 that you may conveniently adjust view properties with the @ref{scale} and
11181 It accepts the following parameters:
11185 The desired packing format. Supported values are:
11190 The views are next to each other (default).
11193 The views are on top of each other.
11196 The views are packed by line.
11199 The views are packed by column.
11202 The views are temporally interleaved.
11211 # Convert left and right views into a frame-sequential video
11212 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11214 # Convert views into a side-by-side video with the same output resolution as the input
11215 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
11220 Change the frame rate by interpolating new video output frames from the source
11223 This filter is not designed to function correctly with interlaced media. If
11224 you wish to change the frame rate of interlaced media then you are required
11225 to deinterlace before this filter and re-interlace after this filter.
11227 A description of the accepted options follows.
11231 Specify the output frames per second. This option can also be specified
11232 as a value alone. The default is @code{50}.
11235 Specify the start of a range where the output frame will be created as a
11236 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11237 the default is @code{15}.
11240 Specify the end of a range where the output frame will be created as a
11241 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11242 the default is @code{240}.
11245 Specify the level at which a scene change is detected as a value between
11246 0 and 100 to indicate a new scene; a low value reflects a low
11247 probability for the current frame to introduce a new scene, while a higher
11248 value means the current frame is more likely to be one.
11249 The default is @code{8.2}.
11252 Specify flags influencing the filter process.
11254 Available value for @var{flags} is:
11257 @item scene_change_detect, scd
11258 Enable scene change detection using the value of the option @var{scene}.
11259 This flag is enabled by default.
11265 Select one frame every N-th frame.
11267 This filter accepts the following option:
11270 Select frame after every @code{step} frames.
11271 Allowed values are positive integers higher than 0. Default value is @code{1}.
11274 @section freezedetect
11276 Detect frozen video.
11278 This filter logs a message and sets frame metadata when it detects that the
11279 input video has no significant change in content during a specified duration.
11280 Video freeze detection calculates the mean average absolute difference of all
11281 the components of video frames and compares it to a noise floor.
11283 The printed times and duration are expressed in seconds. The
11284 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
11285 whose timestamp equals or exceeds the detection duration and it contains the
11286 timestamp of the first frame of the freeze. The
11287 @code{lavfi.freezedetect.freeze_duration} and
11288 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
11291 The filter accepts the following options:
11295 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
11296 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
11300 Set freeze duration until notification (default is 2 seconds).
11303 @section freezeframes
11305 Freeze video frames.
11307 This filter freezes video frames using frame from 2nd input.
11309 The filter accepts the following options:
11313 Set number of first frame from which to start freeze.
11316 Set number of last frame from which to end freeze.
11319 Set number of frame from 2nd input which will be used instead of replaced frames.
11325 Apply a frei0r effect to the input video.
11327 To enable the compilation of this filter, you need to install the frei0r
11328 header and configure FFmpeg with @code{--enable-frei0r}.
11330 It accepts the following parameters:
11335 The name of the frei0r effect to load. If the environment variable
11336 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
11337 directories specified by the colon-separated list in @env{FREI0R_PATH}.
11338 Otherwise, the standard frei0r paths are searched, in this order:
11339 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
11340 @file{/usr/lib/frei0r-1/}.
11342 @item filter_params
11343 A '|'-separated list of parameters to pass to the frei0r effect.
11347 A frei0r effect parameter can be a boolean (its value is either
11348 "y" or "n"), a double, a color (specified as
11349 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
11350 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
11351 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
11352 a position (specified as @var{X}/@var{Y}, where
11353 @var{X} and @var{Y} are floating point numbers) and/or a string.
11355 The number and types of parameters depend on the loaded effect. If an
11356 effect parameter is not specified, the default value is set.
11358 @subsection Examples
11362 Apply the distort0r effect, setting the first two double parameters:
11364 frei0r=filter_name=distort0r:filter_params=0.5|0.01
11368 Apply the colordistance effect, taking a color as the first parameter:
11370 frei0r=colordistance:0.2/0.3/0.4
11371 frei0r=colordistance:violet
11372 frei0r=colordistance:0x112233
11376 Apply the perspective effect, specifying the top left and top right image
11379 frei0r=perspective:0.2/0.2|0.8/0.2
11383 For more information, see
11384 @url{http://frei0r.dyne.org}
11388 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
11390 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
11391 processing filter, one of them is performed once per block, not per pixel.
11392 This allows for much higher speed.
11394 The filter accepts the following options:
11398 Set quality. This option defines the number of levels for averaging. It accepts
11399 an integer in the range 4-5. Default value is @code{4}.
11402 Force a constant quantization parameter. It accepts an integer in range 0-63.
11403 If not set, the filter will use the QP from the video stream (if available).
11406 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
11407 more details but also more artifacts, while higher values make the image smoother
11408 but also blurrier. Default value is @code{0} − PSNR optimal.
11410 @item use_bframe_qp
11411 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
11412 option may cause flicker since the B-Frames have often larger QP. Default is
11413 @code{0} (not enabled).
11419 Apply Gaussian blur filter.
11421 The filter accepts the following options:
11425 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
11428 Set number of steps for Gaussian approximation. Default is @code{1}.
11431 Set which planes to filter. By default all planes are filtered.
11434 Set vertical sigma, if negative it will be same as @code{sigma}.
11435 Default is @code{-1}.
11438 @subsection Commands
11439 This filter supports same commands as options.
11440 The command accepts the same syntax of the corresponding option.
11442 If the specified expression is not valid, it is kept at its current
11447 Apply generic equation to each pixel.
11449 The filter accepts the following options:
11452 @item lum_expr, lum
11453 Set the luminance expression.
11455 Set the chrominance blue expression.
11457 Set the chrominance red expression.
11458 @item alpha_expr, a
11459 Set the alpha expression.
11461 Set the red expression.
11462 @item green_expr, g
11463 Set the green expression.
11465 Set the blue expression.
11468 The colorspace is selected according to the specified options. If one
11469 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
11470 options is specified, the filter will automatically select a YCbCr
11471 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
11472 @option{blue_expr} options is specified, it will select an RGB
11475 If one of the chrominance expression is not defined, it falls back on the other
11476 one. If no alpha expression is specified it will evaluate to opaque value.
11477 If none of chrominance expressions are specified, they will evaluate
11478 to the luminance expression.
11480 The expressions can use the following variables and functions:
11484 The sequential number of the filtered frame, starting from @code{0}.
11488 The coordinates of the current sample.
11492 The width and height of the image.
11496 Width and height scale depending on the currently filtered plane. It is the
11497 ratio between the corresponding luma plane number of pixels and the current
11498 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
11499 @code{0.5,0.5} for chroma planes.
11502 Time of the current frame, expressed in seconds.
11505 Return the value of the pixel at location (@var{x},@var{y}) of the current
11509 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
11513 Return the value of the pixel at location (@var{x},@var{y}) of the
11514 blue-difference chroma plane. Return 0 if there is no such plane.
11517 Return the value of the pixel at location (@var{x},@var{y}) of the
11518 red-difference chroma plane. Return 0 if there is no such plane.
11523 Return the value of the pixel at location (@var{x},@var{y}) of the
11524 red/green/blue component. Return 0 if there is no such component.
11527 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
11528 plane. Return 0 if there is no such plane.
11530 @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)
11531 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
11532 sums of samples within a rectangle. See the functions without the sum postfix.
11534 @item interpolation
11535 Set one of interpolation methods:
11540 Default is bilinear.
11543 For functions, if @var{x} and @var{y} are outside the area, the value will be
11544 automatically clipped to the closer edge.
11546 Please note that this filter can use multiple threads in which case each slice
11547 will have its own expression state. If you want to use only a single expression
11548 state because your expressions depend on previous state then you should limit
11549 the number of filter threads to 1.
11551 @subsection Examples
11555 Flip the image horizontally:
11561 Generate a bidimensional sine wave, with angle @code{PI/3} and a
11562 wavelength of 100 pixels:
11564 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
11568 Generate a fancy enigmatic moving light:
11570 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
11574 Generate a quick emboss effect:
11576 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
11580 Modify RGB components depending on pixel position:
11582 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
11586 Create a radial gradient that is the same size as the input (also see
11587 the @ref{vignette} filter):
11589 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
11595 Fix the banding artifacts that are sometimes introduced into nearly flat
11596 regions by truncation to 8-bit color depth.
11597 Interpolate the gradients that should go where the bands are, and
11600 It is designed for playback only. Do not use it prior to
11601 lossy compression, because compression tends to lose the dither and
11602 bring back the bands.
11604 It accepts the following parameters:
11609 The maximum amount by which the filter will change any one pixel. This is also
11610 the threshold for detecting nearly flat regions. Acceptable values range from
11611 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
11615 The neighborhood to fit the gradient to. A larger radius makes for smoother
11616 gradients, but also prevents the filter from modifying the pixels near detailed
11617 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
11618 values will be clipped to the valid range.
11622 Alternatively, the options can be specified as a flat string:
11623 @var{strength}[:@var{radius}]
11625 @subsection Examples
11629 Apply the filter with a @code{3.5} strength and radius of @code{8}:
11635 Specify radius, omitting the strength (which will fall-back to the default
11643 @anchor{graphmonitor}
11644 @section graphmonitor
11645 Show various filtergraph stats.
11647 With this filter one can debug complete filtergraph.
11648 Especially issues with links filling with queued frames.
11650 The filter accepts the following options:
11654 Set video output size. Default is @var{hd720}.
11657 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
11660 Set output mode, can be @var{fulll} or @var{compact}.
11661 In @var{compact} mode only filters with some queued frames have displayed stats.
11664 Set flags which enable which stats are shown in video.
11666 Available values for flags are:
11669 Display number of queued frames in each link.
11671 @item frame_count_in
11672 Display number of frames taken from filter.
11674 @item frame_count_out
11675 Display number of frames given out from filter.
11678 Display current filtered frame pts.
11681 Display current filtered frame time.
11684 Display time base for filter link.
11687 Display used format for filter link.
11690 Display video size or number of audio channels in case of audio used by filter link.
11693 Display video frame rate or sample rate in case of audio used by filter link.
11697 Set upper limit for video rate of output stream, Default value is @var{25}.
11698 This guarantee that output video frame rate will not be higher than this value.
11702 A color constancy variation filter which estimates scene illumination via grey edge algorithm
11703 and corrects the scene colors accordingly.
11705 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
11707 The filter accepts the following options:
11711 The order of differentiation to be applied on the scene. Must be chosen in the range
11712 [0,2] and default value is 1.
11715 The Minkowski parameter to be used for calculating the Minkowski distance. Must
11716 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
11717 max value instead of calculating Minkowski distance.
11720 The standard deviation of Gaussian blur to be applied on the scene. Must be
11721 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
11722 can't be equal to 0 if @var{difford} is greater than 0.
11725 @subsection Examples
11731 greyedge=difford=1:minknorm=5:sigma=2
11737 greyedge=difford=1:minknorm=0:sigma=2
11745 Apply a Hald CLUT to a video stream.
11747 First input is the video stream to process, and second one is the Hald CLUT.
11748 The Hald CLUT input can be a simple picture or a complete video stream.
11750 The filter accepts the following options:
11754 Force termination when the shortest input terminates. Default is @code{0}.
11756 Continue applying the last CLUT after the end of the stream. A value of
11757 @code{0} disable the filter after the last frame of the CLUT is reached.
11758 Default is @code{1}.
11761 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
11762 filters share the same internals).
11764 This filter also supports the @ref{framesync} options.
11766 More information about the Hald CLUT can be found on Eskil Steenberg's website
11767 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
11769 @subsection Workflow examples
11771 @subsubsection Hald CLUT video stream
11773 Generate an identity Hald CLUT stream altered with various effects:
11775 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
11778 Note: make sure you use a lossless codec.
11780 Then use it with @code{haldclut} to apply it on some random stream:
11782 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
11785 The Hald CLUT will be applied to the 10 first seconds (duration of
11786 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
11787 to the remaining frames of the @code{mandelbrot} stream.
11789 @subsubsection Hald CLUT with preview
11791 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
11792 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
11793 biggest possible square starting at the top left of the picture. The remaining
11794 padding pixels (bottom or right) will be ignored. This area can be used to add
11795 a preview of the Hald CLUT.
11797 Typically, the following generated Hald CLUT will be supported by the
11798 @code{haldclut} filter:
11801 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
11802 pad=iw+320 [padded_clut];
11803 smptebars=s=320x256, split [a][b];
11804 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
11805 [main][b] overlay=W-320" -frames:v 1 clut.png
11808 It contains the original and a preview of the effect of the CLUT: SMPTE color
11809 bars are displayed on the right-top, and below the same color bars processed by
11812 Then, the effect of this Hald CLUT can be visualized with:
11814 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
11819 Flip the input video horizontally.
11821 For example, to horizontally flip the input video with @command{ffmpeg}:
11823 ffmpeg -i in.avi -vf "hflip" out.avi
11827 This filter applies a global color histogram equalization on a
11830 It can be used to correct video that has a compressed range of pixel
11831 intensities. The filter redistributes the pixel intensities to
11832 equalize their distribution across the intensity range. It may be
11833 viewed as an "automatically adjusting contrast filter". This filter is
11834 useful only for correcting degraded or poorly captured source
11837 The filter accepts the following options:
11841 Determine the amount of equalization to be applied. As the strength
11842 is reduced, the distribution of pixel intensities more-and-more
11843 approaches that of the input frame. The value must be a float number
11844 in the range [0,1] and defaults to 0.200.
11847 Set the maximum intensity that can generated and scale the output
11848 values appropriately. The strength should be set as desired and then
11849 the intensity can be limited if needed to avoid washing-out. The value
11850 must be a float number in the range [0,1] and defaults to 0.210.
11853 Set the antibanding level. If enabled the filter will randomly vary
11854 the luminance of output pixels by a small amount to avoid banding of
11855 the histogram. Possible values are @code{none}, @code{weak} or
11856 @code{strong}. It defaults to @code{none}.
11862 Compute and draw a color distribution histogram for the input video.
11864 The computed histogram is a representation of the color component
11865 distribution in an image.
11867 Standard histogram displays the color components distribution in an image.
11868 Displays color graph for each color component. Shows distribution of
11869 the Y, U, V, A or R, G, B components, depending on input format, in the
11870 current frame. Below each graph a color component scale meter is shown.
11872 The filter accepts the following options:
11876 Set height of level. Default value is @code{200}.
11877 Allowed range is [50, 2048].
11880 Set height of color scale. Default value is @code{12}.
11881 Allowed range is [0, 40].
11885 It accepts the following values:
11888 Per color component graphs are placed below each other.
11891 Per color component graphs are placed side by side.
11894 Presents information identical to that in the @code{parade}, except
11895 that the graphs representing color components are superimposed directly
11898 Default is @code{stack}.
11901 Set mode. Can be either @code{linear}, or @code{logarithmic}.
11902 Default is @code{linear}.
11905 Set what color components to display.
11906 Default is @code{7}.
11909 Set foreground opacity. Default is @code{0.7}.
11912 Set background opacity. Default is @code{0.5}.
11915 @subsection Examples
11920 Calculate and draw histogram:
11922 ffplay -i input -vf histogram
11930 This is a high precision/quality 3d denoise filter. It aims to reduce
11931 image noise, producing smooth images and making still images really
11932 still. It should enhance compressibility.
11934 It accepts the following optional parameters:
11938 A non-negative floating point number which specifies spatial luma strength.
11939 It defaults to 4.0.
11941 @item chroma_spatial
11942 A non-negative floating point number which specifies spatial chroma strength.
11943 It defaults to 3.0*@var{luma_spatial}/4.0.
11946 A floating point number which specifies luma temporal strength. It defaults to
11947 6.0*@var{luma_spatial}/4.0.
11950 A floating point number which specifies chroma temporal strength. It defaults to
11951 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
11954 @subsection Commands
11955 This filter supports same @ref{commands} as options.
11956 The command accepts the same syntax of the corresponding option.
11958 If the specified expression is not valid, it is kept at its current
11961 @anchor{hwdownload}
11962 @section hwdownload
11964 Download hardware frames to system memory.
11966 The input must be in hardware frames, and the output a non-hardware format.
11967 Not all formats will be supported on the output - it may be necessary to insert
11968 an additional @option{format} filter immediately following in the graph to get
11969 the output in a supported format.
11973 Map hardware frames to system memory or to another device.
11975 This filter has several different modes of operation; which one is used depends
11976 on the input and output formats:
11979 Hardware frame input, normal frame output
11981 Map the input frames to system memory and pass them to the output. If the
11982 original hardware frame is later required (for example, after overlaying
11983 something else on part of it), the @option{hwmap} filter can be used again
11984 in the next mode to retrieve it.
11986 Normal frame input, hardware frame output
11988 If the input is actually a software-mapped hardware frame, then unmap it -
11989 that is, return the original hardware frame.
11991 Otherwise, a device must be provided. Create new hardware surfaces on that
11992 device for the output, then map them back to the software format at the input
11993 and give those frames to the preceding filter. This will then act like the
11994 @option{hwupload} filter, but may be able to avoid an additional copy when
11995 the input is already in a compatible format.
11997 Hardware frame input and output
11999 A device must be supplied for the output, either directly or with the
12000 @option{derive_device} option. The input and output devices must be of
12001 different types and compatible - the exact meaning of this is
12002 system-dependent, but typically it means that they must refer to the same
12003 underlying hardware context (for example, refer to the same graphics card).
12005 If the input frames were originally created on the output device, then unmap
12006 to retrieve the original frames.
12008 Otherwise, map the frames to the output device - create new hardware frames
12009 on the output corresponding to the frames on the input.
12012 The following additional parameters are accepted:
12016 Set the frame mapping mode. Some combination of:
12019 The mapped frame should be readable.
12021 The mapped frame should be writeable.
12023 The mapping will always overwrite the entire frame.
12025 This may improve performance in some cases, as the original contents of the
12026 frame need not be loaded.
12028 The mapping must not involve any copying.
12030 Indirect mappings to copies of frames are created in some cases where either
12031 direct mapping is not possible or it would have unexpected properties.
12032 Setting this flag ensures that the mapping is direct and will fail if that is
12035 Defaults to @var{read+write} if not specified.
12037 @item derive_device @var{type}
12038 Rather than using the device supplied at initialisation, instead derive a new
12039 device of type @var{type} from the device the input frames exist on.
12042 In a hardware to hardware mapping, map in reverse - create frames in the sink
12043 and map them back to the source. This may be necessary in some cases where
12044 a mapping in one direction is required but only the opposite direction is
12045 supported by the devices being used.
12047 This option is dangerous - it may break the preceding filter in undefined
12048 ways if there are any additional constraints on that filter's output.
12049 Do not use it without fully understanding the implications of its use.
12055 Upload system memory frames to hardware surfaces.
12057 The device to upload to must be supplied when the filter is initialised. If
12058 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12059 option or with the @option{derive_device} option. The input and output devices
12060 must be of different types and compatible - the exact meaning of this is
12061 system-dependent, but typically it means that they must refer to the same
12062 underlying hardware context (for example, refer to the same graphics card).
12064 The following additional parameters are accepted:
12067 @item derive_device @var{type}
12068 Rather than using the device supplied at initialisation, instead derive a new
12069 device of type @var{type} from the device the input frames exist on.
12072 @anchor{hwupload_cuda}
12073 @section hwupload_cuda
12075 Upload system memory frames to a CUDA device.
12077 It accepts the following optional parameters:
12081 The number of the CUDA device to use
12086 Apply a high-quality magnification filter designed for pixel art. This filter
12087 was originally created by Maxim Stepin.
12089 It accepts the following option:
12093 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12094 @code{hq3x} and @code{4} for @code{hq4x}.
12095 Default is @code{3}.
12099 Stack input videos horizontally.
12101 All streams must be of same pixel format and of same height.
12103 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12104 to create same output.
12106 The filter accepts the following option:
12110 Set number of input streams. Default is 2.
12113 If set to 1, force the output to terminate when the shortest input
12114 terminates. Default value is 0.
12119 Modify the hue and/or the saturation of the input.
12121 It accepts the following parameters:
12125 Specify the hue angle as a number of degrees. It accepts an expression,
12126 and defaults to "0".
12129 Specify the saturation in the [-10,10] range. It accepts an expression and
12133 Specify the hue angle as a number of radians. It accepts an
12134 expression, and defaults to "0".
12137 Specify the brightness in the [-10,10] range. It accepts an expression and
12141 @option{h} and @option{H} are mutually exclusive, and can't be
12142 specified at the same time.
12144 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12145 expressions containing the following constants:
12149 frame count of the input frame starting from 0
12152 presentation timestamp of the input frame expressed in time base units
12155 frame rate of the input video, NAN if the input frame rate is unknown
12158 timestamp expressed in seconds, NAN if the input timestamp is unknown
12161 time base of the input video
12164 @subsection Examples
12168 Set the hue to 90 degrees and the saturation to 1.0:
12174 Same command but expressing the hue in radians:
12180 Rotate hue and make the saturation swing between 0
12181 and 2 over a period of 1 second:
12183 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12187 Apply a 3 seconds saturation fade-in effect starting at 0:
12189 hue="s=min(t/3\,1)"
12192 The general fade-in expression can be written as:
12194 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12198 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12200 hue="s=max(0\, min(1\, (8-t)/3))"
12203 The general fade-out expression can be written as:
12205 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12210 @subsection Commands
12212 This filter supports the following commands:
12218 Modify the hue and/or the saturation and/or brightness of the input video.
12219 The command accepts the same syntax of the corresponding option.
12221 If the specified expression is not valid, it is kept at its current
12225 @section hysteresis
12227 Grow first stream into second stream by connecting components.
12228 This makes it possible to build more robust edge masks.
12230 This filter accepts the following options:
12234 Set which planes will be processed as bitmap, unprocessed planes will be
12235 copied from first stream.
12236 By default value 0xf, all planes will be processed.
12239 Set threshold which is used in filtering. If pixel component value is higher than
12240 this value filter algorithm for connecting components is activated.
12241 By default value is 0.
12246 Detect video interlacing type.
12248 This filter tries to detect if the input frames are interlaced, progressive,
12249 top or bottom field first. It will also try to detect fields that are
12250 repeated between adjacent frames (a sign of telecine).
12252 Single frame detection considers only immediately adjacent frames when classifying each frame.
12253 Multiple frame detection incorporates the classification history of previous frames.
12255 The filter will log these metadata values:
12258 @item single.current_frame
12259 Detected type of current frame using single-frame detection. One of:
12260 ``tff'' (top field first), ``bff'' (bottom field first),
12261 ``progressive'', or ``undetermined''
12264 Cumulative number of frames detected as top field first using single-frame detection.
12267 Cumulative number of frames detected as top field first using multiple-frame detection.
12270 Cumulative number of frames detected as bottom field first using single-frame detection.
12272 @item multiple.current_frame
12273 Detected type of current frame using multiple-frame detection. One of:
12274 ``tff'' (top field first), ``bff'' (bottom field first),
12275 ``progressive'', or ``undetermined''
12278 Cumulative number of frames detected as bottom field first using multiple-frame detection.
12280 @item single.progressive
12281 Cumulative number of frames detected as progressive using single-frame detection.
12283 @item multiple.progressive
12284 Cumulative number of frames detected as progressive using multiple-frame detection.
12286 @item single.undetermined
12287 Cumulative number of frames that could not be classified using single-frame detection.
12289 @item multiple.undetermined
12290 Cumulative number of frames that could not be classified using multiple-frame detection.
12292 @item repeated.current_frame
12293 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
12295 @item repeated.neither
12296 Cumulative number of frames with no repeated field.
12299 Cumulative number of frames with the top field repeated from the previous frame's top field.
12301 @item repeated.bottom
12302 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
12305 The filter accepts the following options:
12309 Set interlacing threshold.
12311 Set progressive threshold.
12313 Threshold for repeated field detection.
12315 Number of frames after which a given frame's contribution to the
12316 statistics is halved (i.e., it contributes only 0.5 to its
12317 classification). The default of 0 means that all frames seen are given
12318 full weight of 1.0 forever.
12319 @item analyze_interlaced_flag
12320 When this is not 0 then idet will use the specified number of frames to determine
12321 if the interlaced flag is accurate, it will not count undetermined frames.
12322 If the flag is found to be accurate it will be used without any further
12323 computations, if it is found to be inaccurate it will be cleared without any
12324 further computations. This allows inserting the idet filter as a low computational
12325 method to clean up the interlaced flag
12330 Deinterleave or interleave fields.
12332 This filter allows one to process interlaced images fields without
12333 deinterlacing them. Deinterleaving splits the input frame into 2
12334 fields (so called half pictures). Odd lines are moved to the top
12335 half of the output image, even lines to the bottom half.
12336 You can process (filter) them independently and then re-interleave them.
12338 The filter accepts the following options:
12342 @item chroma_mode, c
12343 @item alpha_mode, a
12344 Available values for @var{luma_mode}, @var{chroma_mode} and
12345 @var{alpha_mode} are:
12351 @item deinterleave, d
12352 Deinterleave fields, placing one above the other.
12354 @item interleave, i
12355 Interleave fields. Reverse the effect of deinterleaving.
12357 Default value is @code{none}.
12359 @item luma_swap, ls
12360 @item chroma_swap, cs
12361 @item alpha_swap, as
12362 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
12365 @subsection Commands
12367 This filter supports the all above options as @ref{commands}.
12371 Apply inflate effect to the video.
12373 This filter replaces the pixel by the local(3x3) average by taking into account
12374 only values higher than the pixel.
12376 It accepts the following options:
12383 Limit the maximum change for each plane, default is 65535.
12384 If 0, plane will remain unchanged.
12387 @subsection Commands
12389 This filter supports the all above options as @ref{commands}.
12393 Simple interlacing filter from progressive contents. This interleaves upper (or
12394 lower) lines from odd frames with lower (or upper) lines from even frames,
12395 halving the frame rate and preserving image height.
12398 Original Original New Frame
12399 Frame 'j' Frame 'j+1' (tff)
12400 ========== =========== ==================
12401 Line 0 --------------------> Frame 'j' Line 0
12402 Line 1 Line 1 ----> Frame 'j+1' Line 1
12403 Line 2 ---------------------> Frame 'j' Line 2
12404 Line 3 Line 3 ----> Frame 'j+1' Line 3
12406 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
12409 It accepts the following optional parameters:
12413 This determines whether the interlaced frame is taken from the even
12414 (tff - default) or odd (bff) lines of the progressive frame.
12417 Vertical lowpass filter to avoid twitter interlacing and
12418 reduce moire patterns.
12422 Disable vertical lowpass filter
12425 Enable linear filter (default)
12428 Enable complex filter. This will slightly less reduce twitter and moire
12429 but better retain detail and subjective sharpness impression.
12436 Deinterlace input video by applying Donald Graft's adaptive kernel
12437 deinterling. Work on interlaced parts of a video to produce
12438 progressive frames.
12440 The description of the accepted parameters follows.
12444 Set the threshold which affects the filter's tolerance when
12445 determining if a pixel line must be processed. It must be an integer
12446 in the range [0,255] and defaults to 10. A value of 0 will result in
12447 applying the process on every pixels.
12450 Paint pixels exceeding the threshold value to white if set to 1.
12454 Set the fields order. Swap fields if set to 1, leave fields alone if
12458 Enable additional sharpening if set to 1. Default is 0.
12461 Enable twoway sharpening if set to 1. Default is 0.
12464 @subsection Examples
12468 Apply default values:
12470 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
12474 Enable additional sharpening:
12480 Paint processed pixels in white:
12488 Slowly update darker pixels.
12490 This filter makes short flashes of light appear longer.
12491 This filter accepts the following options:
12495 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
12498 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
12501 @section lenscorrection
12503 Correct radial lens distortion
12505 This filter can be used to correct for radial distortion as can result from the use
12506 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
12507 one can use tools available for example as part of opencv or simply trial-and-error.
12508 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
12509 and extract the k1 and k2 coefficients from the resulting matrix.
12511 Note that effectively the same filter is available in the open-source tools Krita and
12512 Digikam from the KDE project.
12514 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
12515 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
12516 brightness distribution, so you may want to use both filters together in certain
12517 cases, though you will have to take care of ordering, i.e. whether vignetting should
12518 be applied before or after lens correction.
12520 @subsection Options
12522 The filter accepts the following options:
12526 Relative x-coordinate of the focal point of the image, and thereby the center of the
12527 distortion. This value has a range [0,1] and is expressed as fractions of the image
12528 width. Default is 0.5.
12530 Relative y-coordinate of the focal point of the image, and thereby the center of the
12531 distortion. This value has a range [0,1] and is expressed as fractions of the image
12532 height. Default is 0.5.
12534 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
12535 no correction. Default is 0.
12537 Coefficient of the double quadratic correction term. This value has a range [-1,1].
12538 0 means no correction. Default is 0.
12541 The formula that generates the correction is:
12543 @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)
12545 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
12546 distances from the focal point in the source and target images, respectively.
12550 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
12552 The @code{lensfun} filter requires the camera make, camera model, and lens model
12553 to apply the lens correction. The filter will load the lensfun database and
12554 query it to find the corresponding camera and lens entries in the database. As
12555 long as these entries can be found with the given options, the filter can
12556 perform corrections on frames. Note that incomplete strings will result in the
12557 filter choosing the best match with the given options, and the filter will
12558 output the chosen camera and lens models (logged with level "info"). You must
12559 provide the make, camera model, and lens model as they are required.
12561 The filter accepts the following options:
12565 The make of the camera (for example, "Canon"). This option is required.
12568 The model of the camera (for example, "Canon EOS 100D"). This option is
12572 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
12573 option is required.
12576 The type of correction to apply. The following values are valid options:
12580 Enables fixing lens vignetting.
12583 Enables fixing lens geometry. This is the default.
12586 Enables fixing chromatic aberrations.
12589 Enables fixing lens vignetting and lens geometry.
12592 Enables fixing lens vignetting and chromatic aberrations.
12595 Enables fixing both lens geometry and chromatic aberrations.
12598 Enables all possible corrections.
12602 The focal length of the image/video (zoom; expected constant for video). For
12603 example, a 18--55mm lens has focal length range of [18--55], so a value in that
12604 range should be chosen when using that lens. Default 18.
12607 The aperture of the image/video (expected constant for video). Note that
12608 aperture is only used for vignetting correction. Default 3.5.
12610 @item focus_distance
12611 The focus distance of the image/video (expected constant for video). Note that
12612 focus distance is only used for vignetting and only slightly affects the
12613 vignetting correction process. If unknown, leave it at the default value (which
12617 The scale factor which is applied after transformation. After correction the
12618 video is no longer necessarily rectangular. This parameter controls how much of
12619 the resulting image is visible. The value 0 means that a value will be chosen
12620 automatically such that there is little or no unmapped area in the output
12621 image. 1.0 means that no additional scaling is done. Lower values may result
12622 in more of the corrected image being visible, while higher values may avoid
12623 unmapped areas in the output.
12625 @item target_geometry
12626 The target geometry of the output image/video. The following values are valid
12630 @item rectilinear (default)
12633 @item equirectangular
12634 @item fisheye_orthographic
12635 @item fisheye_stereographic
12636 @item fisheye_equisolid
12637 @item fisheye_thoby
12640 Apply the reverse of image correction (instead of correcting distortion, apply
12643 @item interpolation
12644 The type of interpolation used when correcting distortion. The following values
12649 @item linear (default)
12654 @subsection Examples
12658 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
12659 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
12663 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
12667 Apply the same as before, but only for the first 5 seconds of video.
12670 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
12677 Obtain the VMAF (Video Multi-Method Assessment Fusion)
12678 score between two input videos.
12680 The obtained VMAF score is printed through the logging system.
12682 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
12683 After installing the library it can be enabled using:
12684 @code{./configure --enable-libvmaf --enable-version3}.
12685 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
12687 The filter has following options:
12691 Set the model path which is to be used for SVM.
12692 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
12695 Set the file path to be used to store logs.
12698 Set the format of the log file (xml or json).
12700 @item enable_transform
12701 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
12702 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
12703 Default value: @code{false}
12706 Invokes the phone model which will generate VMAF scores higher than in the
12707 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
12708 Default value: @code{false}
12711 Enables computing psnr along with vmaf.
12712 Default value: @code{false}
12715 Enables computing ssim along with vmaf.
12716 Default value: @code{false}
12719 Enables computing ms_ssim along with vmaf.
12720 Default value: @code{false}
12723 Set the pool method to be used for computing vmaf.
12724 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
12727 Set number of threads to be used when computing vmaf.
12728 Default value: @code{0}, which makes use of all available logical processors.
12731 Set interval for frame subsampling used when computing vmaf.
12732 Default value: @code{1}
12734 @item enable_conf_interval
12735 Enables confidence interval.
12736 Default value: @code{false}
12739 This filter also supports the @ref{framesync} options.
12741 @subsection Examples
12744 On the below examples the input file @file{main.mpg} being processed is
12745 compared with the reference file @file{ref.mpg}.
12748 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
12752 Example with options:
12754 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
12758 Example with options and different containers:
12760 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 -
12766 Limits the pixel components values to the specified range [min, max].
12768 The filter accepts the following options:
12772 Lower bound. Defaults to the lowest allowed value for the input.
12775 Upper bound. Defaults to the highest allowed value for the input.
12778 Specify which planes will be processed. Defaults to all available.
12785 The filter accepts the following options:
12789 Set the number of loops. Setting this value to -1 will result in infinite loops.
12793 Set maximal size in number of frames. Default is 0.
12796 Set first frame of loop. Default is 0.
12799 @subsection Examples
12803 Loop single first frame infinitely:
12805 loop=loop=-1:size=1:start=0
12809 Loop single first frame 10 times:
12811 loop=loop=10:size=1:start=0
12815 Loop 10 first frames 5 times:
12817 loop=loop=5:size=10:start=0
12823 Apply a 1D LUT to an input video.
12825 The filter accepts the following options:
12829 Set the 1D LUT file name.
12831 Currently supported formats:
12840 Select interpolation mode.
12842 Available values are:
12846 Use values from the nearest defined point.
12848 Interpolate values using the linear interpolation.
12850 Interpolate values using the cosine interpolation.
12852 Interpolate values using the cubic interpolation.
12854 Interpolate values using the spline interpolation.
12861 Apply a 3D LUT to an input video.
12863 The filter accepts the following options:
12867 Set the 3D LUT file name.
12869 Currently supported formats:
12883 Select interpolation mode.
12885 Available values are:
12889 Use values from the nearest defined point.
12891 Interpolate values using the 8 points defining a cube.
12893 Interpolate values using a tetrahedron.
12899 Turn certain luma values into transparency.
12901 The filter accepts the following options:
12905 Set the luma which will be used as base for transparency.
12906 Default value is @code{0}.
12909 Set the range of luma values to be keyed out.
12910 Default value is @code{0.01}.
12913 Set the range of softness. Default value is @code{0}.
12914 Use this to control gradual transition from zero to full transparency.
12917 @subsection Commands
12918 This filter supports same @ref{commands} as options.
12919 The command accepts the same syntax of the corresponding option.
12921 If the specified expression is not valid, it is kept at its current
12924 @section lut, lutrgb, lutyuv
12926 Compute a look-up table for binding each pixel component input value
12927 to an output value, and apply it to the input video.
12929 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
12930 to an RGB input video.
12932 These filters accept the following parameters:
12935 set first pixel component expression
12937 set second pixel component expression
12939 set third pixel component expression
12941 set fourth pixel component expression, corresponds to the alpha component
12944 set red component expression
12946 set green component expression
12948 set blue component expression
12950 alpha component expression
12953 set Y/luminance component expression
12955 set U/Cb component expression
12957 set V/Cr component expression
12960 Each of them specifies the expression to use for computing the lookup table for
12961 the corresponding pixel component values.
12963 The exact component associated to each of the @var{c*} options depends on the
12966 The @var{lut} filter requires either YUV or RGB pixel formats in input,
12967 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
12969 The expressions can contain the following constants and functions:
12974 The input width and height.
12977 The input value for the pixel component.
12980 The input value, clipped to the @var{minval}-@var{maxval} range.
12983 The maximum value for the pixel component.
12986 The minimum value for the pixel component.
12989 The negated value for the pixel component value, clipped to the
12990 @var{minval}-@var{maxval} range; it corresponds to the expression
12991 "maxval-clipval+minval".
12994 The computed value in @var{val}, clipped to the
12995 @var{minval}-@var{maxval} range.
12997 @item gammaval(gamma)
12998 The computed gamma correction value of the pixel component value,
12999 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13001 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13005 All expressions default to "val".
13007 @subsection Examples
13011 Negate input video:
13013 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13014 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13017 The above is the same as:
13019 lutrgb="r=negval:g=negval:b=negval"
13020 lutyuv="y=negval:u=negval:v=negval"
13030 Remove chroma components, turning the video into a graytone image:
13032 lutyuv="u=128:v=128"
13036 Apply a luma burning effect:
13042 Remove green and blue components:
13048 Set a constant alpha channel value on input:
13050 format=rgba,lutrgb=a="maxval-minval/2"
13054 Correct luminance gamma by a factor of 0.5:
13056 lutyuv=y=gammaval(0.5)
13060 Discard least significant bits of luma:
13062 lutyuv=y='bitand(val, 128+64+32)'
13066 Technicolor like effect:
13068 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13072 @section lut2, tlut2
13074 The @code{lut2} filter takes two input streams and outputs one
13077 The @code{tlut2} (time lut2) filter takes two consecutive frames
13078 from one single stream.
13080 This filter accepts the following parameters:
13083 set first pixel component expression
13085 set second pixel component expression
13087 set third pixel component expression
13089 set fourth pixel component expression, corresponds to the alpha component
13092 set output bit depth, only available for @code{lut2} filter. By default is 0,
13093 which means bit depth is automatically picked from first input format.
13096 Each of them specifies the expression to use for computing the lookup table for
13097 the corresponding pixel component values.
13099 The exact component associated to each of the @var{c*} options depends on the
13102 The expressions can contain the following constants:
13107 The input width and height.
13110 The first input value for the pixel component.
13113 The second input value for the pixel component.
13116 The first input video bit depth.
13119 The second input video bit depth.
13122 All expressions default to "x".
13124 @subsection Examples
13128 Highlight differences between two RGB video streams:
13130 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)'
13134 Highlight differences between two YUV video streams:
13136 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)'
13140 Show max difference between two video streams:
13142 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)))'
13146 @section maskedclamp
13148 Clamp the first input stream with the second input and third input stream.
13150 Returns the value of first stream to be between second input
13151 stream - @code{undershoot} and third input stream + @code{overshoot}.
13153 This filter accepts the following options:
13156 Default value is @code{0}.
13159 Default value is @code{0}.
13162 Set which planes will be processed as bitmap, unprocessed planes will be
13163 copied from first stream.
13164 By default value 0xf, all planes will be processed.
13169 Merge the second and third input stream into output stream using absolute differences
13170 between second input stream and first input stream and absolute difference between
13171 third input stream and first input stream. The picked value will be from second input
13172 stream if second absolute difference is greater than first one or from third input stream
13175 This filter accepts the following options:
13178 Set which planes will be processed as bitmap, unprocessed planes will be
13179 copied from first stream.
13180 By default value 0xf, all planes will be processed.
13183 @section maskedmerge
13185 Merge the first input stream with the second input stream using per pixel
13186 weights in the third input stream.
13188 A value of 0 in the third stream pixel component means that pixel component
13189 from first stream is returned unchanged, while maximum value (eg. 255 for
13190 8-bit videos) means that pixel component from second stream is returned
13191 unchanged. Intermediate values define the amount of merging between both
13192 input stream's pixel components.
13194 This filter accepts the following options:
13197 Set which planes will be processed as bitmap, unprocessed planes will be
13198 copied from first stream.
13199 By default value 0xf, all planes will be processed.
13204 Merge the second and third input stream into output stream using absolute differences
13205 between second input stream and first input stream and absolute difference between
13206 third input stream and first input stream. The picked value will be from second input
13207 stream if second absolute difference is less than first one or from third input stream
13210 This filter accepts the following options:
13213 Set which planes will be processed as bitmap, unprocessed planes will be
13214 copied from first stream.
13215 By default value 0xf, all planes will be processed.
13219 Create mask from input video.
13221 For example it is useful to create motion masks after @code{tblend} filter.
13223 This filter accepts the following options:
13227 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
13230 Set high threshold. Any pixel component higher than this value will be set to max value
13231 allowed for current pixel format.
13234 Set planes to filter, by default all available planes are filtered.
13237 Fill all frame pixels with this value.
13240 Set max average pixel value for frame. If sum of all pixel components is higher that this
13241 average, output frame will be completely filled with value set by @var{fill} option.
13242 Typically useful for scene changes when used in combination with @code{tblend} filter.
13247 Apply motion-compensation deinterlacing.
13249 It needs one field per frame as input and must thus be used together
13250 with yadif=1/3 or equivalent.
13252 This filter accepts the following options:
13255 Set the deinterlacing mode.
13257 It accepts one of the following values:
13262 use iterative motion estimation
13264 like @samp{slow}, but use multiple reference frames.
13266 Default value is @samp{fast}.
13269 Set the picture field parity assumed for the input video. It must be
13270 one of the following values:
13274 assume top field first
13276 assume bottom field first
13279 Default value is @samp{bff}.
13282 Set per-block quantization parameter (QP) used by the internal
13285 Higher values should result in a smoother motion vector field but less
13286 optimal individual vectors. Default value is 1.
13291 Pick median pixel from certain rectangle defined by radius.
13293 This filter accepts the following options:
13297 Set horizontal radius size. Default value is @code{1}.
13298 Allowed range is integer from 1 to 127.
13301 Set which planes to process. Default is @code{15}, which is all available planes.
13304 Set vertical radius size. Default value is @code{0}.
13305 Allowed range is integer from 0 to 127.
13306 If it is 0, value will be picked from horizontal @code{radius} option.
13309 Set median percentile. Default value is @code{0.5}.
13310 Default value of @code{0.5} will pick always median values, while @code{0} will pick
13311 minimum values, and @code{1} maximum values.
13314 @subsection Commands
13315 This filter supports same @ref{commands} as options.
13316 The command accepts the same syntax of the corresponding option.
13318 If the specified expression is not valid, it is kept at its current
13321 @section mergeplanes
13323 Merge color channel components from several video streams.
13325 The filter accepts up to 4 input streams, and merge selected input
13326 planes to the output video.
13328 This filter accepts the following options:
13331 Set input to output plane mapping. Default is @code{0}.
13333 The mappings is specified as a bitmap. It should be specified as a
13334 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
13335 mapping for the first plane of the output stream. 'A' sets the number of
13336 the input stream to use (from 0 to 3), and 'a' the plane number of the
13337 corresponding input to use (from 0 to 3). The rest of the mappings is
13338 similar, 'Bb' describes the mapping for the output stream second
13339 plane, 'Cc' describes the mapping for the output stream third plane and
13340 'Dd' describes the mapping for the output stream fourth plane.
13343 Set output pixel format. Default is @code{yuva444p}.
13346 @subsection Examples
13350 Merge three gray video streams of same width and height into single video stream:
13352 [a0][a1][a2]mergeplanes=0x001020:yuv444p
13356 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
13358 [a0][a1]mergeplanes=0x00010210:yuva444p
13362 Swap Y and A plane in yuva444p stream:
13364 format=yuva444p,mergeplanes=0x03010200:yuva444p
13368 Swap U and V plane in yuv420p stream:
13370 format=yuv420p,mergeplanes=0x000201:yuv420p
13374 Cast a rgb24 clip to yuv444p:
13376 format=rgb24,mergeplanes=0x000102:yuv444p
13382 Estimate and export motion vectors using block matching algorithms.
13383 Motion vectors are stored in frame side data to be used by other filters.
13385 This filter accepts the following options:
13388 Specify the motion estimation method. Accepts one of the following values:
13392 Exhaustive search algorithm.
13394 Three step search algorithm.
13396 Two dimensional logarithmic search algorithm.
13398 New three step search algorithm.
13400 Four step search algorithm.
13402 Diamond search algorithm.
13404 Hexagon-based search algorithm.
13406 Enhanced predictive zonal search algorithm.
13408 Uneven multi-hexagon search algorithm.
13410 Default value is @samp{esa}.
13413 Macroblock size. Default @code{16}.
13416 Search parameter. Default @code{7}.
13419 @section midequalizer
13421 Apply Midway Image Equalization effect using two video streams.
13423 Midway Image Equalization adjusts a pair of images to have the same
13424 histogram, while maintaining their dynamics as much as possible. It's
13425 useful for e.g. matching exposures from a pair of stereo cameras.
13427 This filter has two inputs and one output, which must be of same pixel format, but
13428 may be of different sizes. The output of filter is first input adjusted with
13429 midway histogram of both inputs.
13431 This filter accepts the following option:
13435 Set which planes to process. Default is @code{15}, which is all available planes.
13438 @section minterpolate
13440 Convert the video to specified frame rate using motion interpolation.
13442 This filter accepts the following options:
13445 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}.
13448 Motion interpolation mode. Following values are accepted:
13451 Duplicate previous or next frame for interpolating new ones.
13453 Blend source frames. Interpolated frame is mean of previous and next frames.
13455 Motion compensated interpolation. Following options are effective when this mode is selected:
13459 Motion compensation mode. Following values are accepted:
13462 Overlapped block motion compensation.
13464 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
13466 Default mode is @samp{obmc}.
13469 Motion estimation mode. Following values are accepted:
13472 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
13474 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
13476 Default mode is @samp{bilat}.
13479 The algorithm to be used for motion estimation. Following values are accepted:
13482 Exhaustive search algorithm.
13484 Three step search algorithm.
13486 Two dimensional logarithmic search algorithm.
13488 New three step search algorithm.
13490 Four step search algorithm.
13492 Diamond search algorithm.
13494 Hexagon-based search algorithm.
13496 Enhanced predictive zonal search algorithm.
13498 Uneven multi-hexagon search algorithm.
13500 Default algorithm is @samp{epzs}.
13503 Macroblock size. Default @code{16}.
13506 Motion estimation search parameter. Default @code{32}.
13509 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).
13514 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:
13517 Disable scene change detection.
13519 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
13521 Default method is @samp{fdiff}.
13523 @item scd_threshold
13524 Scene change detection threshold. Default is @code{5.0}.
13529 Mix several video input streams into one video stream.
13531 A description of the accepted options follows.
13535 The number of inputs. If unspecified, it defaults to 2.
13538 Specify weight of each input video stream as sequence.
13539 Each weight is separated by space. If number of weights
13540 is smaller than number of @var{frames} last specified
13541 weight will be used for all remaining unset weights.
13544 Specify scale, if it is set it will be multiplied with sum
13545 of each weight multiplied with pixel values to give final destination
13546 pixel value. By default @var{scale} is auto scaled to sum of weights.
13549 Specify how end of stream is determined.
13552 The duration of the longest input. (default)
13555 The duration of the shortest input.
13558 The duration of the first input.
13562 @section mpdecimate
13564 Drop frames that do not differ greatly from the previous frame in
13565 order to reduce frame rate.
13567 The main use of this filter is for very-low-bitrate encoding
13568 (e.g. streaming over dialup modem), but it could in theory be used for
13569 fixing movies that were inverse-telecined incorrectly.
13571 A description of the accepted options follows.
13575 Set the maximum number of consecutive frames which can be dropped (if
13576 positive), or the minimum interval between dropped frames (if
13577 negative). If the value is 0, the frame is dropped disregarding the
13578 number of previous sequentially dropped frames.
13580 Default value is 0.
13585 Set the dropping threshold values.
13587 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
13588 represent actual pixel value differences, so a threshold of 64
13589 corresponds to 1 unit of difference for each pixel, or the same spread
13590 out differently over the block.
13592 A frame is a candidate for dropping if no 8x8 blocks differ by more
13593 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
13594 meaning the whole image) differ by more than a threshold of @option{lo}.
13596 Default value for @option{hi} is 64*12, default value for @option{lo} is
13597 64*5, and default value for @option{frac} is 0.33.
13603 Negate (invert) the input video.
13605 It accepts the following option:
13610 With value 1, it negates the alpha component, if present. Default value is 0.
13616 Denoise frames using Non-Local Means algorithm.
13618 Each pixel is adjusted by looking for other pixels with similar contexts. This
13619 context similarity is defined by comparing their surrounding patches of size
13620 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
13623 Note that the research area defines centers for patches, which means some
13624 patches will be made of pixels outside that research area.
13626 The filter accepts the following options.
13630 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
13633 Set patch size. Default is 7. Must be odd number in range [0, 99].
13636 Same as @option{p} but for chroma planes.
13638 The default value is @var{0} and means automatic.
13641 Set research size. Default is 15. Must be odd number in range [0, 99].
13644 Same as @option{r} but for chroma planes.
13646 The default value is @var{0} and means automatic.
13651 Deinterlace video using neural network edge directed interpolation.
13653 This filter accepts the following options:
13657 Mandatory option, without binary file filter can not work.
13658 Currently file can be found here:
13659 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
13662 Set which frames to deinterlace, by default it is @code{all}.
13663 Can be @code{all} or @code{interlaced}.
13666 Set mode of operation.
13668 Can be one of the following:
13672 Use frame flags, both fields.
13674 Use frame flags, single field.
13676 Use top field only.
13678 Use bottom field only.
13680 Use both fields, top first.
13682 Use both fields, bottom first.
13686 Set which planes to process, by default filter process all frames.
13689 Set size of local neighborhood around each pixel, used by the predictor neural
13692 Can be one of the following:
13705 Set the number of neurons in predictor neural network.
13706 Can be one of the following:
13717 Controls the number of different neural network predictions that are blended
13718 together to compute the final output value. Can be @code{fast}, default or
13722 Set which set of weights to use in the predictor.
13723 Can be one of the following:
13727 weights trained to minimize absolute error
13729 weights trained to minimize squared error
13733 Controls whether or not the prescreener neural network is used to decide
13734 which pixels should be processed by the predictor neural network and which
13735 can be handled by simple cubic interpolation.
13736 The prescreener is trained to know whether cubic interpolation will be
13737 sufficient for a pixel or whether it should be predicted by the predictor nn.
13738 The computational complexity of the prescreener nn is much less than that of
13739 the predictor nn. Since most pixels can be handled by cubic interpolation,
13740 using the prescreener generally results in much faster processing.
13741 The prescreener is pretty accurate, so the difference between using it and not
13742 using it is almost always unnoticeable.
13744 Can be one of the following:
13752 Default is @code{new}.
13755 Set various debugging flags.
13760 Force libavfilter not to use any of the specified pixel formats for the
13761 input to the next filter.
13763 It accepts the following parameters:
13767 A '|'-separated list of pixel format names, such as
13768 pix_fmts=yuv420p|monow|rgb24".
13772 @subsection Examples
13776 Force libavfilter to use a format different from @var{yuv420p} for the
13777 input to the vflip filter:
13779 noformat=pix_fmts=yuv420p,vflip
13783 Convert the input video to any of the formats not contained in the list:
13785 noformat=yuv420p|yuv444p|yuv410p
13791 Add noise on video input frame.
13793 The filter accepts the following options:
13801 Set noise seed for specific pixel component or all pixel components in case
13802 of @var{all_seed}. Default value is @code{123457}.
13804 @item all_strength, alls
13805 @item c0_strength, c0s
13806 @item c1_strength, c1s
13807 @item c2_strength, c2s
13808 @item c3_strength, c3s
13809 Set noise strength for specific pixel component or all pixel components in case
13810 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
13812 @item all_flags, allf
13813 @item c0_flags, c0f
13814 @item c1_flags, c1f
13815 @item c2_flags, c2f
13816 @item c3_flags, c3f
13817 Set pixel component flags or set flags for all components if @var{all_flags}.
13818 Available values for component flags are:
13821 averaged temporal noise (smoother)
13823 mix random noise with a (semi)regular pattern
13825 temporal noise (noise pattern changes between frames)
13827 uniform noise (gaussian otherwise)
13831 @subsection Examples
13833 Add temporal and uniform noise to input video:
13835 noise=alls=20:allf=t+u
13840 Normalize RGB video (aka histogram stretching, contrast stretching).
13841 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
13843 For each channel of each frame, the filter computes the input range and maps
13844 it linearly to the user-specified output range. The output range defaults
13845 to the full dynamic range from pure black to pure white.
13847 Temporal smoothing can be used on the input range to reduce flickering (rapid
13848 changes in brightness) caused when small dark or bright objects enter or leave
13849 the scene. This is similar to the auto-exposure (automatic gain control) on a
13850 video camera, and, like a video camera, it may cause a period of over- or
13851 under-exposure of the video.
13853 The R,G,B channels can be normalized independently, which may cause some
13854 color shifting, or linked together as a single channel, which prevents
13855 color shifting. Linked normalization preserves hue. Independent normalization
13856 does not, so it can be used to remove some color casts. Independent and linked
13857 normalization can be combined in any ratio.
13859 The normalize filter accepts the following options:
13864 Colors which define the output range. The minimum input value is mapped to
13865 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
13866 The defaults are black and white respectively. Specifying white for
13867 @var{blackpt} and black for @var{whitept} will give color-inverted,
13868 normalized video. Shades of grey can be used to reduce the dynamic range
13869 (contrast). Specifying saturated colors here can create some interesting
13873 The number of previous frames to use for temporal smoothing. The input range
13874 of each channel is smoothed using a rolling average over the current frame
13875 and the @var{smoothing} previous frames. The default is 0 (no temporal
13879 Controls the ratio of independent (color shifting) channel normalization to
13880 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
13881 independent. Defaults to 1.0 (fully independent).
13884 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
13885 expensive no-op. Defaults to 1.0 (full strength).
13889 @subsection Commands
13890 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
13891 The command accepts the same syntax of the corresponding option.
13893 If the specified expression is not valid, it is kept at its current
13896 @subsection Examples
13898 Stretch video contrast to use the full dynamic range, with no temporal
13899 smoothing; may flicker depending on the source content:
13901 normalize=blackpt=black:whitept=white:smoothing=0
13904 As above, but with 50 frames of temporal smoothing; flicker should be
13905 reduced, depending on the source content:
13907 normalize=blackpt=black:whitept=white:smoothing=50
13910 As above, but with hue-preserving linked channel normalization:
13912 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
13915 As above, but with half strength:
13917 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
13920 Map the darkest input color to red, the brightest input color to cyan:
13922 normalize=blackpt=red:whitept=cyan
13927 Pass the video source unchanged to the output.
13930 Optical Character Recognition
13932 This filter uses Tesseract for optical character recognition. To enable
13933 compilation of this filter, you need to configure FFmpeg with
13934 @code{--enable-libtesseract}.
13936 It accepts the following options:
13940 Set datapath to tesseract data. Default is to use whatever was
13941 set at installation.
13944 Set language, default is "eng".
13947 Set character whitelist.
13950 Set character blacklist.
13953 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
13954 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
13958 Apply a video transform using libopencv.
13960 To enable this filter, install the libopencv library and headers and
13961 configure FFmpeg with @code{--enable-libopencv}.
13963 It accepts the following parameters:
13968 The name of the libopencv filter to apply.
13970 @item filter_params
13971 The parameters to pass to the libopencv filter. If not specified, the default
13972 values are assumed.
13976 Refer to the official libopencv documentation for more precise
13978 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
13980 Several libopencv filters are supported; see the following subsections.
13985 Dilate an image by using a specific structuring element.
13986 It corresponds to the libopencv function @code{cvDilate}.
13988 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
13990 @var{struct_el} represents a structuring element, and has the syntax:
13991 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
13993 @var{cols} and @var{rows} represent the number of columns and rows of
13994 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
13995 point, and @var{shape} the shape for the structuring element. @var{shape}
13996 must be "rect", "cross", "ellipse", or "custom".
13998 If the value for @var{shape} is "custom", it must be followed by a
13999 string of the form "=@var{filename}". The file with name
14000 @var{filename} is assumed to represent a binary image, with each
14001 printable character corresponding to a bright pixel. When a custom
14002 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14003 or columns and rows of the read file are assumed instead.
14005 The default value for @var{struct_el} is "3x3+0x0/rect".
14007 @var{nb_iterations} specifies the number of times the transform is
14008 applied to the image, and defaults to 1.
14012 # Use the default values
14015 # Dilate using a structuring element with a 5x5 cross, iterating two times
14016 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14018 # Read the shape from the file diamond.shape, iterating two times.
14019 # The file diamond.shape may contain a pattern of characters like this
14025 # The specified columns and rows are ignored
14026 # but the anchor point coordinates are not
14027 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14032 Erode an image by using a specific structuring element.
14033 It corresponds to the libopencv function @code{cvErode}.
14035 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14036 with the same syntax and semantics as the @ref{dilate} filter.
14040 Smooth the input video.
14042 The filter takes the following parameters:
14043 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14045 @var{type} is the type of smooth filter to apply, and must be one of
14046 the following values: "blur", "blur_no_scale", "median", "gaussian",
14047 or "bilateral". The default value is "gaussian".
14049 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14050 depends on the smooth type. @var{param1} and
14051 @var{param2} accept integer positive values or 0. @var{param3} and
14052 @var{param4} accept floating point values.
14054 The default value for @var{param1} is 3. The default value for the
14055 other parameters is 0.
14057 These parameters correspond to the parameters assigned to the
14058 libopencv function @code{cvSmooth}.
14060 @section oscilloscope
14062 2D Video Oscilloscope.
14064 Useful to measure spatial impulse, step responses, chroma delays, etc.
14066 It accepts the following parameters:
14070 Set scope center x position.
14073 Set scope center y position.
14076 Set scope size, relative to frame diagonal.
14079 Set scope tilt/rotation.
14085 Set trace center x position.
14088 Set trace center y position.
14091 Set trace width, relative to width of frame.
14094 Set trace height, relative to height of frame.
14097 Set which components to trace. By default it traces first three components.
14100 Draw trace grid. By default is enabled.
14103 Draw some statistics. By default is enabled.
14106 Draw scope. By default is enabled.
14109 @subsection Commands
14110 This filter supports same @ref{commands} as options.
14111 The command accepts the same syntax of the corresponding option.
14113 If the specified expression is not valid, it is kept at its current
14116 @subsection Examples
14120 Inspect full first row of video frame.
14122 oscilloscope=x=0.5:y=0:s=1
14126 Inspect full last row of video frame.
14128 oscilloscope=x=0.5:y=1:s=1
14132 Inspect full 5th line of video frame of height 1080.
14134 oscilloscope=x=0.5:y=5/1080:s=1
14138 Inspect full last column of video frame.
14140 oscilloscope=x=1:y=0.5:s=1:t=1
14148 Overlay one video on top of another.
14150 It takes two inputs and has one output. The first input is the "main"
14151 video on which the second input is overlaid.
14153 It accepts the following parameters:
14155 A description of the accepted options follows.
14160 Set the expression for the x and y coordinates of the overlaid video
14161 on the main video. Default value is "0" for both expressions. In case
14162 the expression is invalid, it is set to a huge value (meaning that the
14163 overlay will not be displayed within the output visible area).
14166 See @ref{framesync}.
14169 Set when the expressions for @option{x}, and @option{y} are evaluated.
14171 It accepts the following values:
14174 only evaluate expressions once during the filter initialization or
14175 when a command is processed
14178 evaluate expressions for each incoming frame
14181 Default value is @samp{frame}.
14184 See @ref{framesync}.
14187 Set the format for the output video.
14189 It accepts the following values:
14192 force YUV420 output
14195 force YUV422 output
14198 force YUV444 output
14201 force packed RGB output
14204 force planar RGB output
14207 automatically pick format
14210 Default value is @samp{yuv420}.
14213 See @ref{framesync}.
14216 Set format of alpha of the overlaid video, it can be @var{straight} or
14217 @var{premultiplied}. Default is @var{straight}.
14220 The @option{x}, and @option{y} expressions can contain the following
14226 The main input width and height.
14230 The overlay input width and height.
14234 The computed values for @var{x} and @var{y}. They are evaluated for
14239 horizontal and vertical chroma subsample values of the output
14240 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
14244 the number of input frame, starting from 0
14247 the position in the file of the input frame, NAN if unknown
14250 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
14254 This filter also supports the @ref{framesync} options.
14256 Note that the @var{n}, @var{pos}, @var{t} variables are available only
14257 when evaluation is done @emph{per frame}, and will evaluate to NAN
14258 when @option{eval} is set to @samp{init}.
14260 Be aware that frames are taken from each input video in timestamp
14261 order, hence, if their initial timestamps differ, it is a good idea
14262 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
14263 have them begin in the same zero timestamp, as the example for
14264 the @var{movie} filter does.
14266 You can chain together more overlays but you should test the
14267 efficiency of such approach.
14269 @subsection Commands
14271 This filter supports the following commands:
14275 Modify the x and y of the overlay input.
14276 The command accepts the same syntax of the corresponding option.
14278 If the specified expression is not valid, it is kept at its current
14282 @subsection Examples
14286 Draw the overlay at 10 pixels from the bottom right corner of the main
14289 overlay=main_w-overlay_w-10:main_h-overlay_h-10
14292 Using named options the example above becomes:
14294 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
14298 Insert a transparent PNG logo in the bottom left corner of the input,
14299 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
14301 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
14305 Insert 2 different transparent PNG logos (second logo on bottom
14306 right corner) using the @command{ffmpeg} tool:
14308 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
14312 Add a transparent color layer on top of the main video; @code{WxH}
14313 must specify the size of the main input to the overlay filter:
14315 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
14319 Play an original video and a filtered version (here with the deshake
14320 filter) side by side using the @command{ffplay} tool:
14322 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
14325 The above command is the same as:
14327 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
14331 Make a sliding overlay appearing from the left to the right top part of the
14332 screen starting since time 2:
14334 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
14338 Compose output by putting two input videos side to side:
14340 ffmpeg -i left.avi -i right.avi -filter_complex "
14341 nullsrc=size=200x100 [background];
14342 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
14343 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
14344 [background][left] overlay=shortest=1 [background+left];
14345 [background+left][right] overlay=shortest=1:x=100 [left+right]
14350 Mask 10-20 seconds of a video by applying the delogo filter to a section
14352 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
14353 -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]'
14358 Chain several overlays in cascade:
14360 nullsrc=s=200x200 [bg];
14361 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
14362 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
14363 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
14364 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
14365 [in3] null, [mid2] overlay=100:100 [out0]
14372 Apply Overcomplete Wavelet denoiser.
14374 The filter accepts the following options:
14380 Larger depth values will denoise lower frequency components more, but
14381 slow down filtering.
14383 Must be an int in the range 8-16, default is @code{8}.
14385 @item luma_strength, ls
14388 Must be a double value in the range 0-1000, default is @code{1.0}.
14390 @item chroma_strength, cs
14391 Set chroma strength.
14393 Must be a double value in the range 0-1000, default is @code{1.0}.
14399 Add paddings to the input image, and place the original input at the
14400 provided @var{x}, @var{y} coordinates.
14402 It accepts the following parameters:
14407 Specify an expression for the size of the output image with the
14408 paddings added. If the value for @var{width} or @var{height} is 0, the
14409 corresponding input size is used for the output.
14411 The @var{width} expression can reference the value set by the
14412 @var{height} expression, and vice versa.
14414 The default value of @var{width} and @var{height} is 0.
14418 Specify the offsets to place the input image at within the padded area,
14419 with respect to the top/left border of the output image.
14421 The @var{x} expression can reference the value set by the @var{y}
14422 expression, and vice versa.
14424 The default value of @var{x} and @var{y} is 0.
14426 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
14427 so the input image is centered on the padded area.
14430 Specify the color of the padded area. For the syntax of this option,
14431 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
14432 manual,ffmpeg-utils}.
14434 The default value of @var{color} is "black".
14437 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
14439 It accepts the following values:
14443 Only evaluate expressions once during the filter initialization or when
14444 a command is processed.
14447 Evaluate expressions for each incoming frame.
14451 Default value is @samp{init}.
14454 Pad to aspect instead to a resolution.
14458 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
14459 options are expressions containing the following constants:
14464 The input video width and height.
14468 These are the same as @var{in_w} and @var{in_h}.
14472 The output width and height (the size of the padded area), as
14473 specified by the @var{width} and @var{height} expressions.
14477 These are the same as @var{out_w} and @var{out_h}.
14481 The x and y offsets as specified by the @var{x} and @var{y}
14482 expressions, or NAN if not yet specified.
14485 same as @var{iw} / @var{ih}
14488 input sample aspect ratio
14491 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
14495 The horizontal and vertical chroma subsample values. For example for the
14496 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14499 @subsection Examples
14503 Add paddings with the color "violet" to the input video. The output video
14504 size is 640x480, and the top-left corner of the input video is placed at
14507 pad=640:480:0:40:violet
14510 The example above is equivalent to the following command:
14512 pad=width=640:height=480:x=0:y=40:color=violet
14516 Pad the input to get an output with dimensions increased by 3/2,
14517 and put the input video at the center of the padded area:
14519 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
14523 Pad the input to get a squared output with size equal to the maximum
14524 value between the input width and height, and put the input video at
14525 the center of the padded area:
14527 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
14531 Pad the input to get a final w/h ratio of 16:9:
14533 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
14537 In case of anamorphic video, in order to set the output display aspect
14538 correctly, it is necessary to use @var{sar} in the expression,
14539 according to the relation:
14541 (ih * X / ih) * sar = output_dar
14542 X = output_dar / sar
14545 Thus the previous example needs to be modified to:
14547 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
14551 Double the output size and put the input video in the bottom-right
14552 corner of the output padded area:
14554 pad="2*iw:2*ih:ow-iw:oh-ih"
14558 @anchor{palettegen}
14559 @section palettegen
14561 Generate one palette for a whole video stream.
14563 It accepts the following options:
14567 Set the maximum number of colors to quantize in the palette.
14568 Note: the palette will still contain 256 colors; the unused palette entries
14571 @item reserve_transparent
14572 Create a palette of 255 colors maximum and reserve the last one for
14573 transparency. Reserving the transparency color is useful for GIF optimization.
14574 If not set, the maximum of colors in the palette will be 256. You probably want
14575 to disable this option for a standalone image.
14578 @item transparency_color
14579 Set the color that will be used as background for transparency.
14582 Set statistics mode.
14584 It accepts the following values:
14587 Compute full frame histograms.
14589 Compute histograms only for the part that differs from previous frame. This
14590 might be relevant to give more importance to the moving part of your input if
14591 the background is static.
14593 Compute new histogram for each frame.
14596 Default value is @var{full}.
14599 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
14600 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
14601 color quantization of the palette. This information is also visible at
14602 @var{info} logging level.
14604 @subsection Examples
14608 Generate a representative palette of a given video using @command{ffmpeg}:
14610 ffmpeg -i input.mkv -vf palettegen palette.png
14614 @section paletteuse
14616 Use a palette to downsample an input video stream.
14618 The filter takes two inputs: one video stream and a palette. The palette must
14619 be a 256 pixels image.
14621 It accepts the following options:
14625 Select dithering mode. Available algorithms are:
14628 Ordered 8x8 bayer dithering (deterministic)
14630 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
14631 Note: this dithering is sometimes considered "wrong" and is included as a
14633 @item floyd_steinberg
14634 Floyd and Steingberg dithering (error diffusion)
14636 Frankie Sierra dithering v2 (error diffusion)
14638 Frankie Sierra dithering v2 "Lite" (error diffusion)
14641 Default is @var{sierra2_4a}.
14644 When @var{bayer} dithering is selected, this option defines the scale of the
14645 pattern (how much the crosshatch pattern is visible). A low value means more
14646 visible pattern for less banding, and higher value means less visible pattern
14647 at the cost of more banding.
14649 The option must be an integer value in the range [0,5]. Default is @var{2}.
14652 If set, define the zone to process
14656 Only the changing rectangle will be reprocessed. This is similar to GIF
14657 cropping/offsetting compression mechanism. This option can be useful for speed
14658 if only a part of the image is changing, and has use cases such as limiting the
14659 scope of the error diffusal @option{dither} to the rectangle that bounds the
14660 moving scene (it leads to more deterministic output if the scene doesn't change
14661 much, and as a result less moving noise and better GIF compression).
14664 Default is @var{none}.
14667 Take new palette for each output frame.
14669 @item alpha_threshold
14670 Sets the alpha threshold for transparency. Alpha values above this threshold
14671 will be treated as completely opaque, and values below this threshold will be
14672 treated as completely transparent.
14674 The option must be an integer value in the range [0,255]. Default is @var{128}.
14677 @subsection Examples
14681 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
14682 using @command{ffmpeg}:
14684 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
14688 @section perspective
14690 Correct perspective of video not recorded perpendicular to the screen.
14692 A description of the accepted parameters follows.
14703 Set coordinates expression for top left, top right, bottom left and bottom right corners.
14704 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
14705 If the @code{sense} option is set to @code{source}, then the specified points will be sent
14706 to the corners of the destination. If the @code{sense} option is set to @code{destination},
14707 then the corners of the source will be sent to the specified coordinates.
14709 The expressions can use the following variables:
14714 the width and height of video frame.
14718 Output frame count.
14721 @item interpolation
14722 Set interpolation for perspective correction.
14724 It accepts the following values:
14730 Default value is @samp{linear}.
14733 Set interpretation of coordinate options.
14735 It accepts the following values:
14739 Send point in the source specified by the given coordinates to
14740 the corners of the destination.
14742 @item 1, destination
14744 Send the corners of the source to the point in the destination specified
14745 by the given coordinates.
14747 Default value is @samp{source}.
14751 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
14753 It accepts the following values:
14756 only evaluate expressions once during the filter initialization or
14757 when a command is processed
14760 evaluate expressions for each incoming frame
14763 Default value is @samp{init}.
14768 Delay interlaced video by one field time so that the field order changes.
14770 The intended use is to fix PAL movies that have been captured with the
14771 opposite field order to the film-to-video transfer.
14773 A description of the accepted parameters follows.
14779 It accepts the following values:
14782 Capture field order top-first, transfer bottom-first.
14783 Filter will delay the bottom field.
14786 Capture field order bottom-first, transfer top-first.
14787 Filter will delay the top field.
14790 Capture and transfer with the same field order. This mode only exists
14791 for the documentation of the other options to refer to, but if you
14792 actually select it, the filter will faithfully do nothing.
14795 Capture field order determined automatically by field flags, transfer
14797 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
14798 basis using field flags. If no field information is available,
14799 then this works just like @samp{u}.
14802 Capture unknown or varying, transfer opposite.
14803 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
14804 analyzing the images and selecting the alternative that produces best
14805 match between the fields.
14808 Capture top-first, transfer unknown or varying.
14809 Filter selects among @samp{t} and @samp{p} using image analysis.
14812 Capture bottom-first, transfer unknown or varying.
14813 Filter selects among @samp{b} and @samp{p} using image analysis.
14816 Capture determined by field flags, transfer unknown or varying.
14817 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
14818 image analysis. If no field information is available, then this works just
14819 like @samp{U}. This is the default mode.
14822 Both capture and transfer unknown or varying.
14823 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
14827 @section photosensitivity
14828 Reduce various flashes in video, so to help users with epilepsy.
14830 It accepts the following options:
14833 Set how many frames to use when filtering. Default is 30.
14836 Set detection threshold factor. Default is 1.
14840 Set how many pixels to skip when sampling frames. Default is 1.
14841 Allowed range is from 1 to 1024.
14844 Leave frames unchanged. Default is disabled.
14847 @section pixdesctest
14849 Pixel format descriptor test filter, mainly useful for internal
14850 testing. The output video should be equal to the input video.
14854 format=monow, pixdesctest
14857 can be used to test the monowhite pixel format descriptor definition.
14861 Display sample values of color channels. Mainly useful for checking color
14862 and levels. Minimum supported resolution is 640x480.
14864 The filters accept the following options:
14868 Set scope X position, relative offset on X axis.
14871 Set scope Y position, relative offset on Y axis.
14880 Set window opacity. This window also holds statistics about pixel area.
14883 Set window X position, relative offset on X axis.
14886 Set window Y position, relative offset on Y axis.
14891 Enable the specified chain of postprocessing subfilters using libpostproc. This
14892 library should be automatically selected with a GPL build (@code{--enable-gpl}).
14893 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
14894 Each subfilter and some options have a short and a long name that can be used
14895 interchangeably, i.e. dr/dering are the same.
14897 The filters accept the following options:
14901 Set postprocessing subfilters string.
14904 All subfilters share common options to determine their scope:
14908 Honor the quality commands for this subfilter.
14911 Do chrominance filtering, too (default).
14914 Do luminance filtering only (no chrominance).
14917 Do chrominance filtering only (no luminance).
14920 These options can be appended after the subfilter name, separated by a '|'.
14922 Available subfilters are:
14925 @item hb/hdeblock[|difference[|flatness]]
14926 Horizontal deblocking filter
14929 Difference factor where higher values mean more deblocking (default: @code{32}).
14931 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14934 @item vb/vdeblock[|difference[|flatness]]
14935 Vertical deblocking filter
14938 Difference factor where higher values mean more deblocking (default: @code{32}).
14940 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14943 @item ha/hadeblock[|difference[|flatness]]
14944 Accurate horizontal deblocking filter
14947 Difference factor where higher values mean more deblocking (default: @code{32}).
14949 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14952 @item va/vadeblock[|difference[|flatness]]
14953 Accurate vertical deblocking filter
14956 Difference factor where higher values mean more deblocking (default: @code{32}).
14958 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14962 The horizontal and vertical deblocking filters share the difference and
14963 flatness values so you cannot set different horizontal and vertical
14967 @item h1/x1hdeblock
14968 Experimental horizontal deblocking filter
14970 @item v1/x1vdeblock
14971 Experimental vertical deblocking filter
14976 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
14979 larger -> stronger filtering
14981 larger -> stronger filtering
14983 larger -> stronger filtering
14986 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
14989 Stretch luminance to @code{0-255}.
14992 @item lb/linblenddeint
14993 Linear blend deinterlacing filter that deinterlaces the given block by
14994 filtering all lines with a @code{(1 2 1)} filter.
14996 @item li/linipoldeint
14997 Linear interpolating deinterlacing filter that deinterlaces the given block by
14998 linearly interpolating every second line.
15000 @item ci/cubicipoldeint
15001 Cubic interpolating deinterlacing filter deinterlaces the given block by
15002 cubically interpolating every second line.
15004 @item md/mediandeint
15005 Median deinterlacing filter that deinterlaces the given block by applying a
15006 median filter to every second line.
15008 @item fd/ffmpegdeint
15009 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15010 second line with a @code{(-1 4 2 4 -1)} filter.
15013 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15014 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15016 @item fq/forceQuant[|quantizer]
15017 Overrides the quantizer table from the input with the constant quantizer you
15025 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15028 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15031 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15034 @subsection Examples
15038 Apply horizontal and vertical deblocking, deringing and automatic
15039 brightness/contrast:
15045 Apply default filters without brightness/contrast correction:
15051 Apply default filters and temporal denoiser:
15053 pp=default/tmpnoise|1|2|3
15057 Apply deblocking on luminance only, and switch vertical deblocking on or off
15058 automatically depending on available CPU time:
15065 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15066 similar to spp = 6 with 7 point DCT, where only the center sample is
15069 The filter accepts the following options:
15073 Force a constant quantization parameter. It accepts an integer in range
15074 0 to 63. If not set, the filter will use the QP from the video stream
15078 Set thresholding mode. Available modes are:
15082 Set hard thresholding.
15084 Set soft thresholding (better de-ringing effect, but likely blurrier).
15086 Set medium thresholding (good results, default).
15090 @section premultiply
15091 Apply alpha premultiply effect to input video stream using first plane
15092 of second stream as alpha.
15094 Both streams must have same dimensions and same pixel format.
15096 The filter accepts the following option:
15100 Set which planes will be processed, unprocessed planes will be copied.
15101 By default value 0xf, all planes will be processed.
15104 Do not require 2nd input for processing, instead use alpha plane from input stream.
15108 Apply prewitt operator to input video stream.
15110 The filter accepts the following option:
15114 Set which planes will be processed, unprocessed planes will be copied.
15115 By default value 0xf, all planes will be processed.
15118 Set value which will be multiplied with filtered result.
15121 Set value which will be added to filtered result.
15124 @section pseudocolor
15126 Alter frame colors in video with pseudocolors.
15128 This filter accepts the following options:
15132 set pixel first component expression
15135 set pixel second component expression
15138 set pixel third component expression
15141 set pixel fourth component expression, corresponds to the alpha component
15144 set component to use as base for altering colors
15147 Each of them specifies the expression to use for computing the lookup table for
15148 the corresponding pixel component values.
15150 The expressions can contain the following constants and functions:
15155 The input width and height.
15158 The input value for the pixel component.
15160 @item ymin, umin, vmin, amin
15161 The minimum allowed component value.
15163 @item ymax, umax, vmax, amax
15164 The maximum allowed component value.
15167 All expressions default to "val".
15169 @subsection Examples
15173 Change too high luma values to gradient:
15175 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'"
15181 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
15182 Ratio) between two input videos.
15184 This filter takes in input two input videos, the first input is
15185 considered the "main" source and is passed unchanged to the
15186 output. The second input is used as a "reference" video for computing
15189 Both video inputs must have the same resolution and pixel format for
15190 this filter to work correctly. Also it assumes that both inputs
15191 have the same number of frames, which are compared one by one.
15193 The obtained average PSNR is printed through the logging system.
15195 The filter stores the accumulated MSE (mean squared error) of each
15196 frame, and at the end of the processing it is averaged across all frames
15197 equally, and the following formula is applied to obtain the PSNR:
15200 PSNR = 10*log10(MAX^2/MSE)
15203 Where MAX is the average of the maximum values of each component of the
15206 The description of the accepted parameters follows.
15209 @item stats_file, f
15210 If specified the filter will use the named file to save the PSNR of
15211 each individual frame. When filename equals "-" the data is sent to
15214 @item stats_version
15215 Specifies which version of the stats file format to use. Details of
15216 each format are written below.
15217 Default value is 1.
15219 @item stats_add_max
15220 Determines whether the max value is output to the stats log.
15221 Default value is 0.
15222 Requires stats_version >= 2. If this is set and stats_version < 2,
15223 the filter will return an error.
15226 This filter also supports the @ref{framesync} options.
15228 The file printed if @var{stats_file} is selected, contains a sequence of
15229 key/value pairs of the form @var{key}:@var{value} for each compared
15232 If a @var{stats_version} greater than 1 is specified, a header line precedes
15233 the list of per-frame-pair stats, with key value pairs following the frame
15234 format with the following parameters:
15237 @item psnr_log_version
15238 The version of the log file format. Will match @var{stats_version}.
15241 A comma separated list of the per-frame-pair parameters included in
15245 A description of each shown per-frame-pair parameter follows:
15249 sequential number of the input frame, starting from 1
15252 Mean Square Error pixel-by-pixel average difference of the compared
15253 frames, averaged over all the image components.
15255 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
15256 Mean Square Error pixel-by-pixel average difference of the compared
15257 frames for the component specified by the suffix.
15259 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
15260 Peak Signal to Noise ratio of the compared frames for the component
15261 specified by the suffix.
15263 @item max_avg, max_y, max_u, max_v
15264 Maximum allowed value for each channel, and average over all
15268 @subsection Examples
15273 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
15274 [main][ref] psnr="stats_file=stats.log" [out]
15277 On this example the input file being processed is compared with the
15278 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
15279 is stored in @file{stats.log}.
15282 Another example with different containers:
15284 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 -
15291 Pulldown reversal (inverse telecine) filter, capable of handling mixed
15292 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
15295 The pullup filter is designed to take advantage of future context in making
15296 its decisions. This filter is stateless in the sense that it does not lock
15297 onto a pattern to follow, but it instead looks forward to the following
15298 fields in order to identify matches and rebuild progressive frames.
15300 To produce content with an even framerate, insert the fps filter after
15301 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
15302 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
15304 The filter accepts the following options:
15311 These options set the amount of "junk" to ignore at the left, right, top, and
15312 bottom of the image, respectively. Left and right are in units of 8 pixels,
15313 while top and bottom are in units of 2 lines.
15314 The default is 8 pixels on each side.
15317 Set the strict breaks. Setting this option to 1 will reduce the chances of
15318 filter generating an occasional mismatched frame, but it may also cause an
15319 excessive number of frames to be dropped during high motion sequences.
15320 Conversely, setting it to -1 will make filter match fields more easily.
15321 This may help processing of video where there is slight blurring between
15322 the fields, but may also cause there to be interlaced frames in the output.
15323 Default value is @code{0}.
15326 Set the metric plane to use. It accepts the following values:
15332 Use chroma blue plane.
15335 Use chroma red plane.
15338 This option may be set to use chroma plane instead of the default luma plane
15339 for doing filter's computations. This may improve accuracy on very clean
15340 source material, but more likely will decrease accuracy, especially if there
15341 is chroma noise (rainbow effect) or any grayscale video.
15342 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
15343 load and make pullup usable in realtime on slow machines.
15346 For best results (without duplicated frames in the output file) it is
15347 necessary to change the output frame rate. For example, to inverse
15348 telecine NTSC input:
15350 ffmpeg -i input -vf pullup -r 24000/1001 ...
15355 Change video quantization parameters (QP).
15357 The filter accepts the following option:
15361 Set expression for quantization parameter.
15364 The expression is evaluated through the eval API and can contain, among others,
15365 the following constants:
15369 1 if index is not 129, 0 otherwise.
15372 Sequential index starting from -129 to 128.
15375 @subsection Examples
15379 Some equation like:
15387 Flush video frames from internal cache of frames into a random order.
15388 No frame is discarded.
15389 Inspired by @ref{frei0r} nervous filter.
15393 Set size in number of frames of internal cache, in range from @code{2} to
15394 @code{512}. Default is @code{30}.
15397 Set seed for random number generator, must be an integer included between
15398 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
15399 less than @code{0}, the filter will try to use a good random seed on a
15403 @section readeia608
15405 Read closed captioning (EIA-608) information from the top lines of a video frame.
15407 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
15408 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
15409 with EIA-608 data (starting from 0). A description of each metadata value follows:
15412 @item lavfi.readeia608.X.cc
15413 The two bytes stored as EIA-608 data (printed in hexadecimal).
15415 @item lavfi.readeia608.X.line
15416 The number of the line on which the EIA-608 data was identified and read.
15419 This filter accepts the following options:
15423 Set the line to start scanning for EIA-608 data. Default is @code{0}.
15426 Set the line to end scanning for EIA-608 data. Default is @code{29}.
15429 Set the ratio of width reserved for sync code detection.
15430 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
15433 Enable checking the parity bit. In the event of a parity error, the filter will output
15434 @code{0x00} for that character. Default is false.
15437 Lowpass lines prior to further processing. Default is enabled.
15440 @subsection Examples
15444 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
15446 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
15452 Read vertical interval timecode (VITC) information from the top lines of a
15455 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
15456 timecode value, if a valid timecode has been detected. Further metadata key
15457 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
15458 timecode data has been found or not.
15460 This filter accepts the following options:
15464 Set the maximum number of lines to scan for VITC data. If the value is set to
15465 @code{-1} the full video frame is scanned. Default is @code{45}.
15468 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
15469 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
15472 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
15473 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
15476 @subsection Examples
15480 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
15481 draw @code{--:--:--:--} as a placeholder:
15483 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
15489 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
15491 Destination pixel at position (X, Y) will be picked from source (x, y) position
15492 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
15493 value for pixel will be used for destination pixel.
15495 Xmap and Ymap input video streams must be of same dimensions. Output video stream
15496 will have Xmap/Ymap video stream dimensions.
15497 Xmap and Ymap input video streams are 16bit depth, single channel.
15501 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
15502 Default is @code{color}.
15505 Specify the color of the unmapped pixels. For the syntax of this option,
15506 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15507 manual,ffmpeg-utils}. Default color is @code{black}.
15510 @section removegrain
15512 The removegrain filter is a spatial denoiser for progressive video.
15516 Set mode for the first plane.
15519 Set mode for the second plane.
15522 Set mode for the third plane.
15525 Set mode for the fourth plane.
15528 Range of mode is from 0 to 24. Description of each mode follows:
15532 Leave input plane unchanged. Default.
15535 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
15538 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
15541 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
15544 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
15545 This is equivalent to a median filter.
15548 Line-sensitive clipping giving the minimal change.
15551 Line-sensitive clipping, intermediate.
15554 Line-sensitive clipping, intermediate.
15557 Line-sensitive clipping, intermediate.
15560 Line-sensitive clipping on a line where the neighbours pixels are the closest.
15563 Replaces the target pixel with the closest neighbour.
15566 [1 2 1] horizontal and vertical kernel blur.
15572 Bob mode, interpolates top field from the line where the neighbours
15573 pixels are the closest.
15576 Bob mode, interpolates bottom field from the line where the neighbours
15577 pixels are the closest.
15580 Bob mode, interpolates top field. Same as 13 but with a more complicated
15581 interpolation formula.
15584 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
15585 interpolation formula.
15588 Clips the pixel with the minimum and maximum of respectively the maximum and
15589 minimum of each pair of opposite neighbour pixels.
15592 Line-sensitive clipping using opposite neighbours whose greatest distance from
15593 the current pixel is minimal.
15596 Replaces the pixel with the average of its 8 neighbours.
15599 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
15602 Clips pixels using the averages of opposite neighbour.
15605 Same as mode 21 but simpler and faster.
15608 Small edge and halo removal, but reputed useless.
15614 @section removelogo
15616 Suppress a TV station logo, using an image file to determine which
15617 pixels comprise the logo. It works by filling in the pixels that
15618 comprise the logo with neighboring pixels.
15620 The filter accepts the following options:
15624 Set the filter bitmap file, which can be any image format supported by
15625 libavformat. The width and height of the image file must match those of the
15626 video stream being processed.
15629 Pixels in the provided bitmap image with a value of zero are not
15630 considered part of the logo, non-zero pixels are considered part of
15631 the logo. If you use white (255) for the logo and black (0) for the
15632 rest, you will be safe. For making the filter bitmap, it is
15633 recommended to take a screen capture of a black frame with the logo
15634 visible, and then using a threshold filter followed by the erode
15635 filter once or twice.
15637 If needed, little splotches can be fixed manually. Remember that if
15638 logo pixels are not covered, the filter quality will be much
15639 reduced. Marking too many pixels as part of the logo does not hurt as
15640 much, but it will increase the amount of blurring needed to cover over
15641 the image and will destroy more information than necessary, and extra
15642 pixels will slow things down on a large logo.
15644 @section repeatfields
15646 This filter uses the repeat_field flag from the Video ES headers and hard repeats
15647 fields based on its value.
15651 Reverse a video clip.
15653 Warning: This filter requires memory to buffer the entire clip, so trimming
15656 @subsection Examples
15660 Take the first 5 seconds of a clip, and reverse it.
15667 Shift R/G/B/A pixels horizontally and/or vertically.
15669 The filter accepts the following options:
15672 Set amount to shift red horizontally.
15674 Set amount to shift red vertically.
15676 Set amount to shift green horizontally.
15678 Set amount to shift green vertically.
15680 Set amount to shift blue horizontally.
15682 Set amount to shift blue vertically.
15684 Set amount to shift alpha horizontally.
15686 Set amount to shift alpha vertically.
15688 Set edge mode, can be @var{smear}, default, or @var{warp}.
15691 @subsection Commands
15693 This filter supports the all above options as @ref{commands}.
15696 Apply roberts cross operator to input video stream.
15698 The filter accepts the following option:
15702 Set which planes will be processed, unprocessed planes will be copied.
15703 By default value 0xf, all planes will be processed.
15706 Set value which will be multiplied with filtered result.
15709 Set value which will be added to filtered result.
15714 Rotate video by an arbitrary angle expressed in radians.
15716 The filter accepts the following options:
15718 A description of the optional parameters follows.
15721 Set an expression for the angle by which to rotate the input video
15722 clockwise, expressed as a number of radians. A negative value will
15723 result in a counter-clockwise rotation. By default it is set to "0".
15725 This expression is evaluated for each frame.
15728 Set the output width expression, default value is "iw".
15729 This expression is evaluated just once during configuration.
15732 Set the output height expression, default value is "ih".
15733 This expression is evaluated just once during configuration.
15736 Enable bilinear interpolation if set to 1, a value of 0 disables
15737 it. Default value is 1.
15740 Set the color used to fill the output area not covered by the rotated
15741 image. For the general syntax of this option, check the
15742 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
15743 If the special value "none" is selected then no
15744 background is printed (useful for example if the background is never shown).
15746 Default value is "black".
15749 The expressions for the angle and the output size can contain the
15750 following constants and functions:
15754 sequential number of the input frame, starting from 0. It is always NAN
15755 before the first frame is filtered.
15758 time in seconds of the input frame, it is set to 0 when the filter is
15759 configured. It is always NAN before the first frame is filtered.
15763 horizontal and vertical chroma subsample values. For example for the
15764 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15768 the input video width and height
15772 the output width and height, that is the size of the padded area as
15773 specified by the @var{width} and @var{height} expressions
15777 the minimal width/height required for completely containing the input
15778 video rotated by @var{a} radians.
15780 These are only available when computing the @option{out_w} and
15781 @option{out_h} expressions.
15784 @subsection Examples
15788 Rotate the input by PI/6 radians clockwise:
15794 Rotate the input by PI/6 radians counter-clockwise:
15800 Rotate the input by 45 degrees clockwise:
15806 Apply a constant rotation with period T, starting from an angle of PI/3:
15808 rotate=PI/3+2*PI*t/T
15812 Make the input video rotation oscillating with a period of T
15813 seconds and an amplitude of A radians:
15815 rotate=A*sin(2*PI/T*t)
15819 Rotate the video, output size is chosen so that the whole rotating
15820 input video is always completely contained in the output:
15822 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
15826 Rotate the video, reduce the output size so that no background is ever
15829 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
15833 @subsection Commands
15835 The filter supports the following commands:
15839 Set the angle expression.
15840 The command accepts the same syntax of the corresponding option.
15842 If the specified expression is not valid, it is kept at its current
15848 Apply Shape Adaptive Blur.
15850 The filter accepts the following options:
15853 @item luma_radius, lr
15854 Set luma blur filter strength, must be a value in range 0.1-4.0, default
15855 value is 1.0. A greater value will result in a more blurred image, and
15856 in slower processing.
15858 @item luma_pre_filter_radius, lpfr
15859 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
15862 @item luma_strength, ls
15863 Set luma maximum difference between pixels to still be considered, must
15864 be a value in the 0.1-100.0 range, default value is 1.0.
15866 @item chroma_radius, cr
15867 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
15868 greater value will result in a more blurred image, and in slower
15871 @item chroma_pre_filter_radius, cpfr
15872 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
15874 @item chroma_strength, cs
15875 Set chroma maximum difference between pixels to still be considered,
15876 must be a value in the -0.9-100.0 range.
15879 Each chroma option value, if not explicitly specified, is set to the
15880 corresponding luma option value.
15885 Scale (resize) the input video, using the libswscale library.
15887 The scale filter forces the output display aspect ratio to be the same
15888 of the input, by changing the output sample aspect ratio.
15890 If the input image format is different from the format requested by
15891 the next filter, the scale filter will convert the input to the
15894 @subsection Options
15895 The filter accepts the following options, or any of the options
15896 supported by the libswscale scaler.
15898 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
15899 the complete list of scaler options.
15904 Set the output video dimension expression. Default value is the input
15907 If the @var{width} or @var{w} value is 0, the input width is used for
15908 the output. If the @var{height} or @var{h} value is 0, the input height
15909 is used for the output.
15911 If one and only one of the values is -n with n >= 1, the scale filter
15912 will use a value that maintains the aspect ratio of the input image,
15913 calculated from the other specified dimension. After that it will,
15914 however, make sure that the calculated dimension is divisible by n and
15915 adjust the value if necessary.
15917 If both values are -n with n >= 1, the behavior will be identical to
15918 both values being set to 0 as previously detailed.
15920 See below for the list of accepted constants for use in the dimension
15924 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
15928 Only evaluate expressions once during the filter initialization or when a command is processed.
15931 Evaluate expressions for each incoming frame.
15935 Default value is @samp{init}.
15939 Set the interlacing mode. It accepts the following values:
15943 Force interlaced aware scaling.
15946 Do not apply interlaced scaling.
15949 Select interlaced aware scaling depending on whether the source frames
15950 are flagged as interlaced or not.
15953 Default value is @samp{0}.
15956 Set libswscale scaling flags. See
15957 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15958 complete list of values. If not explicitly specified the filter applies
15962 @item param0, param1
15963 Set libswscale input parameters for scaling algorithms that need them. See
15964 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15965 complete documentation. If not explicitly specified the filter applies
15971 Set the video size. For the syntax of this option, check the
15972 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15974 @item in_color_matrix
15975 @item out_color_matrix
15976 Set in/output YCbCr color space type.
15978 This allows the autodetected value to be overridden as well as allows forcing
15979 a specific value used for the output and encoder.
15981 If not specified, the color space type depends on the pixel format.
15987 Choose automatically.
15990 Format conforming to International Telecommunication Union (ITU)
15991 Recommendation BT.709.
15994 Set color space conforming to the United States Federal Communications
15995 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16000 Set color space conforming to:
16004 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16007 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16010 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16015 Set color space conforming to SMPTE ST 240:1999.
16018 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16023 Set in/output YCbCr sample range.
16025 This allows the autodetected value to be overridden as well as allows forcing
16026 a specific value used for the output and encoder. If not specified, the
16027 range depends on the pixel format. Possible values:
16031 Choose automatically.
16034 Set full range (0-255 in case of 8-bit luma).
16036 @item mpeg/limited/tv
16037 Set "MPEG" range (16-235 in case of 8-bit luma).
16040 @item force_original_aspect_ratio
16041 Enable decreasing or increasing output video width or height if necessary to
16042 keep the original aspect ratio. Possible values:
16046 Scale the video as specified and disable this feature.
16049 The output video dimensions will automatically be decreased if needed.
16052 The output video dimensions will automatically be increased if needed.
16056 One useful instance of this option is that when you know a specific device's
16057 maximum allowed resolution, you can use this to limit the output video to
16058 that, while retaining the aspect ratio. For example, device A allows
16059 1280x720 playback, and your video is 1920x800. Using this option (set it to
16060 decrease) and specifying 1280x720 to the command line makes the output
16063 Please note that this is a different thing than specifying -1 for @option{w}
16064 or @option{h}, you still need to specify the output resolution for this option
16067 @item force_divisible_by
16068 Ensures that both the output dimensions, width and height, are divisible by the
16069 given integer when used together with @option{force_original_aspect_ratio}. This
16070 works similar to using @code{-n} in the @option{w} and @option{h} options.
16072 This option respects the value set for @option{force_original_aspect_ratio},
16073 increasing or decreasing the resolution accordingly. The video's aspect ratio
16074 may be slightly modified.
16076 This option can be handy if you need to have a video fit within or exceed
16077 a defined resolution using @option{force_original_aspect_ratio} but also have
16078 encoder restrictions on width or height divisibility.
16082 The values of the @option{w} and @option{h} options are expressions
16083 containing the following constants:
16088 The input width and height
16092 These are the same as @var{in_w} and @var{in_h}.
16096 The output (scaled) width and height
16100 These are the same as @var{out_w} and @var{out_h}
16103 The same as @var{iw} / @var{ih}
16106 input sample aspect ratio
16109 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16113 horizontal and vertical input chroma subsample values. For example for the
16114 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16118 horizontal and vertical output chroma subsample values. For example for the
16119 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16122 The (sequential) number of the input frame, starting from 0.
16123 Only available with @code{eval=frame}.
16126 The presentation timestamp of the input frame, expressed as a number of
16127 seconds. Only available with @code{eval=frame}.
16130 The position (byte offset) of the frame in the input stream, or NaN if
16131 this information is unavailable and/or meaningless (for example in case of synthetic video).
16132 Only available with @code{eval=frame}.
16135 @subsection Examples
16139 Scale the input video to a size of 200x100
16144 This is equivalent to:
16155 Specify a size abbreviation for the output size:
16160 which can also be written as:
16166 Scale the input to 2x:
16168 scale=w=2*iw:h=2*ih
16172 The above is the same as:
16174 scale=2*in_w:2*in_h
16178 Scale the input to 2x with forced interlaced scaling:
16180 scale=2*iw:2*ih:interl=1
16184 Scale the input to half size:
16186 scale=w=iw/2:h=ih/2
16190 Increase the width, and set the height to the same size:
16196 Seek Greek harmony:
16203 Increase the height, and set the width to 3/2 of the height:
16205 scale=w=3/2*oh:h=3/5*ih
16209 Increase the size, making the size a multiple of the chroma
16212 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
16216 Increase the width to a maximum of 500 pixels,
16217 keeping the same aspect ratio as the input:
16219 scale=w='min(500\, iw*3/2):h=-1'
16223 Make pixels square by combining scale and setsar:
16225 scale='trunc(ih*dar):ih',setsar=1/1
16229 Make pixels square by combining scale and setsar,
16230 making sure the resulting resolution is even (required by some codecs):
16232 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
16236 @subsection Commands
16238 This filter supports the following commands:
16242 Set the output video dimension expression.
16243 The command accepts the same syntax of the corresponding option.
16245 If the specified expression is not valid, it is kept at its current
16251 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
16252 format conversion on CUDA video frames. Setting the output width and height
16253 works in the same way as for the @var{scale} filter.
16255 The following additional options are accepted:
16258 The pixel format of the output CUDA frames. If set to the string "same" (the
16259 default), the input format will be kept. Note that automatic format negotiation
16260 and conversion is not yet supported for hardware frames
16263 The interpolation algorithm used for resizing. One of the following:
16270 @item cubic2p_bspline
16271 2-parameter cubic (B=1, C=0)
16273 @item cubic2p_catmullrom
16274 2-parameter cubic (B=0, C=1/2)
16276 @item cubic2p_b05c03
16277 2-parameter cubic (B=1/2, C=3/10)
16285 @item force_original_aspect_ratio
16286 Enable decreasing or increasing output video width or height if necessary to
16287 keep the original aspect ratio. Possible values:
16291 Scale the video as specified and disable this feature.
16294 The output video dimensions will automatically be decreased if needed.
16297 The output video dimensions will automatically be increased if needed.
16301 One useful instance of this option is that when you know a specific device's
16302 maximum allowed resolution, you can use this to limit the output video to
16303 that, while retaining the aspect ratio. For example, device A allows
16304 1280x720 playback, and your video is 1920x800. Using this option (set it to
16305 decrease) and specifying 1280x720 to the command line makes the output
16308 Please note that this is a different thing than specifying -1 for @option{w}
16309 or @option{h}, you still need to specify the output resolution for this option
16312 @item force_divisible_by
16313 Ensures that both the output dimensions, width and height, are divisible by the
16314 given integer when used together with @option{force_original_aspect_ratio}. This
16315 works similar to using @code{-n} in the @option{w} and @option{h} options.
16317 This option respects the value set for @option{force_original_aspect_ratio},
16318 increasing or decreasing the resolution accordingly. The video's aspect ratio
16319 may be slightly modified.
16321 This option can be handy if you need to have a video fit within or exceed
16322 a defined resolution using @option{force_original_aspect_ratio} but also have
16323 encoder restrictions on width or height divisibility.
16329 Scale (resize) the input video, based on a reference video.
16331 See the scale filter for available options, scale2ref supports the same but
16332 uses the reference video instead of the main input as basis. scale2ref also
16333 supports the following additional constants for the @option{w} and
16334 @option{h} options:
16339 The main input video's width and height
16342 The same as @var{main_w} / @var{main_h}
16345 The main input video's sample aspect ratio
16347 @item main_dar, mdar
16348 The main input video's display aspect ratio. Calculated from
16349 @code{(main_w / main_h) * main_sar}.
16353 The main input video's horizontal and vertical chroma subsample values.
16354 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
16358 The (sequential) number of the main input frame, starting from 0.
16359 Only available with @code{eval=frame}.
16362 The presentation timestamp of the main input frame, expressed as a number of
16363 seconds. Only available with @code{eval=frame}.
16366 The position (byte offset) of the frame in the main input stream, or NaN if
16367 this information is unavailable and/or meaningless (for example in case of synthetic video).
16368 Only available with @code{eval=frame}.
16371 @subsection Examples
16375 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
16377 'scale2ref[b][a];[a][b]overlay'
16381 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
16383 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
16387 @subsection Commands
16389 This filter supports the following commands:
16393 Set the output video dimension expression.
16394 The command accepts the same syntax of the corresponding option.
16396 If the specified expression is not valid, it is kept at its current
16401 Scroll input video horizontally and/or vertically by constant speed.
16403 The filter accepts the following options:
16405 @item horizontal, h
16406 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
16407 Negative values changes scrolling direction.
16410 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
16411 Negative values changes scrolling direction.
16414 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
16417 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
16420 @subsection Commands
16422 This filter supports the following @ref{commands}:
16424 @item horizontal, h
16425 Set the horizontal scrolling speed.
16427 Set the vertical scrolling speed.
16430 @anchor{selectivecolor}
16431 @section selectivecolor
16433 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
16434 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
16435 by the "purity" of the color (that is, how saturated it already is).
16437 This filter is similar to the Adobe Photoshop Selective Color tool.
16439 The filter accepts the following options:
16442 @item correction_method
16443 Select color correction method.
16445 Available values are:
16448 Specified adjustments are applied "as-is" (added/subtracted to original pixel
16451 Specified adjustments are relative to the original component value.
16453 Default is @code{absolute}.
16455 Adjustments for red pixels (pixels where the red component is the maximum)
16457 Adjustments for yellow pixels (pixels where the blue component is the minimum)
16459 Adjustments for green pixels (pixels where the green component is the maximum)
16461 Adjustments for cyan pixels (pixels where the red component is the minimum)
16463 Adjustments for blue pixels (pixels where the blue component is the maximum)
16465 Adjustments for magenta pixels (pixels where the green component is the minimum)
16467 Adjustments for white pixels (pixels where all components are greater than 128)
16469 Adjustments for all pixels except pure black and pure white
16471 Adjustments for black pixels (pixels where all components are lesser than 128)
16473 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
16476 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
16477 4 space separated floating point adjustment values in the [-1,1] range,
16478 respectively to adjust the amount of cyan, magenta, yellow and black for the
16479 pixels of its range.
16481 @subsection Examples
16485 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
16486 increase magenta by 27% in blue areas:
16488 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
16492 Use a Photoshop selective color preset:
16494 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
16498 @anchor{separatefields}
16499 @section separatefields
16501 The @code{separatefields} takes a frame-based video input and splits
16502 each frame into its components fields, producing a new half height clip
16503 with twice the frame rate and twice the frame count.
16505 This filter use field-dominance information in frame to decide which
16506 of each pair of fields to place first in the output.
16507 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
16509 @section setdar, setsar
16511 The @code{setdar} filter sets the Display Aspect Ratio for the filter
16514 This is done by changing the specified Sample (aka Pixel) Aspect
16515 Ratio, according to the following equation:
16517 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
16520 Keep in mind that the @code{setdar} filter does not modify the pixel
16521 dimensions of the video frame. Also, the display aspect ratio set by
16522 this filter may be changed by later filters in the filterchain,
16523 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
16526 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
16527 the filter output video.
16529 Note that as a consequence of the application of this filter, the
16530 output display aspect ratio will change according to the equation
16533 Keep in mind that the sample aspect ratio set by the @code{setsar}
16534 filter may be changed by later filters in the filterchain, e.g. if
16535 another "setsar" or a "setdar" filter is applied.
16537 It accepts the following parameters:
16540 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
16541 Set the aspect ratio used by the filter.
16543 The parameter can be a floating point number string, an expression, or
16544 a string of the form @var{num}:@var{den}, where @var{num} and
16545 @var{den} are the numerator and denominator of the aspect ratio. If
16546 the parameter is not specified, it is assumed the value "0".
16547 In case the form "@var{num}:@var{den}" is used, the @code{:} character
16551 Set the maximum integer value to use for expressing numerator and
16552 denominator when reducing the expressed aspect ratio to a rational.
16553 Default value is @code{100}.
16557 The parameter @var{sar} is an expression containing
16558 the following constants:
16562 These are approximated values for the mathematical constants e
16563 (Euler's number), pi (Greek pi), and phi (the golden ratio).
16566 The input width and height.
16569 These are the same as @var{w} / @var{h}.
16572 The input sample aspect ratio.
16575 The input display aspect ratio. It is the same as
16576 (@var{w} / @var{h}) * @var{sar}.
16579 Horizontal and vertical chroma subsample values. For example, for the
16580 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16583 @subsection Examples
16588 To change the display aspect ratio to 16:9, specify one of the following:
16595 To change the sample aspect ratio to 10:11, specify:
16601 To set a display aspect ratio of 16:9, and specify a maximum integer value of
16602 1000 in the aspect ratio reduction, use the command:
16604 setdar=ratio=16/9:max=1000
16612 Force field for the output video frame.
16614 The @code{setfield} filter marks the interlace type field for the
16615 output frames. It does not change the input frame, but only sets the
16616 corresponding property, which affects how the frame is treated by
16617 following filters (e.g. @code{fieldorder} or @code{yadif}).
16619 The filter accepts the following options:
16624 Available values are:
16628 Keep the same field property.
16631 Mark the frame as bottom-field-first.
16634 Mark the frame as top-field-first.
16637 Mark the frame as progressive.
16644 Force frame parameter for the output video frame.
16646 The @code{setparams} filter marks interlace and color range for the
16647 output frames. It does not change the input frame, but only sets the
16648 corresponding property, which affects how the frame is treated by
16653 Available values are:
16657 Keep the same field property (default).
16660 Mark the frame as bottom-field-first.
16663 Mark the frame as top-field-first.
16666 Mark the frame as progressive.
16670 Available values are:
16674 Keep the same color range property (default).
16676 @item unspecified, unknown
16677 Mark the frame as unspecified color range.
16679 @item limited, tv, mpeg
16680 Mark the frame as limited range.
16682 @item full, pc, jpeg
16683 Mark the frame as full range.
16686 @item color_primaries
16687 Set the color primaries.
16688 Available values are:
16692 Keep the same color primaries property (default).
16709 Set the color transfer.
16710 Available values are:
16714 Keep the same color trc property (default).
16736 Set the colorspace.
16737 Available values are:
16741 Keep the same colorspace property (default).
16754 @item chroma-derived-nc
16755 @item chroma-derived-c
16762 Show a line containing various information for each input video frame.
16763 The input video is not modified.
16765 This filter supports the following options:
16769 Calculate checksums of each plane. By default enabled.
16772 The shown line contains a sequence of key/value pairs of the form
16773 @var{key}:@var{value}.
16775 The following values are shown in the output:
16779 The (sequential) number of the input frame, starting from 0.
16782 The Presentation TimeStamp of the input frame, expressed as a number of
16783 time base units. The time base unit depends on the filter input pad.
16786 The Presentation TimeStamp of the input frame, expressed as a number of
16790 The position of the frame in the input stream, or -1 if this information is
16791 unavailable and/or meaningless (for example in case of synthetic video).
16794 The pixel format name.
16797 The sample aspect ratio of the input frame, expressed in the form
16798 @var{num}/@var{den}.
16801 The size of the input frame. For the syntax of this option, check the
16802 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16805 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
16806 for bottom field first).
16809 This is 1 if the frame is a key frame, 0 otherwise.
16812 The picture type of the input frame ("I" for an I-frame, "P" for a
16813 P-frame, "B" for a B-frame, or "?" for an unknown type).
16814 Also refer to the documentation of the @code{AVPictureType} enum and of
16815 the @code{av_get_picture_type_char} function defined in
16816 @file{libavutil/avutil.h}.
16819 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
16821 @item plane_checksum
16822 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
16823 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
16826 The mean value of pixels in each plane of the input frame, expressed in the form
16827 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
16830 The standard deviation of pixel values in each plane of the input frame, expressed
16831 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
16835 @section showpalette
16837 Displays the 256 colors palette of each frame. This filter is only relevant for
16838 @var{pal8} pixel format frames.
16840 It accepts the following option:
16844 Set the size of the box used to represent one palette color entry. Default is
16845 @code{30} (for a @code{30x30} pixel box).
16848 @section shuffleframes
16850 Reorder and/or duplicate and/or drop video frames.
16852 It accepts the following parameters:
16856 Set the destination indexes of input frames.
16857 This is space or '|' separated list of indexes that maps input frames to output
16858 frames. Number of indexes also sets maximal value that each index may have.
16859 '-1' index have special meaning and that is to drop frame.
16862 The first frame has the index 0. The default is to keep the input unchanged.
16864 @subsection Examples
16868 Swap second and third frame of every three frames of the input:
16870 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
16874 Swap 10th and 1st frame of every ten frames of the input:
16876 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
16880 @section shuffleplanes
16882 Reorder and/or duplicate video planes.
16884 It accepts the following parameters:
16889 The index of the input plane to be used as the first output plane.
16892 The index of the input plane to be used as the second output plane.
16895 The index of the input plane to be used as the third output plane.
16898 The index of the input plane to be used as the fourth output plane.
16902 The first plane has the index 0. The default is to keep the input unchanged.
16904 @subsection Examples
16908 Swap the second and third planes of the input:
16910 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
16914 @anchor{signalstats}
16915 @section signalstats
16916 Evaluate various visual metrics that assist in determining issues associated
16917 with the digitization of analog video media.
16919 By default the filter will log these metadata values:
16923 Display the minimal Y value contained within the input frame. Expressed in
16927 Display the Y value at the 10% percentile within the input frame. Expressed in
16931 Display the average Y value within the input frame. Expressed in range of
16935 Display the Y value at the 90% percentile within the input frame. Expressed in
16939 Display the maximum Y value contained within the input frame. Expressed in
16943 Display the minimal U value contained within the input frame. Expressed in
16947 Display the U value at the 10% percentile within the input frame. Expressed in
16951 Display the average U value within the input frame. Expressed in range of
16955 Display the U value at the 90% percentile within the input frame. Expressed in
16959 Display the maximum U value contained within the input frame. Expressed in
16963 Display the minimal V value contained within the input frame. Expressed in
16967 Display the V value at the 10% percentile within the input frame. Expressed in
16971 Display the average V value within the input frame. Expressed in range of
16975 Display the V value at the 90% percentile within the input frame. Expressed in
16979 Display the maximum V value contained within the input frame. Expressed in
16983 Display the minimal saturation value contained within the input frame.
16984 Expressed in range of [0-~181.02].
16987 Display the saturation value at the 10% percentile within the input frame.
16988 Expressed in range of [0-~181.02].
16991 Display the average saturation value within the input frame. Expressed in range
16995 Display the saturation value at the 90% percentile within the input frame.
16996 Expressed in range of [0-~181.02].
16999 Display the maximum saturation value contained within the input frame.
17000 Expressed in range of [0-~181.02].
17003 Display the median value for hue within the input frame. Expressed in range of
17007 Display the average value for hue within the input frame. Expressed in range of
17011 Display the average of sample value difference between all values of the Y
17012 plane in the current frame and corresponding values of the previous input frame.
17013 Expressed in range of [0-255].
17016 Display the average of sample value difference between all values of the U
17017 plane in the current frame and corresponding values of the previous input frame.
17018 Expressed in range of [0-255].
17021 Display the average of sample value difference between all values of the V
17022 plane in the current frame and corresponding values of the previous input frame.
17023 Expressed in range of [0-255].
17026 Display bit depth of Y plane in current frame.
17027 Expressed in range of [0-16].
17030 Display bit depth of U plane in current frame.
17031 Expressed in range of [0-16].
17034 Display bit depth of V plane in current frame.
17035 Expressed in range of [0-16].
17038 The filter accepts the following options:
17044 @option{stat} specify an additional form of image analysis.
17045 @option{out} output video with the specified type of pixel highlighted.
17047 Both options accept the following values:
17051 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17052 unlike the neighboring pixels of the same field. Examples of temporal outliers
17053 include the results of video dropouts, head clogs, or tape tracking issues.
17056 Identify @var{vertical line repetition}. Vertical line repetition includes
17057 similar rows of pixels within a frame. In born-digital video vertical line
17058 repetition is common, but this pattern is uncommon in video digitized from an
17059 analog source. When it occurs in video that results from the digitization of an
17060 analog source it can indicate concealment from a dropout compensator.
17063 Identify pixels that fall outside of legal broadcast range.
17067 Set the highlight color for the @option{out} option. The default color is
17071 @subsection Examples
17075 Output data of various video metrics:
17077 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17081 Output specific data about the minimum and maximum values of the Y plane per frame:
17083 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17087 Playback video while highlighting pixels that are outside of broadcast range in red.
17089 ffplay example.mov -vf signalstats="out=brng:color=red"
17093 Playback video with signalstats metadata drawn over the frame.
17095 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17098 The contents of signalstat_drawtext.txt used in the command are:
17101 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17102 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17103 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17104 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17112 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17113 input. In this case the matching between the inputs can be calculated additionally.
17114 The filter always passes through the first input. The signature of each stream can
17115 be written into a file.
17117 It accepts the following options:
17121 Enable or disable the matching process.
17123 Available values are:
17127 Disable the calculation of a matching (default).
17129 Calculate the matching for the whole video and output whether the whole video
17130 matches or only parts.
17132 Calculate only until a matching is found or the video ends. Should be faster in
17137 Set the number of inputs. The option value must be a non negative integer.
17138 Default value is 1.
17141 Set the path to which the output is written. If there is more than one input,
17142 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
17143 integer), that will be replaced with the input number. If no filename is
17144 specified, no output will be written. This is the default.
17147 Choose the output format.
17149 Available values are:
17153 Use the specified binary representation (default).
17155 Use the specified xml representation.
17159 Set threshold to detect one word as similar. The option value must be an integer
17160 greater than zero. The default value is 9000.
17163 Set threshold to detect all words as similar. The option value must be an integer
17164 greater than zero. The default value is 60000.
17167 Set threshold to detect frames as similar. The option value must be an integer
17168 greater than zero. The default value is 116.
17171 Set the minimum length of a sequence in frames to recognize it as matching
17172 sequence. The option value must be a non negative integer value.
17173 The default value is 0.
17176 Set the minimum relation, that matching frames to all frames must have.
17177 The option value must be a double value between 0 and 1. The default value is 0.5.
17180 @subsection Examples
17184 To calculate the signature of an input video and store it in signature.bin:
17186 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
17190 To detect whether two videos match and store the signatures in XML format in
17191 signature0.xml and signature1.xml:
17193 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 -
17201 Blur the input video without impacting the outlines.
17203 It accepts the following options:
17206 @item luma_radius, lr
17207 Set the luma radius. The option value must be a float number in
17208 the range [0.1,5.0] that specifies the variance of the gaussian filter
17209 used to blur the image (slower if larger). Default value is 1.0.
17211 @item luma_strength, ls
17212 Set the luma strength. The option value must be a float number
17213 in the range [-1.0,1.0] that configures the blurring. A value included
17214 in [0.0,1.0] will blur the image whereas a value included in
17215 [-1.0,0.0] will sharpen the image. Default value is 1.0.
17217 @item luma_threshold, lt
17218 Set the luma threshold used as a coefficient to determine
17219 whether a pixel should be blurred or not. The option value must be an
17220 integer in the range [-30,30]. A value of 0 will filter all the image,
17221 a value included in [0,30] will filter flat areas and a value included
17222 in [-30,0] will filter edges. Default value is 0.
17224 @item chroma_radius, cr
17225 Set the chroma radius. The option value must be a float number in
17226 the range [0.1,5.0] that specifies the variance of the gaussian filter
17227 used to blur the image (slower if larger). Default value is @option{luma_radius}.
17229 @item chroma_strength, cs
17230 Set the chroma strength. The option value must be a float number
17231 in the range [-1.0,1.0] that configures the blurring. A value included
17232 in [0.0,1.0] will blur the image whereas a value included in
17233 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
17235 @item chroma_threshold, ct
17236 Set the chroma threshold used as a coefficient to determine
17237 whether a pixel should be blurred or not. The option value must be an
17238 integer in the range [-30,30]. A value of 0 will filter all the image,
17239 a value included in [0,30] will filter flat areas and a value included
17240 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
17243 If a chroma option is not explicitly set, the corresponding luma value
17247 Apply sobel operator to input video stream.
17249 The filter accepts the following option:
17253 Set which planes will be processed, unprocessed planes will be copied.
17254 By default value 0xf, all planes will be processed.
17257 Set value which will be multiplied with filtered result.
17260 Set value which will be added to filtered result.
17266 Apply a simple postprocessing filter that compresses and decompresses the image
17267 at several (or - in the case of @option{quality} level @code{6} - all) shifts
17268 and average the results.
17270 The filter accepts the following options:
17274 Set quality. This option defines the number of levels for averaging. It accepts
17275 an integer in the range 0-6. If set to @code{0}, the filter will have no
17276 effect. A value of @code{6} means the higher quality. For each increment of
17277 that value the speed drops by a factor of approximately 2. Default value is
17281 Force a constant quantization parameter. If not set, the filter will use the QP
17282 from the video stream (if available).
17285 Set thresholding mode. Available modes are:
17289 Set hard thresholding (default).
17291 Set soft thresholding (better de-ringing effect, but likely blurrier).
17294 @item use_bframe_qp
17295 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
17296 option may cause flicker since the B-Frames have often larger QP. Default is
17297 @code{0} (not enabled).
17300 @subsection Commands
17302 This filter supports the following commands:
17304 @item quality, level
17305 Set quality level. The value @code{max} can be used to set the maximum level,
17306 currently @code{6}.
17311 Scale the input by applying one of the super-resolution methods based on
17312 convolutional neural networks. Supported models:
17316 Super-Resolution Convolutional Neural Network model (SRCNN).
17317 See @url{https://arxiv.org/abs/1501.00092}.
17320 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
17321 See @url{https://arxiv.org/abs/1609.05158}.
17324 Training scripts as well as scripts for model file (.pb) saving can be found at
17325 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
17326 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
17328 Native model files (.model) can be generated from TensorFlow model
17329 files (.pb) by using tools/python/convert.py
17331 The filter accepts the following options:
17335 Specify which DNN backend to use for model loading and execution. This option accepts
17336 the following values:
17340 Native implementation of DNN loading and execution.
17343 TensorFlow backend. To enable this backend you
17344 need to install the TensorFlow for C library (see
17345 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
17346 @code{--enable-libtensorflow}
17349 Default value is @samp{native}.
17352 Set path to model file specifying network architecture and its parameters.
17353 Note that different backends use different file formats. TensorFlow backend
17354 can load files for both formats, while native backend can load files for only
17358 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
17359 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
17360 input upscaled using bicubic upscaling with proper scale factor.
17365 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
17367 This filter takes in input two input videos, the first input is
17368 considered the "main" source and is passed unchanged to the
17369 output. The second input is used as a "reference" video for computing
17372 Both video inputs must have the same resolution and pixel format for
17373 this filter to work correctly. Also it assumes that both inputs
17374 have the same number of frames, which are compared one by one.
17376 The filter stores the calculated SSIM of each frame.
17378 The description of the accepted parameters follows.
17381 @item stats_file, f
17382 If specified the filter will use the named file to save the SSIM of
17383 each individual frame. When filename equals "-" the data is sent to
17387 The file printed if @var{stats_file} is selected, contains a sequence of
17388 key/value pairs of the form @var{key}:@var{value} for each compared
17391 A description of each shown parameter follows:
17395 sequential number of the input frame, starting from 1
17397 @item Y, U, V, R, G, B
17398 SSIM of the compared frames for the component specified by the suffix.
17401 SSIM of the compared frames for the whole frame.
17404 Same as above but in dB representation.
17407 This filter also supports the @ref{framesync} options.
17409 @subsection Examples
17414 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
17415 [main][ref] ssim="stats_file=stats.log" [out]
17418 On this example the input file being processed is compared with the
17419 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
17420 is stored in @file{stats.log}.
17423 Another example with both psnr and ssim at same time:
17425 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
17429 Another example with different containers:
17431 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 -
17437 Convert between different stereoscopic image formats.
17439 The filters accept the following options:
17443 Set stereoscopic image format of input.
17445 Available values for input image formats are:
17448 side by side parallel (left eye left, right eye right)
17451 side by side crosseye (right eye left, left eye right)
17454 side by side parallel with half width resolution
17455 (left eye left, right eye right)
17458 side by side crosseye with half width resolution
17459 (right eye left, left eye right)
17463 above-below (left eye above, right eye below)
17467 above-below (right eye above, left eye below)
17471 above-below with half height resolution
17472 (left eye above, right eye below)
17476 above-below with half height resolution
17477 (right eye above, left eye below)
17480 alternating frames (left eye first, right eye second)
17483 alternating frames (right eye first, left eye second)
17486 interleaved rows (left eye has top row, right eye starts on next row)
17489 interleaved rows (right eye has top row, left eye starts on next row)
17492 interleaved columns, left eye first
17495 interleaved columns, right eye first
17497 Default value is @samp{sbsl}.
17501 Set stereoscopic image format of output.
17505 side by side parallel (left eye left, right eye right)
17508 side by side crosseye (right eye left, left eye right)
17511 side by side parallel with half width resolution
17512 (left eye left, right eye right)
17515 side by side crosseye with half width resolution
17516 (right eye left, left eye right)
17520 above-below (left eye above, right eye below)
17524 above-below (right eye above, left eye below)
17528 above-below with half height resolution
17529 (left eye above, right eye below)
17533 above-below with half height resolution
17534 (right eye above, left eye below)
17537 alternating frames (left eye first, right eye second)
17540 alternating frames (right eye first, left eye second)
17543 interleaved rows (left eye has top row, right eye starts on next row)
17546 interleaved rows (right eye has top row, left eye starts on next row)
17549 anaglyph red/blue gray
17550 (red filter on left eye, blue filter on right eye)
17553 anaglyph red/green gray
17554 (red filter on left eye, green filter on right eye)
17557 anaglyph red/cyan gray
17558 (red filter on left eye, cyan filter on right eye)
17561 anaglyph red/cyan half colored
17562 (red filter on left eye, cyan filter on right eye)
17565 anaglyph red/cyan color
17566 (red filter on left eye, cyan filter on right eye)
17569 anaglyph red/cyan color optimized with the least squares projection of dubois
17570 (red filter on left eye, cyan filter on right eye)
17573 anaglyph green/magenta gray
17574 (green filter on left eye, magenta filter on right eye)
17577 anaglyph green/magenta half colored
17578 (green filter on left eye, magenta filter on right eye)
17581 anaglyph green/magenta colored
17582 (green filter on left eye, magenta filter on right eye)
17585 anaglyph green/magenta color optimized with the least squares projection of dubois
17586 (green filter on left eye, magenta filter on right eye)
17589 anaglyph yellow/blue gray
17590 (yellow filter on left eye, blue filter on right eye)
17593 anaglyph yellow/blue half colored
17594 (yellow filter on left eye, blue filter on right eye)
17597 anaglyph yellow/blue colored
17598 (yellow filter on left eye, blue filter on right eye)
17601 anaglyph yellow/blue color optimized with the least squares projection of dubois
17602 (yellow filter on left eye, blue filter on right eye)
17605 mono output (left eye only)
17608 mono output (right eye only)
17611 checkerboard, left eye first
17614 checkerboard, right eye first
17617 interleaved columns, left eye first
17620 interleaved columns, right eye first
17626 Default value is @samp{arcd}.
17629 @subsection Examples
17633 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
17639 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
17645 @section streamselect, astreamselect
17646 Select video or audio streams.
17648 The filter accepts the following options:
17652 Set number of inputs. Default is 2.
17655 Set input indexes to remap to outputs.
17658 @subsection Commands
17660 The @code{streamselect} and @code{astreamselect} filter supports the following
17665 Set input indexes to remap to outputs.
17668 @subsection Examples
17672 Select first 5 seconds 1st stream and rest of time 2nd stream:
17674 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
17678 Same as above, but for audio:
17680 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
17687 Draw subtitles on top of input video using the libass library.
17689 To enable compilation of this filter you need to configure FFmpeg with
17690 @code{--enable-libass}. This filter also requires a build with libavcodec and
17691 libavformat to convert the passed subtitles file to ASS (Advanced Substation
17692 Alpha) subtitles format.
17694 The filter accepts the following options:
17698 Set the filename of the subtitle file to read. It must be specified.
17700 @item original_size
17701 Specify the size of the original video, the video for which the ASS file
17702 was composed. For the syntax of this option, check the
17703 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17704 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
17705 correctly scale the fonts if the aspect ratio has been changed.
17708 Set a directory path containing fonts that can be used by the filter.
17709 These fonts will be used in addition to whatever the font provider uses.
17712 Process alpha channel, by default alpha channel is untouched.
17715 Set subtitles input character encoding. @code{subtitles} filter only. Only
17716 useful if not UTF-8.
17718 @item stream_index, si
17719 Set subtitles stream index. @code{subtitles} filter only.
17722 Override default style or script info parameters of the subtitles. It accepts a
17723 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
17726 If the first key is not specified, it is assumed that the first value
17727 specifies the @option{filename}.
17729 For example, to render the file @file{sub.srt} on top of the input
17730 video, use the command:
17735 which is equivalent to:
17737 subtitles=filename=sub.srt
17740 To render the default subtitles stream from file @file{video.mkv}, use:
17742 subtitles=video.mkv
17745 To render the second subtitles stream from that file, use:
17747 subtitles=video.mkv:si=1
17750 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
17751 @code{DejaVu Serif}, use:
17753 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
17756 @section super2xsai
17758 Scale the input by 2x and smooth using the Super2xSaI (Scale and
17759 Interpolate) pixel art scaling algorithm.
17761 Useful for enlarging pixel art images without reducing sharpness.
17765 Swap two rectangular objects in video.
17767 This filter accepts the following options:
17777 Set 1st rect x coordinate.
17780 Set 1st rect y coordinate.
17783 Set 2nd rect x coordinate.
17786 Set 2nd rect y coordinate.
17788 All expressions are evaluated once for each frame.
17791 The all options are expressions containing the following constants:
17796 The input width and height.
17799 same as @var{w} / @var{h}
17802 input sample aspect ratio
17805 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
17808 The number of the input frame, starting from 0.
17811 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
17814 the position in the file of the input frame, NAN if unknown
17822 Apply telecine process to the video.
17824 This filter accepts the following options:
17833 The default value is @code{top}.
17837 A string of numbers representing the pulldown pattern you wish to apply.
17838 The default value is @code{23}.
17842 Some typical patterns:
17847 24p: 2332 (preferred)
17854 24p: 222222222223 ("Euro pulldown")
17859 @section thistogram
17861 Compute and draw a color distribution histogram for the input video across time.
17863 Unlike @ref{histogram} video filter which only shows histogram of single input frame
17864 at certain time, this filter shows also past histograms of number of frames defined
17865 by @code{width} option.
17867 The computed histogram is a representation of the color component
17868 distribution in an image.
17870 The filter accepts the following options:
17874 Set width of single color component output. Default value is @code{0}.
17875 Value of @code{0} means width will be picked from input video.
17876 This also set number of passed histograms to keep.
17877 Allowed range is [0, 8192].
17879 @item display_mode, d
17881 It accepts the following values:
17884 Per color component graphs are placed below each other.
17887 Per color component graphs are placed side by side.
17890 Presents information identical to that in the @code{parade}, except
17891 that the graphs representing color components are superimposed directly
17894 Default is @code{stack}.
17896 @item levels_mode, m
17897 Set mode. Can be either @code{linear}, or @code{logarithmic}.
17898 Default is @code{linear}.
17900 @item components, c
17901 Set what color components to display.
17902 Default is @code{7}.
17905 Set background opacity. Default is @code{0.9}.
17908 Show envelope. Default is disabled.
17911 Set envelope color. Default is @code{gold}.
17916 Apply threshold effect to video stream.
17918 This filter needs four video streams to perform thresholding.
17919 First stream is stream we are filtering.
17920 Second stream is holding threshold values, third stream is holding min values,
17921 and last, fourth stream is holding max values.
17923 The filter accepts the following option:
17927 Set which planes will be processed, unprocessed planes will be copied.
17928 By default value 0xf, all planes will be processed.
17931 For example if first stream pixel's component value is less then threshold value
17932 of pixel component from 2nd threshold stream, third stream value will picked,
17933 otherwise fourth stream pixel component value will be picked.
17935 Using color source filter one can perform various types of thresholding:
17937 @subsection Examples
17941 Binary threshold, using gray color as threshold:
17943 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
17947 Inverted binary threshold, using gray color as threshold:
17949 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
17953 Truncate binary threshold, using gray color as threshold:
17955 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
17959 Threshold to zero, using gray color as threshold:
17961 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
17965 Inverted threshold to zero, using gray color as threshold:
17967 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
17972 Select the most representative frame in a given sequence of consecutive frames.
17974 The filter accepts the following options:
17978 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
17979 will pick one of them, and then handle the next batch of @var{n} frames until
17980 the end. Default is @code{100}.
17983 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
17984 value will result in a higher memory usage, so a high value is not recommended.
17986 @subsection Examples
17990 Extract one picture each 50 frames:
17996 Complete example of a thumbnail creation with @command{ffmpeg}:
17998 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18004 Tile several successive frames together.
18006 The filter accepts the following options:
18011 Set the grid size (i.e. the number of lines and columns). For the syntax of
18012 this option, check the
18013 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18016 Set the maximum number of frames to render in the given area. It must be less
18017 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18018 the area will be used.
18021 Set the outer border margin in pixels.
18024 Set the inner border thickness (i.e. the number of pixels between frames). For
18025 more advanced padding options (such as having different values for the edges),
18026 refer to the pad video filter.
18029 Specify the color of the unused area. For the syntax of this option, check the
18030 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18031 The default value of @var{color} is "black".
18034 Set the number of frames to overlap when tiling several successive frames together.
18035 The value must be between @code{0} and @var{nb_frames - 1}.
18038 Set the number of frames to initially be empty before displaying first output frame.
18039 This controls how soon will one get first output frame.
18040 The value must be between @code{0} and @var{nb_frames - 1}.
18043 @subsection Examples
18047 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18049 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18051 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18052 duplicating each output frame to accommodate the originally detected frame
18056 Display @code{5} pictures in an area of @code{3x2} frames,
18057 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18058 mixed flat and named options:
18060 tile=3x2:nb_frames=5:padding=7:margin=2
18064 @section tinterlace
18066 Perform various types of temporal field interlacing.
18068 Frames are counted starting from 1, so the first input frame is
18071 The filter accepts the following options:
18076 Specify the mode of the interlacing. This option can also be specified
18077 as a value alone. See below for a list of values for this option.
18079 Available values are:
18083 Move odd frames into the upper field, even into the lower field,
18084 generating a double height frame at half frame rate.
18088 Frame 1 Frame 2 Frame 3 Frame 4
18090 11111 22222 33333 44444
18091 11111 22222 33333 44444
18092 11111 22222 33333 44444
18093 11111 22222 33333 44444
18107 Only output odd frames, even frames are dropped, generating a frame with
18108 unchanged height at half frame rate.
18113 Frame 1 Frame 2 Frame 3 Frame 4
18115 11111 22222 33333 44444
18116 11111 22222 33333 44444
18117 11111 22222 33333 44444
18118 11111 22222 33333 44444
18128 Only output even frames, odd frames are dropped, generating a frame with
18129 unchanged height at half frame rate.
18134 Frame 1 Frame 2 Frame 3 Frame 4
18136 11111 22222 33333 44444
18137 11111 22222 33333 44444
18138 11111 22222 33333 44444
18139 11111 22222 33333 44444
18149 Expand each frame to full height, but pad alternate lines with black,
18150 generating a frame with double height at the same input frame rate.
18155 Frame 1 Frame 2 Frame 3 Frame 4
18157 11111 22222 33333 44444
18158 11111 22222 33333 44444
18159 11111 22222 33333 44444
18160 11111 22222 33333 44444
18163 11111 ..... 33333 .....
18164 ..... 22222 ..... 44444
18165 11111 ..... 33333 .....
18166 ..... 22222 ..... 44444
18167 11111 ..... 33333 .....
18168 ..... 22222 ..... 44444
18169 11111 ..... 33333 .....
18170 ..... 22222 ..... 44444
18174 @item interleave_top, 4
18175 Interleave the upper field from odd frames with the lower field from
18176 even frames, generating a frame with unchanged height at half frame rate.
18181 Frame 1 Frame 2 Frame 3 Frame 4
18183 11111<- 22222 33333<- 44444
18184 11111 22222<- 33333 44444<-
18185 11111<- 22222 33333<- 44444
18186 11111 22222<- 33333 44444<-
18196 @item interleave_bottom, 5
18197 Interleave the lower field from odd frames with the upper field from
18198 even frames, generating a frame with unchanged height at half frame rate.
18203 Frame 1 Frame 2 Frame 3 Frame 4
18205 11111 22222<- 33333 44444<-
18206 11111<- 22222 33333<- 44444
18207 11111 22222<- 33333 44444<-
18208 11111<- 22222 33333<- 44444
18218 @item interlacex2, 6
18219 Double frame rate with unchanged height. Frames are inserted each
18220 containing the second temporal field from the previous input frame and
18221 the first temporal field from the next input frame. This mode relies on
18222 the top_field_first flag. Useful for interlaced video displays with no
18223 field synchronisation.
18228 Frame 1 Frame 2 Frame 3 Frame 4
18230 11111 22222 33333 44444
18231 11111 22222 33333 44444
18232 11111 22222 33333 44444
18233 11111 22222 33333 44444
18236 11111 22222 22222 33333 33333 44444 44444
18237 11111 11111 22222 22222 33333 33333 44444
18238 11111 22222 22222 33333 33333 44444 44444
18239 11111 11111 22222 22222 33333 33333 44444
18244 Move odd frames into the upper field, even into the lower field,
18245 generating a double height frame at same frame rate.
18250 Frame 1 Frame 2 Frame 3 Frame 4
18252 11111 22222 33333 44444
18253 11111 22222 33333 44444
18254 11111 22222 33333 44444
18255 11111 22222 33333 44444
18258 11111 33333 33333 55555
18259 22222 22222 44444 44444
18260 11111 33333 33333 55555
18261 22222 22222 44444 44444
18262 11111 33333 33333 55555
18263 22222 22222 44444 44444
18264 11111 33333 33333 55555
18265 22222 22222 44444 44444
18270 Numeric values are deprecated but are accepted for backward
18271 compatibility reasons.
18273 Default mode is @code{merge}.
18276 Specify flags influencing the filter process.
18278 Available value for @var{flags} is:
18281 @item low_pass_filter, vlpf
18282 Enable linear vertical low-pass filtering in the filter.
18283 Vertical low-pass filtering is required when creating an interlaced
18284 destination from a progressive source which contains high-frequency
18285 vertical detail. Filtering will reduce interlace 'twitter' and Moire
18288 @item complex_filter, cvlpf
18289 Enable complex vertical low-pass filtering.
18290 This will slightly less reduce interlace 'twitter' and Moire
18291 patterning but better retain detail and subjective sharpness impression.
18294 Bypass already interlaced frames, only adjust the frame rate.
18297 Vertical low-pass filtering and bypassing already interlaced frames can only be
18298 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
18304 Mix successive video frames.
18306 A description of the accepted options follows.
18310 The number of successive frames to mix. If unspecified, it defaults to 3.
18313 Specify weight of each input video frame.
18314 Each weight is separated by space. If number of weights is smaller than
18315 number of @var{frames} last specified weight will be used for all remaining
18319 Specify scale, if it is set it will be multiplied with sum
18320 of each weight multiplied with pixel values to give final destination
18321 pixel value. By default @var{scale} is auto scaled to sum of weights.
18324 @subsection Examples
18328 Average 7 successive frames:
18330 tmix=frames=7:weights="1 1 1 1 1 1 1"
18334 Apply simple temporal convolution:
18336 tmix=frames=3:weights="-1 3 -1"
18340 Similar as above but only showing temporal differences:
18342 tmix=frames=3:weights="-1 2 -1":scale=1
18348 Tone map colors from different dynamic ranges.
18350 This filter expects data in single precision floating point, as it needs to
18351 operate on (and can output) out-of-range values. Another filter, such as
18352 @ref{zscale}, is needed to convert the resulting frame to a usable format.
18354 The tonemapping algorithms implemented only work on linear light, so input
18355 data should be linearized beforehand (and possibly correctly tagged).
18358 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
18361 @subsection Options
18362 The filter accepts the following options.
18366 Set the tone map algorithm to use.
18368 Possible values are:
18371 Do not apply any tone map, only desaturate overbright pixels.
18374 Hard-clip any out-of-range values. Use it for perfect color accuracy for
18375 in-range values, while distorting out-of-range values.
18378 Stretch the entire reference gamut to a linear multiple of the display.
18381 Fit a logarithmic transfer between the tone curves.
18384 Preserve overall image brightness with a simple curve, using nonlinear
18385 contrast, which results in flattening details and degrading color accuracy.
18388 Preserve both dark and bright details better than @var{reinhard}, at the cost
18389 of slightly darkening everything. Use it when detail preservation is more
18390 important than color and brightness accuracy.
18393 Smoothly map out-of-range values, while retaining contrast and colors for
18394 in-range material as much as possible. Use it when color accuracy is more
18395 important than detail preservation.
18401 Tune the tone mapping algorithm.
18403 This affects the following algorithms:
18409 Specifies the scale factor to use while stretching.
18413 Specifies the exponent of the function.
18417 Specify an extra linear coefficient to multiply into the signal before clipping.
18421 Specify the local contrast coefficient at the display peak.
18422 Default to 0.5, which means that in-gamut values will be about half as bright
18429 Specify the transition point from linear to mobius transform. Every value
18430 below this point is guaranteed to be mapped 1:1. The higher the value, the
18431 more accurate the result will be, at the cost of losing bright details.
18432 Default to 0.3, which due to the steep initial slope still preserves in-range
18433 colors fairly accurately.
18437 Apply desaturation for highlights that exceed this level of brightness. The
18438 higher the parameter, the more color information will be preserved. This
18439 setting helps prevent unnaturally blown-out colors for super-highlights, by
18440 (smoothly) turning into white instead. This makes images feel more natural,
18441 at the cost of reducing information about out-of-range colors.
18443 The default of 2.0 is somewhat conservative and will mostly just apply to
18444 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
18446 This option works only if the input frame has a supported color tag.
18449 Override signal/nominal/reference peak with this value. Useful when the
18450 embedded peak information in display metadata is not reliable or when tone
18451 mapping from a lower range to a higher range.
18456 Temporarily pad video frames.
18458 The filter accepts the following options:
18462 Specify number of delay frames before input video stream.
18465 Specify number of padding frames after input video stream.
18466 Set to -1 to pad indefinitely.
18469 Set kind of frames added to beginning of stream.
18470 Can be either @var{add} or @var{clone}.
18471 With @var{add} frames of solid-color are added.
18472 With @var{clone} frames are clones of first frame.
18475 Set kind of frames added to end of stream.
18476 Can be either @var{add} or @var{clone}.
18477 With @var{add} frames of solid-color are added.
18478 With @var{clone} frames are clones of last frame.
18480 @item start_duration, stop_duration
18481 Specify the duration of the start/stop delay. See
18482 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18483 for the accepted syntax.
18484 These options override @var{start} and @var{stop}.
18487 Specify the color of the padded area. For the syntax of this option,
18488 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
18489 manual,ffmpeg-utils}.
18491 The default value of @var{color} is "black".
18497 Transpose rows with columns in the input video and optionally flip it.
18499 It accepts the following parameters:
18504 Specify the transposition direction.
18506 Can assume the following values:
18508 @item 0, 4, cclock_flip
18509 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
18517 Rotate by 90 degrees clockwise, that is:
18525 Rotate by 90 degrees counterclockwise, that is:
18532 @item 3, 7, clock_flip
18533 Rotate by 90 degrees clockwise and vertically flip, that is:
18541 For values between 4-7, the transposition is only done if the input
18542 video geometry is portrait and not landscape. These values are
18543 deprecated, the @code{passthrough} option should be used instead.
18545 Numerical values are deprecated, and should be dropped in favor of
18546 symbolic constants.
18549 Do not apply the transposition if the input geometry matches the one
18550 specified by the specified value. It accepts the following values:
18553 Always apply transposition.
18555 Preserve portrait geometry (when @var{height} >= @var{width}).
18557 Preserve landscape geometry (when @var{width} >= @var{height}).
18560 Default value is @code{none}.
18563 For example to rotate by 90 degrees clockwise and preserve portrait
18566 transpose=dir=1:passthrough=portrait
18569 The command above can also be specified as:
18571 transpose=1:portrait
18574 @section transpose_npp
18576 Transpose rows with columns in the input video and optionally flip it.
18577 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
18579 It accepts the following parameters:
18584 Specify the transposition direction.
18586 Can assume the following values:
18589 Rotate by 90 degrees counterclockwise and vertically flip. (default)
18592 Rotate by 90 degrees clockwise.
18595 Rotate by 90 degrees counterclockwise.
18598 Rotate by 90 degrees clockwise and vertically flip.
18602 Do not apply the transposition if the input geometry matches the one
18603 specified by the specified value. It accepts the following values:
18606 Always apply transposition. (default)
18608 Preserve portrait geometry (when @var{height} >= @var{width}).
18610 Preserve landscape geometry (when @var{width} >= @var{height}).
18616 Trim the input so that the output contains one continuous subpart of the input.
18618 It accepts the following parameters:
18621 Specify the time of the start of the kept section, i.e. the frame with the
18622 timestamp @var{start} will be the first frame in the output.
18625 Specify the time of the first frame that will be dropped, i.e. the frame
18626 immediately preceding the one with the timestamp @var{end} will be the last
18627 frame in the output.
18630 This is the same as @var{start}, except this option sets the start timestamp
18631 in timebase units instead of seconds.
18634 This is the same as @var{end}, except this option sets the end timestamp
18635 in timebase units instead of seconds.
18638 The maximum duration of the output in seconds.
18641 The number of the first frame that should be passed to the output.
18644 The number of the first frame that should be dropped.
18647 @option{start}, @option{end}, and @option{duration} are expressed as time
18648 duration specifications; see
18649 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18650 for the accepted syntax.
18652 Note that the first two sets of the start/end options and the @option{duration}
18653 option look at the frame timestamp, while the _frame variants simply count the
18654 frames that pass through the filter. Also note that this filter does not modify
18655 the timestamps. If you wish for the output timestamps to start at zero, insert a
18656 setpts filter after the trim filter.
18658 If multiple start or end options are set, this filter tries to be greedy and
18659 keep all the frames that match at least one of the specified constraints. To keep
18660 only the part that matches all the constraints at once, chain multiple trim
18663 The defaults are such that all the input is kept. So it is possible to set e.g.
18664 just the end values to keep everything before the specified time.
18669 Drop everything except the second minute of input:
18671 ffmpeg -i INPUT -vf trim=60:120
18675 Keep only the first second:
18677 ffmpeg -i INPUT -vf trim=duration=1
18682 @section unpremultiply
18683 Apply alpha unpremultiply effect to input video stream using first plane
18684 of second stream as alpha.
18686 Both streams must have same dimensions and same pixel format.
18688 The filter accepts the following option:
18692 Set which planes will be processed, unprocessed planes will be copied.
18693 By default value 0xf, all planes will be processed.
18695 If the format has 1 or 2 components, then luma is bit 0.
18696 If the format has 3 or 4 components:
18697 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
18698 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
18699 If present, the alpha channel is always the last bit.
18702 Do not require 2nd input for processing, instead use alpha plane from input stream.
18708 Sharpen or blur the input video.
18710 It accepts the following parameters:
18713 @item luma_msize_x, lx
18714 Set the luma matrix horizontal size. It must be an odd integer between
18715 3 and 23. The default value is 5.
18717 @item luma_msize_y, ly
18718 Set the luma matrix vertical size. It must be an odd integer between 3
18719 and 23. The default value is 5.
18721 @item luma_amount, la
18722 Set the luma effect strength. It must be a floating point number, reasonable
18723 values lay between -1.5 and 1.5.
18725 Negative values will blur the input video, while positive values will
18726 sharpen it, a value of zero will disable the effect.
18728 Default value is 1.0.
18730 @item chroma_msize_x, cx
18731 Set the chroma matrix horizontal size. It must be an odd integer
18732 between 3 and 23. The default value is 5.
18734 @item chroma_msize_y, cy
18735 Set the chroma matrix vertical size. It must be an odd integer
18736 between 3 and 23. The default value is 5.
18738 @item chroma_amount, ca
18739 Set the chroma effect strength. It must be a floating point number, reasonable
18740 values lay between -1.5 and 1.5.
18742 Negative values will blur the input video, while positive values will
18743 sharpen it, a value of zero will disable the effect.
18745 Default value is 0.0.
18749 All parameters are optional and default to the equivalent of the
18750 string '5:5:1.0:5:5:0.0'.
18752 @subsection Examples
18756 Apply strong luma sharpen effect:
18758 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
18762 Apply a strong blur of both luma and chroma parameters:
18764 unsharp=7:7:-2:7:7:-2
18770 Apply ultra slow/simple postprocessing filter that compresses and decompresses
18771 the image at several (or - in the case of @option{quality} level @code{8} - all)
18772 shifts and average the results.
18774 The way this differs from the behavior of spp is that uspp actually encodes &
18775 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
18776 DCT similar to MJPEG.
18778 The filter accepts the following options:
18782 Set quality. This option defines the number of levels for averaging. It accepts
18783 an integer in the range 0-8. If set to @code{0}, the filter will have no
18784 effect. A value of @code{8} means the higher quality. For each increment of
18785 that value the speed drops by a factor of approximately 2. Default value is
18789 Force a constant quantization parameter. If not set, the filter will use the QP
18790 from the video stream (if available).
18795 Convert 360 videos between various formats.
18797 The filter accepts the following options:
18803 Set format of the input/output video.
18811 Equirectangular projection.
18816 Cubemap with 3x2/6x1/1x6 layout.
18818 Format specific options:
18823 Set padding proportion for the input/output cubemap. Values in decimals.
18830 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)
18833 Default value is @b{@samp{0}}.
18837 Set fixed padding for the input/output cubemap. Values in pixels.
18839 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
18843 Set order of faces for the input/output cubemap. Choose one direction for each position.
18845 Designation of directions:
18861 Default value is @b{@samp{rludfb}}.
18865 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
18867 Designation of angles:
18870 0 degrees clockwise
18872 90 degrees clockwise
18874 180 degrees clockwise
18876 270 degrees clockwise
18879 Default value is @b{@samp{000000}}.
18883 Equi-Angular Cubemap.
18890 Format specific options:
18895 Set output horizontal/vertical/diagonal field of view. Values in degrees.
18897 If diagonal field of view is set it overrides horizontal and vertical field of view.
18902 Set input horizontal/vertical/diagonal field of view. Values in degrees.
18904 If diagonal field of view is set it overrides horizontal and vertical field of view.
18910 Format specific options:
18914 Set padding proportion. Values in decimals.
18924 Default value is @b{@samp{0}}.
18929 Facebook's 360 format.
18932 Stereographic format.
18934 Format specific options:
18939 Set output horizontal/vertical/diagonal field of view. Values in degrees.
18941 If diagonal field of view is set it overrides horizontal and vertical field of view.
18946 Set input horizontal/vertical/diagonal field of view. Values in degrees.
18948 If diagonal field of view is set it overrides horizontal and vertical field of view.
18955 Ball format, gives significant distortion toward the back.
18958 Hammer-Aitoff map projection format.
18961 Sinusoidal map projection format.
18964 Fisheye projection.
18966 Format specific options:
18971 Set output horizontal/vertical/diagonal field of view. Values in degrees.
18973 If diagonal field of view is set it overrides horizontal and vertical field of view.
18978 Set input horizontal/vertical/diagonal field of view. Values in degrees.
18980 If diagonal field of view is set it overrides horizontal and vertical field of view.
18984 Pannini projection. @i{(output only)}
18986 Format specific options:
18989 Set pannini parameter.
18993 Cylindrical projection.
18995 Format specific options:
19000 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19002 If diagonal field of view is set it overrides horizontal and vertical field of view.
19007 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19009 If diagonal field of view is set it overrides horizontal and vertical field of view.
19013 Perspective projection. @i{(output only)}
19015 Format specific options:
19018 Set perspective parameter.
19022 Tetrahedron projection.
19026 Set interpolation method.@*
19027 @i{Note: more complex interpolation methods require much more memory to run.}
19037 Bilinear interpolation.
19040 Bicubic interpolation.
19043 Lanczos interpolation.
19046 Spline16 interpolation.
19049 Gaussian interpolation.
19052 Default value is @b{@samp{line}}.
19056 Set the output video resolution.
19058 Default resolution depends on formats.
19062 Set the input/output stereo format.
19073 Default value is @b{@samp{2d}} for input and output format.
19078 Set rotation for the output video. Values in degrees.
19081 Set rotation order for the output video. Choose one item for each position.
19092 Default value is @b{@samp{ypr}}.
19097 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
19101 Set if input video is flipped horizontally/vertically. Boolean values.
19104 Set if input video is transposed. Boolean value, by default disabled.
19107 Set if output video needs to be transposed. Boolean value, by default disabled.
19110 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
19113 @subsection Examples
19117 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
19119 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
19122 Extract back view of Equi-Angular Cubemap:
19124 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
19127 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
19129 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
19133 @section vaguedenoiser
19135 Apply a wavelet based denoiser.
19137 It transforms each frame from the video input into the wavelet domain,
19138 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
19139 the obtained coefficients. It does an inverse wavelet transform after.
19140 Due to wavelet properties, it should give a nice smoothed result, and
19141 reduced noise, without blurring picture features.
19143 This filter accepts the following options:
19147 The filtering strength. The higher, the more filtered the video will be.
19148 Hard thresholding can use a higher threshold than soft thresholding
19149 before the video looks overfiltered. Default value is 2.
19152 The filtering method the filter will use.
19154 It accepts the following values:
19157 All values under the threshold will be zeroed.
19160 All values under the threshold will be zeroed. All values above will be
19161 reduced by the threshold.
19164 Scales or nullifies coefficients - intermediary between (more) soft and
19165 (less) hard thresholding.
19168 Default is garrote.
19171 Number of times, the wavelet will decompose the picture. Picture can't
19172 be decomposed beyond a particular point (typically, 8 for a 640x480
19173 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
19176 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
19179 A list of the planes to process. By default all planes are processed.
19182 @section vectorscope
19184 Display 2 color component values in the two dimensional graph (which is called
19187 This filter accepts the following options:
19191 Set vectorscope mode.
19193 It accepts the following values:
19197 Gray values are displayed on graph, higher brightness means more pixels have
19198 same component color value on location in graph. This is the default mode.
19201 Gray values are displayed on graph. Surrounding pixels values which are not
19202 present in video frame are drawn in gradient of 2 color components which are
19203 set by option @code{x} and @code{y}. The 3rd color component is static.
19206 Actual color components values present in video frame are displayed on graph.
19209 Similar as color2 but higher frequency of same values @code{x} and @code{y}
19210 on graph increases value of another color component, which is luminance by
19211 default values of @code{x} and @code{y}.
19214 Actual colors present in video frame are displayed on graph. If two different
19215 colors map to same position on graph then color with higher value of component
19216 not present in graph is picked.
19219 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
19220 component picked from radial gradient.
19224 Set which color component will be represented on X-axis. Default is @code{1}.
19227 Set which color component will be represented on Y-axis. Default is @code{2}.
19230 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
19231 of color component which represents frequency of (X, Y) location in graph.
19236 No envelope, this is default.
19239 Instant envelope, even darkest single pixel will be clearly highlighted.
19242 Hold maximum and minimum values presented in graph over time. This way you
19243 can still spot out of range values without constantly looking at vectorscope.
19246 Peak and instant envelope combined together.
19250 Set what kind of graticule to draw.
19259 Set graticule opacity.
19262 Set graticule flags.
19266 Draw graticule for white point.
19269 Draw graticule for black point.
19272 Draw color points short names.
19276 Set background opacity.
19278 @item lthreshold, l
19279 Set low threshold for color component not represented on X or Y axis.
19280 Values lower than this value will be ignored. Default is 0.
19281 Note this value is multiplied with actual max possible value one pixel component
19282 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
19285 @item hthreshold, h
19286 Set high threshold for color component not represented on X or Y axis.
19287 Values higher than this value will be ignored. Default is 1.
19288 Note this value is multiplied with actual max possible value one pixel component
19289 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
19290 is 0.9 * 255 = 230.
19292 @item colorspace, c
19293 Set what kind of colorspace to use when drawing graticule.
19303 Set color tint for gray/tint vectorscope mode. By default both options are zero.
19304 This means no tint, and output will remain gray.
19307 @anchor{vidstabdetect}
19308 @section vidstabdetect
19310 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
19311 @ref{vidstabtransform} for pass 2.
19313 This filter generates a file with relative translation and rotation
19314 transform information about subsequent frames, which is then used by
19315 the @ref{vidstabtransform} filter.
19317 To enable compilation of this filter you need to configure FFmpeg with
19318 @code{--enable-libvidstab}.
19320 This filter accepts the following options:
19324 Set the path to the file used to write the transforms information.
19325 Default value is @file{transforms.trf}.
19328 Set how shaky the video is and how quick the camera is. It accepts an
19329 integer in the range 1-10, a value of 1 means little shakiness, a
19330 value of 10 means strong shakiness. Default value is 5.
19333 Set the accuracy of the detection process. It must be a value in the
19334 range 1-15. A value of 1 means low accuracy, a value of 15 means high
19335 accuracy. Default value is 15.
19338 Set stepsize of the search process. The region around minimum is
19339 scanned with 1 pixel resolution. Default value is 6.
19342 Set minimum contrast. Below this value a local measurement field is
19343 discarded. Must be a floating point value in the range 0-1. Default
19347 Set reference frame number for tripod mode.
19349 If enabled, the motion of the frames is compared to a reference frame
19350 in the filtered stream, identified by the specified number. The idea
19351 is to compensate all movements in a more-or-less static scene and keep
19352 the camera view absolutely still.
19354 If set to 0, it is disabled. The frames are counted starting from 1.
19357 Show fields and transforms in the resulting frames. It accepts an
19358 integer in the range 0-2. Default value is 0, which disables any
19362 @subsection Examples
19366 Use default values:
19372 Analyze strongly shaky movie and put the results in file
19373 @file{mytransforms.trf}:
19375 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
19379 Visualize the result of internal transformations in the resulting
19382 vidstabdetect=show=1
19386 Analyze a video with medium shakiness using @command{ffmpeg}:
19388 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
19392 @anchor{vidstabtransform}
19393 @section vidstabtransform
19395 Video stabilization/deshaking: pass 2 of 2,
19396 see @ref{vidstabdetect} for pass 1.
19398 Read a file with transform information for each frame and
19399 apply/compensate them. Together with the @ref{vidstabdetect}
19400 filter this can be used to deshake videos. See also
19401 @url{http://public.hronopik.de/vid.stab}. It is important to also use
19402 the @ref{unsharp} filter, see below.
19404 To enable compilation of this filter you need to configure FFmpeg with
19405 @code{--enable-libvidstab}.
19407 @subsection Options
19411 Set path to the file used to read the transforms. Default value is
19412 @file{transforms.trf}.
19415 Set the number of frames (value*2 + 1) used for lowpass filtering the
19416 camera movements. Default value is 10.
19418 For example a number of 10 means that 21 frames are used (10 in the
19419 past and 10 in the future) to smoothen the motion in the video. A
19420 larger value leads to a smoother video, but limits the acceleration of
19421 the camera (pan/tilt movements). 0 is a special case where a static
19422 camera is simulated.
19425 Set the camera path optimization algorithm.
19427 Accepted values are:
19430 gaussian kernel low-pass filter on camera motion (default)
19432 averaging on transformations
19436 Set maximal number of pixels to translate frames. Default value is -1,
19440 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
19441 value is -1, meaning no limit.
19444 Specify how to deal with borders that may be visible due to movement
19447 Available values are:
19450 keep image information from previous frame (default)
19452 fill the border black
19456 Invert transforms if set to 1. Default value is 0.
19459 Consider transforms as relative to previous frame if set to 1,
19460 absolute if set to 0. Default value is 0.
19463 Set percentage to zoom. A positive value will result in a zoom-in
19464 effect, a negative value in a zoom-out effect. Default value is 0 (no
19468 Set optimal zooming to avoid borders.
19470 Accepted values are:
19475 optimal static zoom value is determined (only very strong movements
19476 will lead to visible borders) (default)
19478 optimal adaptive zoom value is determined (no borders will be
19479 visible), see @option{zoomspeed}
19482 Note that the value given at zoom is added to the one calculated here.
19485 Set percent to zoom maximally each frame (enabled when
19486 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
19490 Specify type of interpolation.
19492 Available values are:
19497 linear only horizontal
19499 linear in both directions (default)
19501 cubic in both directions (slow)
19505 Enable virtual tripod mode if set to 1, which is equivalent to
19506 @code{relative=0:smoothing=0}. Default value is 0.
19508 Use also @code{tripod} option of @ref{vidstabdetect}.
19511 Increase log verbosity if set to 1. Also the detected global motions
19512 are written to the temporary file @file{global_motions.trf}. Default
19516 @subsection Examples
19520 Use @command{ffmpeg} for a typical stabilization with default values:
19522 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
19525 Note the use of the @ref{unsharp} filter which is always recommended.
19528 Zoom in a bit more and load transform data from a given file:
19530 vidstabtransform=zoom=5:input="mytransforms.trf"
19534 Smoothen the video even more:
19536 vidstabtransform=smoothing=30
19542 Flip the input video vertically.
19544 For example, to vertically flip a video with @command{ffmpeg}:
19546 ffmpeg -i in.avi -vf "vflip" out.avi
19551 Detect variable frame rate video.
19553 This filter tries to detect if the input is variable or constant frame rate.
19555 At end it will output number of frames detected as having variable delta pts,
19556 and ones with constant delta pts.
19557 If there was frames with variable delta, than it will also show min, max and
19558 average delta encountered.
19562 Boost or alter saturation.
19564 The filter accepts the following options:
19567 Set strength of boost if positive value or strength of alter if negative value.
19568 Default is 0. Allowed range is from -2 to 2.
19571 Set the red balance. Default is 1. Allowed range is from -10 to 10.
19574 Set the green balance. Default is 1. Allowed range is from -10 to 10.
19577 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
19580 Set the red luma coefficient.
19583 Set the green luma coefficient.
19586 Set the blue luma coefficient.
19589 If @code{intensity} is negative and this is set to 1, colors will change,
19590 otherwise colors will be less saturated, more towards gray.
19593 @subsection Commands
19595 This filter supports the all above options as @ref{commands}.
19600 Make or reverse a natural vignetting effect.
19602 The filter accepts the following options:
19606 Set lens angle expression as a number of radians.
19608 The value is clipped in the @code{[0,PI/2]} range.
19610 Default value: @code{"PI/5"}
19614 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
19618 Set forward/backward mode.
19620 Available modes are:
19623 The larger the distance from the central point, the darker the image becomes.
19626 The larger the distance from the central point, the brighter the image becomes.
19627 This can be used to reverse a vignette effect, though there is no automatic
19628 detection to extract the lens @option{angle} and other settings (yet). It can
19629 also be used to create a burning effect.
19632 Default value is @samp{forward}.
19635 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
19637 It accepts the following values:
19640 Evaluate expressions only once during the filter initialization.
19643 Evaluate expressions for each incoming frame. This is way slower than the
19644 @samp{init} mode since it requires all the scalers to be re-computed, but it
19645 allows advanced dynamic expressions.
19648 Default value is @samp{init}.
19651 Set dithering to reduce the circular banding effects. Default is @code{1}
19655 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
19656 Setting this value to the SAR of the input will make a rectangular vignetting
19657 following the dimensions of the video.
19659 Default is @code{1/1}.
19662 @subsection Expressions
19664 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
19665 following parameters.
19670 input width and height
19673 the number of input frame, starting from 0
19676 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
19677 @var{TB} units, NAN if undefined
19680 frame rate of the input video, NAN if the input frame rate is unknown
19683 the PTS (Presentation TimeStamp) of the filtered video frame,
19684 expressed in seconds, NAN if undefined
19687 time base of the input video
19691 @subsection Examples
19695 Apply simple strong vignetting effect:
19701 Make a flickering vignetting:
19703 vignette='PI/4+random(1)*PI/50':eval=frame
19708 @section vmafmotion
19710 Obtain the average VMAF motion score of a video.
19711 It is one of the component metrics of VMAF.
19713 The obtained average motion score is printed through the logging system.
19715 The filter accepts the following options:
19719 If specified, the filter will use the named file to save the motion score of
19720 each frame with respect to the previous frame.
19721 When filename equals "-" the data is sent to standard output.
19726 ffmpeg -i ref.mpg -vf vmafmotion -f null -
19730 Stack input videos vertically.
19732 All streams must be of same pixel format and of same width.
19734 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
19735 to create same output.
19737 The filter accepts the following options:
19741 Set number of input streams. Default is 2.
19744 If set to 1, force the output to terminate when the shortest input
19745 terminates. Default value is 0.
19750 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
19751 Deinterlacing Filter").
19753 Based on the process described by Martin Weston for BBC R&D, and
19754 implemented based on the de-interlace algorithm written by Jim
19755 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
19756 uses filter coefficients calculated by BBC R&D.
19758 This filter uses field-dominance information in frame to decide which
19759 of each pair of fields to place first in the output.
19760 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
19762 There are two sets of filter coefficients, so called "simple"
19763 and "complex". Which set of filter coefficients is used can
19764 be set by passing an optional parameter:
19768 Set the interlacing filter coefficients. Accepts one of the following values:
19772 Simple filter coefficient set.
19774 More-complex filter coefficient set.
19776 Default value is @samp{complex}.
19779 Specify which frames to deinterlace. Accepts one of the following values:
19783 Deinterlace all frames,
19785 Only deinterlace frames marked as interlaced.
19788 Default value is @samp{all}.
19792 Video waveform monitor.
19794 The waveform monitor plots color component intensity. By default luminance
19795 only. Each column of the waveform corresponds to a column of pixels in the
19798 It accepts the following options:
19802 Can be either @code{row}, or @code{column}. Default is @code{column}.
19803 In row mode, the graph on the left side represents color component value 0 and
19804 the right side represents value = 255. In column mode, the top side represents
19805 color component value = 0 and bottom side represents value = 255.
19808 Set intensity. Smaller values are useful to find out how many values of the same
19809 luminance are distributed across input rows/columns.
19810 Default value is @code{0.04}. Allowed range is [0, 1].
19813 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
19814 In mirrored mode, higher values will be represented on the left
19815 side for @code{row} mode and at the top for @code{column} mode. Default is
19816 @code{1} (mirrored).
19820 It accepts the following values:
19823 Presents information identical to that in the @code{parade}, except
19824 that the graphs representing color components are superimposed directly
19827 This display mode makes it easier to spot relative differences or similarities
19828 in overlapping areas of the color components that are supposed to be identical,
19829 such as neutral whites, grays, or blacks.
19832 Display separate graph for the color components side by side in
19833 @code{row} mode or one below the other in @code{column} mode.
19836 Display separate graph for the color components side by side in
19837 @code{column} mode or one below the other in @code{row} mode.
19839 Using this display mode makes it easy to spot color casts in the highlights
19840 and shadows of an image, by comparing the contours of the top and the bottom
19841 graphs of each waveform. Since whites, grays, and blacks are characterized
19842 by exactly equal amounts of red, green, and blue, neutral areas of the picture
19843 should display three waveforms of roughly equal width/height. If not, the
19844 correction is easy to perform by making level adjustments the three waveforms.
19846 Default is @code{stack}.
19848 @item components, c
19849 Set which color components to display. Default is 1, which means only luminance
19850 or red color component if input is in RGB colorspace. If is set for example to
19851 7 it will display all 3 (if) available color components.
19856 No envelope, this is default.
19859 Instant envelope, minimum and maximum values presented in graph will be easily
19860 visible even with small @code{step} value.
19863 Hold minimum and maximum values presented in graph across time. This way you
19864 can still spot out of range values without constantly looking at waveforms.
19867 Peak and instant envelope combined together.
19873 No filtering, this is default.
19876 Luma and chroma combined together.
19879 Similar as above, but shows difference between blue and red chroma.
19882 Similar as above, but use different colors.
19885 Similar as above, but again with different colors.
19888 Displays only chroma.
19891 Displays actual color value on waveform.
19894 Similar as above, but with luma showing frequency of chroma values.
19898 Set which graticule to display.
19902 Do not display graticule.
19905 Display green graticule showing legal broadcast ranges.
19908 Display orange graticule showing legal broadcast ranges.
19911 Display invert graticule showing legal broadcast ranges.
19915 Set graticule opacity.
19918 Set graticule flags.
19922 Draw numbers above lines. By default enabled.
19925 Draw dots instead of lines.
19929 Set scale used for displaying graticule.
19936 Default is digital.
19939 Set background opacity.
19943 Set tint for output.
19944 Only used with lowpass filter and when display is not overlay and input
19945 pixel formats are not RGB.
19948 @section weave, doubleweave
19950 The @code{weave} takes a field-based video input and join
19951 each two sequential fields into single frame, producing a new double
19952 height clip with half the frame rate and half the frame count.
19954 The @code{doubleweave} works same as @code{weave} but without
19955 halving frame rate and frame count.
19957 It accepts the following option:
19961 Set first field. Available values are:
19965 Set the frame as top-field-first.
19968 Set the frame as bottom-field-first.
19972 @subsection Examples
19976 Interlace video using @ref{select} and @ref{separatefields} filter:
19978 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
19983 Apply the xBR high-quality magnification filter which is designed for pixel
19984 art. It follows a set of edge-detection rules, see
19985 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
19987 It accepts the following option:
19991 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
19992 @code{3xBR} and @code{4} for @code{4xBR}.
19993 Default is @code{3}.
19998 Apply cross fade from one input video stream to another input video stream.
19999 The cross fade is applied for specified duration.
20001 The filter accepts the following options:
20005 Set one of available transition effects:
20041 Default transition effect is fade.
20044 Set cross fade duration in seconds.
20045 Default duration is 1 second.
20048 Set cross fade start relative to first input stream in seconds.
20049 Default offset is 0.
20052 Set expression for custom transition effect.
20054 The expressions can use the following variables and functions:
20059 The coordinates of the current sample.
20063 The width and height of the image.
20066 Progress of transition effect.
20069 Currently processed plane.
20072 Return value of first input at current location and plane.
20075 Return value of second input at current location and plane.
20081 Return the value of the pixel at location (@var{x},@var{y}) of the
20082 first/second/third/fourth component of first input.
20088 Return the value of the pixel at location (@var{x},@var{y}) of the
20089 first/second/third/fourth component of second input.
20093 @subsection Examples
20097 Cross fade from one input video to another input video, with fade transition and duration of transition
20098 of 2 seconds starting at offset of 5 seconds:
20100 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
20105 Pick median pixels from several input videos.
20107 The filter accepts the following options:
20111 Set number of inputs.
20112 Default is 3. Allowed range is from 3 to 255.
20113 If number of inputs is even number, than result will be mean value between two median values.
20116 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
20119 Set median percentile. Default value is @code{0.5}.
20120 Default value of @code{0.5} will pick always median values, while @code{0} will pick
20121 minimum values, and @code{1} maximum values.
20125 Stack video inputs into custom layout.
20127 All streams must be of same pixel format.
20129 The filter accepts the following options:
20133 Set number of input streams. Default is 2.
20136 Specify layout of inputs.
20137 This option requires the desired layout configuration to be explicitly set by the user.
20138 This sets position of each video input in output. Each input
20139 is separated by '|'.
20140 The first number represents the column, and the second number represents the row.
20141 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
20142 where X is video input from which to take width or height.
20143 Multiple values can be used when separated by '+'. In such
20144 case values are summed together.
20146 Note that if inputs are of different sizes gaps may appear, as not all of
20147 the output video frame will be filled. Similarly, videos can overlap each
20148 other if their position doesn't leave enough space for the full frame of
20151 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
20152 a layout must be set by the user.
20155 If set to 1, force the output to terminate when the shortest input
20156 terminates. Default value is 0.
20159 If set to valid color, all unused pixels will be filled with that color.
20160 By default fill is set to none, so it is disabled.
20163 @subsection Examples
20167 Display 4 inputs into 2x2 grid.
20171 input1(0, 0) | input3(w0, 0)
20172 input2(0, h0) | input4(w0, h0)
20176 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
20179 Note that if inputs are of different sizes, gaps or overlaps may occur.
20182 Display 4 inputs into 1x4 grid.
20189 input4(0, h0+h1+h2)
20193 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
20196 Note that if inputs are of different widths, unused space will appear.
20199 Display 9 inputs into 3x3 grid.
20203 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
20204 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
20205 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
20209 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
20212 Note that if inputs are of different sizes, gaps or overlaps may occur.
20215 Display 16 inputs into 4x4 grid.
20219 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
20220 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
20221 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
20222 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
20226 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|
20227 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
20230 Note that if inputs are of different sizes, gaps or overlaps may occur.
20237 Deinterlace the input video ("yadif" means "yet another deinterlacing
20240 It accepts the following parameters:
20246 The interlacing mode to adopt. It accepts one of the following values:
20249 @item 0, send_frame
20250 Output one frame for each frame.
20251 @item 1, send_field
20252 Output one frame for each field.
20253 @item 2, send_frame_nospatial
20254 Like @code{send_frame}, but it skips the spatial interlacing check.
20255 @item 3, send_field_nospatial
20256 Like @code{send_field}, but it skips the spatial interlacing check.
20259 The default value is @code{send_frame}.
20262 The picture field parity assumed for the input interlaced video. It accepts one
20263 of the following values:
20267 Assume the top field is first.
20269 Assume the bottom field is first.
20271 Enable automatic detection of field parity.
20274 The default value is @code{auto}.
20275 If the interlacing is unknown or the decoder does not export this information,
20276 top field first will be assumed.
20279 Specify which frames to deinterlace. Accepts one of the following
20284 Deinterlace all frames.
20285 @item 1, interlaced
20286 Only deinterlace frames marked as interlaced.
20289 The default value is @code{all}.
20292 @section yadif_cuda
20294 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
20295 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
20298 It accepts the following parameters:
20304 The interlacing mode to adopt. It accepts one of the following values:
20307 @item 0, send_frame
20308 Output one frame for each frame.
20309 @item 1, send_field
20310 Output one frame for each field.
20311 @item 2, send_frame_nospatial
20312 Like @code{send_frame}, but it skips the spatial interlacing check.
20313 @item 3, send_field_nospatial
20314 Like @code{send_field}, but it skips the spatial interlacing check.
20317 The default value is @code{send_frame}.
20320 The picture field parity assumed for the input interlaced video. It accepts one
20321 of the following values:
20325 Assume the top field is first.
20327 Assume the bottom field is first.
20329 Enable automatic detection of field parity.
20332 The default value is @code{auto}.
20333 If the interlacing is unknown or the decoder does not export this information,
20334 top field first will be assumed.
20337 Specify which frames to deinterlace. Accepts one of the following
20342 Deinterlace all frames.
20343 @item 1, interlaced
20344 Only deinterlace frames marked as interlaced.
20347 The default value is @code{all}.
20352 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
20353 The algorithm is described in
20354 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
20356 It accepts the following parameters:
20360 Set the window radius. Default value is 3.
20363 Set which planes to filter. Default is only the first plane.
20366 Set blur strength. Default value is 128.
20369 @subsection Commands
20370 This filter supports same @ref{commands} as options.
20374 Apply Zoom & Pan effect.
20376 This filter accepts the following options:
20380 Set the zoom expression. Range is 1-10. Default is 1.
20384 Set the x and y expression. Default is 0.
20387 Set the duration expression in number of frames.
20388 This sets for how many number of frames effect will last for
20389 single input image.
20392 Set the output image size, default is 'hd720'.
20395 Set the output frame rate, default is '25'.
20398 Each expression can contain the following constants:
20417 Output frame count.
20421 Last calculated 'x' and 'y' position from 'x' and 'y' expression
20422 for current input frame.
20426 'x' and 'y' of last output frame of previous input frame or 0 when there was
20427 not yet such frame (first input frame).
20430 Last calculated zoom from 'z' expression for current input frame.
20433 Last calculated zoom of last output frame of previous input frame.
20436 Number of output frames for current input frame. Calculated from 'd' expression
20437 for each input frame.
20440 number of output frames created for previous input frame
20443 Rational number: input width / input height
20446 sample aspect ratio
20449 display aspect ratio
20453 @subsection Examples
20457 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
20459 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
20463 Zoom-in up to 1.5 and pan always at center of picture:
20465 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
20469 Same as above but without pausing:
20471 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
20477 Scale (resize) the input video, using the z.lib library:
20478 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
20479 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
20481 The zscale filter forces the output display aspect ratio to be the same
20482 as the input, by changing the output sample aspect ratio.
20484 If the input image format is different from the format requested by
20485 the next filter, the zscale filter will convert the input to the
20488 @subsection Options
20489 The filter accepts the following options.
20494 Set the output video dimension expression. Default value is the input
20497 If the @var{width} or @var{w} value is 0, the input width is used for
20498 the output. If the @var{height} or @var{h} value is 0, the input height
20499 is used for the output.
20501 If one and only one of the values is -n with n >= 1, the zscale filter
20502 will use a value that maintains the aspect ratio of the input image,
20503 calculated from the other specified dimension. After that it will,
20504 however, make sure that the calculated dimension is divisible by n and
20505 adjust the value if necessary.
20507 If both values are -n with n >= 1, the behavior will be identical to
20508 both values being set to 0 as previously detailed.
20510 See below for the list of accepted constants for use in the dimension
20514 Set the video size. For the syntax of this option, check the
20515 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20518 Set the dither type.
20520 Possible values are:
20525 @item error_diffusion
20531 Set the resize filter type.
20533 Possible values are:
20543 Default is bilinear.
20546 Set the color range.
20548 Possible values are:
20555 Default is same as input.
20558 Set the color primaries.
20560 Possible values are:
20570 Default is same as input.
20573 Set the transfer characteristics.
20575 Possible values are:
20589 Default is same as input.
20592 Set the colorspace matrix.
20594 Possible value are:
20605 Default is same as input.
20608 Set the input color range.
20610 Possible values are:
20617 Default is same as input.
20619 @item primariesin, pin
20620 Set the input color primaries.
20622 Possible values are:
20632 Default is same as input.
20634 @item transferin, tin
20635 Set the input transfer characteristics.
20637 Possible values are:
20648 Default is same as input.
20650 @item matrixin, min
20651 Set the input colorspace matrix.
20653 Possible value are:
20665 Set the output chroma location.
20667 Possible values are:
20678 @item chromalin, cin
20679 Set the input chroma location.
20681 Possible values are:
20693 Set the nominal peak luminance.
20696 The values of the @option{w} and @option{h} options are expressions
20697 containing the following constants:
20702 The input width and height
20706 These are the same as @var{in_w} and @var{in_h}.
20710 The output (scaled) width and height
20714 These are the same as @var{out_w} and @var{out_h}
20717 The same as @var{iw} / @var{ih}
20720 input sample aspect ratio
20723 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
20727 horizontal and vertical input chroma subsample values. For example for the
20728 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
20732 horizontal and vertical output chroma subsample values. For example for the
20733 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
20736 @subsection Commands
20738 This filter supports the following commands:
20742 Set the output video dimension expression.
20743 The command accepts the same syntax of the corresponding option.
20745 If the specified expression is not valid, it is kept at its current
20749 @c man end VIDEO FILTERS
20751 @chapter OpenCL Video Filters
20752 @c man begin OPENCL VIDEO FILTERS
20754 Below is a description of the currently available OpenCL video filters.
20756 To enable compilation of these filters you need to configure FFmpeg with
20757 @code{--enable-opencl}.
20759 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
20762 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
20763 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
20764 given device parameters.
20766 @item -filter_hw_device @var{name}
20767 Pass the hardware device called @var{name} to all filters in any filter graph.
20771 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
20775 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
20777 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
20781 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.
20783 @section avgblur_opencl
20785 Apply average blur filter.
20787 The filter accepts the following options:
20791 Set horizontal radius size.
20792 Range is @code{[1, 1024]} and default value is @code{1}.
20795 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20798 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
20801 @subsection Example
20805 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.
20807 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
20811 @section boxblur_opencl
20813 Apply a boxblur algorithm to the input video.
20815 It accepts the following parameters:
20819 @item luma_radius, lr
20820 @item luma_power, lp
20821 @item chroma_radius, cr
20822 @item chroma_power, cp
20823 @item alpha_radius, ar
20824 @item alpha_power, ap
20828 A description of the accepted options follows.
20831 @item luma_radius, lr
20832 @item chroma_radius, cr
20833 @item alpha_radius, ar
20834 Set an expression for the box radius in pixels used for blurring the
20835 corresponding input plane.
20837 The radius value must be a non-negative number, and must not be
20838 greater than the value of the expression @code{min(w,h)/2} for the
20839 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
20842 Default value for @option{luma_radius} is "2". If not specified,
20843 @option{chroma_radius} and @option{alpha_radius} default to the
20844 corresponding value set for @option{luma_radius}.
20846 The expressions can contain the following constants:
20850 The input width and height in pixels.
20854 The input chroma image width and height in pixels.
20858 The horizontal and vertical chroma subsample values. For example, for the
20859 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
20862 @item luma_power, lp
20863 @item chroma_power, cp
20864 @item alpha_power, ap
20865 Specify how many times the boxblur filter is applied to the
20866 corresponding plane.
20868 Default value for @option{luma_power} is 2. If not specified,
20869 @option{chroma_power} and @option{alpha_power} default to the
20870 corresponding value set for @option{luma_power}.
20872 A value of 0 will disable the effect.
20875 @subsection Examples
20877 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.
20881 Apply a boxblur filter with the luma, chroma, and alpha radius
20882 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.
20884 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
20885 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
20889 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.
20891 For the luma plane, a 2x2 box radius will be run once.
20893 For the chroma plane, a 4x4 box radius will be run 5 times.
20895 For the alpha plane, a 3x3 box radius will be run 7 times.
20897 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
20901 @section colorkey_opencl
20902 RGB colorspace color keying.
20904 The filter accepts the following options:
20908 The color which will be replaced with transparency.
20911 Similarity percentage with the key color.
20913 0.01 matches only the exact key color, while 1.0 matches everything.
20918 0.0 makes pixels either fully transparent, or not transparent at all.
20920 Higher values result in semi-transparent pixels, with a higher transparency
20921 the more similar the pixels color is to the key color.
20924 @subsection Examples
20928 Make every semi-green pixel in the input transparent with some slight blending:
20930 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
20934 @section convolution_opencl
20936 Apply convolution of 3x3, 5x5, 7x7 matrix.
20938 The filter accepts the following options:
20945 Set matrix for each plane.
20946 Matrix is sequence of 9, 25 or 49 signed numbers.
20947 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
20953 Set multiplier for calculated value for each plane.
20954 If unset or 0, it will be sum of all matrix elements.
20955 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
20961 Set bias for each plane. This value is added to the result of the multiplication.
20962 Useful for making the overall image brighter or darker.
20963 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
20967 @subsection Examples
20973 -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
20979 -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
20983 Apply edge enhance:
20985 -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
20991 -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
20995 Apply laplacian edge detector which includes diagonals:
20997 -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
21003 -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
21007 @section erosion_opencl
21009 Apply erosion effect to the video.
21011 This filter replaces the pixel by the local(3x3) minimum.
21013 It accepts the following options:
21020 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21021 If @code{0}, plane will remain unchanged.
21024 Flag which specifies the pixel to refer to.
21025 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21027 Flags to local 3x3 coordinates region centered on @code{x}:
21036 @subsection Example
21040 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.
21042 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21046 @section deshake_opencl
21047 Feature-point based video stabilization filter.
21049 The filter accepts the following options:
21053 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
21056 Whether or not additional debug info should be displayed, both in the processed output and in the console.
21058 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
21060 Viewing point matches in the output video is only supported for RGB input.
21062 Defaults to @code{0}.
21064 @item adaptive_crop
21065 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
21067 Defaults to @code{1}.
21069 @item refine_features
21070 Whether or not feature points should be refined at a sub-pixel level.
21072 This can be turned off for a slight performance gain at the cost of precision.
21074 Defaults to @code{1}.
21076 @item smooth_strength
21077 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
21079 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
21081 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
21083 Defaults to @code{0.0}.
21085 @item smooth_window_multiplier
21086 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
21088 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
21090 Acceptable values range from @code{0.1} to @code{10.0}.
21092 Larger values increase the amount of motion data available for determining how to smooth the camera path,
21093 potentially improving smoothness, but also increase latency and memory usage.
21095 Defaults to @code{2.0}.
21099 @subsection Examples
21103 Stabilize a video with a fixed, medium smoothing strength:
21105 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
21109 Stabilize a video with debugging (both in console and in rendered video):
21111 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
21115 @section dilation_opencl
21117 Apply dilation effect to the video.
21119 This filter replaces the pixel by the local(3x3) maximum.
21121 It accepts the following options:
21128 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21129 If @code{0}, plane will remain unchanged.
21132 Flag which specifies the pixel to refer to.
21133 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21135 Flags to local 3x3 coordinates region centered on @code{x}:
21144 @subsection Example
21148 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.
21150 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21154 @section nlmeans_opencl
21156 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
21158 @section overlay_opencl
21160 Overlay one video on top of another.
21162 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
21163 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
21165 The filter accepts the following options:
21170 Set the x coordinate of the overlaid video on the main video.
21171 Default value is @code{0}.
21174 Set the y coordinate of the overlaid video on the main video.
21175 Default value is @code{0}.
21179 @subsection Examples
21183 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
21185 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21188 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
21190 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21195 @section pad_opencl
21197 Add paddings to the input image, and place the original input at the
21198 provided @var{x}, @var{y} coordinates.
21200 It accepts the following options:
21205 Specify an expression for the size of the output image with the
21206 paddings added. If the value for @var{width} or @var{height} is 0, the
21207 corresponding input size is used for the output.
21209 The @var{width} expression can reference the value set by the
21210 @var{height} expression, and vice versa.
21212 The default value of @var{width} and @var{height} is 0.
21216 Specify the offsets to place the input image at within the padded area,
21217 with respect to the top/left border of the output image.
21219 The @var{x} expression can reference the value set by the @var{y}
21220 expression, and vice versa.
21222 The default value of @var{x} and @var{y} is 0.
21224 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
21225 so the input image is centered on the padded area.
21228 Specify the color of the padded area. For the syntax of this option,
21229 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
21230 manual,ffmpeg-utils}.
21233 Pad to an aspect instead to a resolution.
21236 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
21237 options are expressions containing the following constants:
21242 The input video width and height.
21246 These are the same as @var{in_w} and @var{in_h}.
21250 The output width and height (the size of the padded area), as
21251 specified by the @var{width} and @var{height} expressions.
21255 These are the same as @var{out_w} and @var{out_h}.
21259 The x and y offsets as specified by the @var{x} and @var{y}
21260 expressions, or NAN if not yet specified.
21263 same as @var{iw} / @var{ih}
21266 input sample aspect ratio
21269 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
21272 @section prewitt_opencl
21274 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
21276 The filter accepts the following option:
21280 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21283 Set value which will be multiplied with filtered result.
21284 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21287 Set value which will be added to filtered result.
21288 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21291 @subsection Example
21295 Apply the Prewitt operator with scale set to 2 and delta set to 10.
21297 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
21301 @anchor{program_opencl}
21302 @section program_opencl
21304 Filter video using an OpenCL program.
21309 OpenCL program source file.
21312 Kernel name in program.
21315 Number of inputs to the filter. Defaults to 1.
21318 Size of output frames. Defaults to the same as the first input.
21322 The program source file must contain a kernel function with the given name,
21323 which will be run once for each plane of the output. Each run on a plane
21324 gets enqueued as a separate 2D global NDRange with one work-item for each
21325 pixel to be generated. The global ID offset for each work-item is therefore
21326 the coordinates of a pixel in the destination image.
21328 The kernel function needs to take the following arguments:
21331 Destination image, @var{__write_only image2d_t}.
21333 This image will become the output; the kernel should write all of it.
21335 Frame index, @var{unsigned int}.
21337 This is a counter starting from zero and increasing by one for each frame.
21339 Source images, @var{__read_only image2d_t}.
21341 These are the most recent images on each input. The kernel may read from
21342 them to generate the output, but they can't be written to.
21349 Copy the input to the output (output must be the same size as the input).
21351 __kernel void copy(__write_only image2d_t destination,
21352 unsigned int index,
21353 __read_only image2d_t source)
21355 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
21357 int2 location = (int2)(get_global_id(0), get_global_id(1));
21359 float4 value = read_imagef(source, sampler, location);
21361 write_imagef(destination, location, value);
21366 Apply a simple transformation, rotating the input by an amount increasing
21367 with the index counter. Pixel values are linearly interpolated by the
21368 sampler, and the output need not have the same dimensions as the input.
21370 __kernel void rotate_image(__write_only image2d_t dst,
21371 unsigned int index,
21372 __read_only image2d_t src)
21374 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21375 CLK_FILTER_LINEAR);
21377 float angle = (float)index / 100.0f;
21379 float2 dst_dim = convert_float2(get_image_dim(dst));
21380 float2 src_dim = convert_float2(get_image_dim(src));
21382 float2 dst_cen = dst_dim / 2.0f;
21383 float2 src_cen = src_dim / 2.0f;
21385 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
21387 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
21389 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
21390 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
21392 src_pos = src_pos * src_dim / dst_dim;
21394 float2 src_loc = src_pos + src_cen;
21396 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
21397 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
21398 write_imagef(dst, dst_loc, 0.5f);
21400 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
21405 Blend two inputs together, with the amount of each input used varying
21406 with the index counter.
21408 __kernel void blend_images(__write_only image2d_t dst,
21409 unsigned int index,
21410 __read_only image2d_t src1,
21411 __read_only image2d_t src2)
21413 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21414 CLK_FILTER_LINEAR);
21416 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
21418 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
21419 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
21420 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
21422 float4 val1 = read_imagef(src1, sampler, src1_loc);
21423 float4 val2 = read_imagef(src2, sampler, src2_loc);
21425 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
21431 @section roberts_opencl
21432 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
21434 The filter accepts the following option:
21438 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21441 Set value which will be multiplied with filtered result.
21442 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21445 Set value which will be added to filtered result.
21446 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21449 @subsection Example
21453 Apply the Roberts cross operator with scale set to 2 and delta set to 10
21455 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
21459 @section sobel_opencl
21461 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
21463 The filter accepts the following option:
21467 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21470 Set value which will be multiplied with filtered result.
21471 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21474 Set value which will be added to filtered result.
21475 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21478 @subsection Example
21482 Apply sobel operator with scale set to 2 and delta set to 10
21484 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
21488 @section tonemap_opencl
21490 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
21492 It accepts the following parameters:
21496 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
21499 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
21502 Apply desaturation for highlights that exceed this level of brightness. The
21503 higher the parameter, the more color information will be preserved. This
21504 setting helps prevent unnaturally blown-out colors for super-highlights, by
21505 (smoothly) turning into white instead. This makes images feel more natural,
21506 at the cost of reducing information about out-of-range colors.
21508 The default value is 0.5, and the algorithm here is a little different from
21509 the cpu version tonemap currently. A setting of 0.0 disables this option.
21512 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
21513 is used to detect whether the scene has changed or not. If the distance between
21514 the current frame average brightness and the current running average exceeds
21515 a threshold value, we would re-calculate scene average and peak brightness.
21516 The default value is 0.2.
21519 Specify the output pixel format.
21521 Currently supported formats are:
21528 Set the output color range.
21530 Possible values are:
21536 Default is same as input.
21539 Set the output color primaries.
21541 Possible values are:
21547 Default is same as input.
21550 Set the output transfer characteristics.
21552 Possible values are:
21561 Set the output colorspace matrix.
21563 Possible value are:
21569 Default is same as input.
21573 @subsection Example
21577 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
21579 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
21583 @section unsharp_opencl
21585 Sharpen or blur the input video.
21587 It accepts the following parameters:
21590 @item luma_msize_x, lx
21591 Set the luma matrix horizontal size.
21592 Range is @code{[1, 23]} and default value is @code{5}.
21594 @item luma_msize_y, ly
21595 Set the luma matrix vertical size.
21596 Range is @code{[1, 23]} and default value is @code{5}.
21598 @item luma_amount, la
21599 Set the luma effect strength.
21600 Range is @code{[-10, 10]} and default value is @code{1.0}.
21602 Negative values will blur the input video, while positive values will
21603 sharpen it, a value of zero will disable the effect.
21605 @item chroma_msize_x, cx
21606 Set the chroma matrix horizontal size.
21607 Range is @code{[1, 23]} and default value is @code{5}.
21609 @item chroma_msize_y, cy
21610 Set the chroma matrix vertical size.
21611 Range is @code{[1, 23]} and default value is @code{5}.
21613 @item chroma_amount, ca
21614 Set the chroma effect strength.
21615 Range is @code{[-10, 10]} and default value is @code{0.0}.
21617 Negative values will blur the input video, while positive values will
21618 sharpen it, a value of zero will disable the effect.
21622 All parameters are optional and default to the equivalent of the
21623 string '5:5:1.0:5:5:0.0'.
21625 @subsection Examples
21629 Apply strong luma sharpen effect:
21631 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
21635 Apply a strong blur of both luma and chroma parameters:
21637 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
21641 @section xfade_opencl
21643 Cross fade two videos with custom transition effect by using OpenCL.
21645 It accepts the following options:
21649 Set one of possible transition effects.
21653 Select custom transition effect, the actual transition description
21654 will be picked from source and kernel options.
21666 Default transition is fade.
21670 OpenCL program source file for custom transition.
21673 Set name of kernel to use for custom transition from program source file.
21676 Set duration of video transition.
21679 Set time of start of transition relative to first video.
21682 The program source file must contain a kernel function with the given name,
21683 which will be run once for each plane of the output. Each run on a plane
21684 gets enqueued as a separate 2D global NDRange with one work-item for each
21685 pixel to be generated. The global ID offset for each work-item is therefore
21686 the coordinates of a pixel in the destination image.
21688 The kernel function needs to take the following arguments:
21691 Destination image, @var{__write_only image2d_t}.
21693 This image will become the output; the kernel should write all of it.
21696 First Source image, @var{__read_only image2d_t}.
21697 Second Source image, @var{__read_only image2d_t}.
21699 These are the most recent images on each input. The kernel may read from
21700 them to generate the output, but they can't be written to.
21703 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
21710 Apply dots curtain transition effect:
21712 __kernel void blend_images(__write_only image2d_t dst,
21713 __read_only image2d_t src1,
21714 __read_only image2d_t src2,
21717 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21718 CLK_FILTER_LINEAR);
21719 int2 p = (int2)(get_global_id(0), get_global_id(1));
21720 float2 rp = (float2)(get_global_id(0), get_global_id(1));
21721 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
21724 float2 dots = (float2)(20.0, 20.0);
21725 float2 center = (float2)(0,0);
21728 float4 val1 = read_imagef(src1, sampler, p);
21729 float4 val2 = read_imagef(src2, sampler, p);
21730 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
21732 write_imagef(dst, p, next ? val1 : val2);
21738 @c man end OPENCL VIDEO FILTERS
21740 @chapter VAAPI Video Filters
21741 @c man begin VAAPI VIDEO FILTERS
21743 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
21745 To enable compilation of these filters you need to configure FFmpeg with
21746 @code{--enable-vaapi}.
21748 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}
21750 @section tonemap_vaapi
21752 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
21753 It maps the dynamic range of HDR10 content to the SDR content.
21754 It currently only accepts HDR10 as input.
21756 It accepts the following parameters:
21760 Specify the output pixel format.
21762 Currently supported formats are:
21771 Set the output color primaries.
21773 Default is same as input.
21776 Set the output transfer characteristics.
21781 Set the output colorspace matrix.
21783 Default is same as input.
21787 @subsection Example
21791 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
21793 tonemap_vaapi=format=p010:t=bt2020-10
21797 @c man end VAAPI VIDEO FILTERS
21799 @chapter Video Sources
21800 @c man begin VIDEO SOURCES
21802 Below is a description of the currently available video sources.
21806 Buffer video frames, and make them available to the filter chain.
21808 This source is mainly intended for a programmatic use, in particular
21809 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
21811 It accepts the following parameters:
21816 Specify the size (width and height) of the buffered video frames. For the
21817 syntax of this option, check the
21818 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21821 The input video width.
21824 The input video height.
21827 A string representing the pixel format of the buffered video frames.
21828 It may be a number corresponding to a pixel format, or a pixel format
21832 Specify the timebase assumed by the timestamps of the buffered frames.
21835 Specify the frame rate expected for the video stream.
21837 @item pixel_aspect, sar
21838 The sample (pixel) aspect ratio of the input video.
21841 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
21842 to the filtergraph description to specify swscale flags for automatically
21843 inserted scalers. See @ref{Filtergraph syntax}.
21845 @item hw_frames_ctx
21846 When using a hardware pixel format, this should be a reference to an
21847 AVHWFramesContext describing input frames.
21852 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
21855 will instruct the source to accept video frames with size 320x240 and
21856 with format "yuv410p", assuming 1/24 as the timestamps timebase and
21857 square pixels (1:1 sample aspect ratio).
21858 Since the pixel format with name "yuv410p" corresponds to the number 6
21859 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
21860 this example corresponds to:
21862 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
21865 Alternatively, the options can be specified as a flat string, but this
21866 syntax is deprecated:
21868 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
21872 Create a pattern generated by an elementary cellular automaton.
21874 The initial state of the cellular automaton can be defined through the
21875 @option{filename} and @option{pattern} options. If such options are
21876 not specified an initial state is created randomly.
21878 At each new frame a new row in the video is filled with the result of
21879 the cellular automaton next generation. The behavior when the whole
21880 frame is filled is defined by the @option{scroll} option.
21882 This source accepts the following options:
21886 Read the initial cellular automaton state, i.e. the starting row, from
21887 the specified file.
21888 In the file, each non-whitespace character is considered an alive
21889 cell, a newline will terminate the row, and further characters in the
21890 file will be ignored.
21893 Read the initial cellular automaton state, i.e. the starting row, from
21894 the specified string.
21896 Each non-whitespace character in the string is considered an alive
21897 cell, a newline will terminate the row, and further characters in the
21898 string will be ignored.
21901 Set the video rate, that is the number of frames generated per second.
21904 @item random_fill_ratio, ratio
21905 Set the random fill ratio for the initial cellular automaton row. It
21906 is a floating point number value ranging from 0 to 1, defaults to
21909 This option is ignored when a file or a pattern is specified.
21911 @item random_seed, seed
21912 Set the seed for filling randomly the initial row, must be an integer
21913 included between 0 and UINT32_MAX. If not specified, or if explicitly
21914 set to -1, the filter will try to use a good random seed on a best
21918 Set the cellular automaton rule, it is a number ranging from 0 to 255.
21919 Default value is 110.
21922 Set the size of the output video. For the syntax of this option, check the
21923 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21925 If @option{filename} or @option{pattern} is specified, the size is set
21926 by default to the width of the specified initial state row, and the
21927 height is set to @var{width} * PHI.
21929 If @option{size} is set, it must contain the width of the specified
21930 pattern string, and the specified pattern will be centered in the
21933 If a filename or a pattern string is not specified, the size value
21934 defaults to "320x518" (used for a randomly generated initial state).
21937 If set to 1, scroll the output upward when all the rows in the output
21938 have been already filled. If set to 0, the new generated row will be
21939 written over the top row just after the bottom row is filled.
21942 @item start_full, full
21943 If set to 1, completely fill the output with generated rows before
21944 outputting the first frame.
21945 This is the default behavior, for disabling set the value to 0.
21948 If set to 1, stitch the left and right row edges together.
21949 This is the default behavior, for disabling set the value to 0.
21952 @subsection Examples
21956 Read the initial state from @file{pattern}, and specify an output of
21959 cellauto=f=pattern:s=200x400
21963 Generate a random initial row with a width of 200 cells, with a fill
21966 cellauto=ratio=2/3:s=200x200
21970 Create a pattern generated by rule 18 starting by a single alive cell
21971 centered on an initial row with width 100:
21973 cellauto=p=@@:s=100x400:full=0:rule=18
21977 Specify a more elaborated initial pattern:
21979 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
21984 @anchor{coreimagesrc}
21985 @section coreimagesrc
21986 Video source generated on GPU using Apple's CoreImage API on OSX.
21988 This video source is a specialized version of the @ref{coreimage} video filter.
21989 Use a core image generator at the beginning of the applied filterchain to
21990 generate the content.
21992 The coreimagesrc video source accepts the following options:
21994 @item list_generators
21995 List all available generators along with all their respective options as well as
21996 possible minimum and maximum values along with the default values.
21998 list_generators=true
22002 Specify the size of the sourced video. For the syntax of this option, check the
22003 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22004 The default value is @code{320x240}.
22007 Specify the frame rate of the sourced video, as the number of frames
22008 generated per second. It has to be a string in the format
22009 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22010 number or a valid video frame rate abbreviation. The default value is
22014 Set the sample aspect ratio of the sourced video.
22017 Set the duration of the sourced video. See
22018 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22019 for the accepted syntax.
22021 If not specified, or the expressed duration is negative, the video is
22022 supposed to be generated forever.
22025 Additionally, all options of the @ref{coreimage} video filter are accepted.
22026 A complete filterchain can be used for further processing of the
22027 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
22028 and examples for details.
22030 @subsection Examples
22035 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
22036 given as complete and escaped command-line for Apple's standard bash shell:
22038 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
22040 This example is equivalent to the QRCode example of @ref{coreimage} without the
22041 need for a nullsrc video source.
22045 @section mandelbrot
22047 Generate a Mandelbrot set fractal, and progressively zoom towards the
22048 point specified with @var{start_x} and @var{start_y}.
22050 This source accepts the following options:
22055 Set the terminal pts value. Default value is 400.
22058 Set the terminal scale value.
22059 Must be a floating point value. Default value is 0.3.
22062 Set the inner coloring mode, that is the algorithm used to draw the
22063 Mandelbrot fractal internal region.
22065 It shall assume one of the following values:
22070 Show time until convergence.
22072 Set color based on point closest to the origin of the iterations.
22077 Default value is @var{mincol}.
22080 Set the bailout value. Default value is 10.0.
22083 Set the maximum of iterations performed by the rendering
22084 algorithm. Default value is 7189.
22087 Set outer coloring mode.
22088 It shall assume one of following values:
22090 @item iteration_count
22091 Set iteration count mode.
22092 @item normalized_iteration_count
22093 set normalized iteration count mode.
22095 Default value is @var{normalized_iteration_count}.
22098 Set frame rate, expressed as number of frames per second. Default
22102 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22103 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22106 Set the initial scale value. Default value is 3.0.
22109 Set the initial x position. Must be a floating point value between
22110 -100 and 100. Default value is -0.743643887037158704752191506114774.
22113 Set the initial y position. Must be a floating point value between
22114 -100 and 100. Default value is -0.131825904205311970493132056385139.
22119 Generate various test patterns, as generated by the MPlayer test filter.
22121 The size of the generated video is fixed, and is 256x256.
22122 This source is useful in particular for testing encoding features.
22124 This source accepts the following options:
22129 Specify the frame rate of the sourced video, as the number of frames
22130 generated per second. It has to be a string in the format
22131 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22132 number or a valid video frame rate abbreviation. The default value is
22136 Set the duration of the sourced video. See
22137 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22138 for the accepted syntax.
22140 If not specified, or the expressed duration is negative, the video is
22141 supposed to be generated forever.
22145 Set the number or the name of the test to perform. Supported tests are:
22159 @item max_frames, m
22160 Set the maximum number of frames generated for each test, default value is 30.
22164 Default value is "all", which will cycle through the list of all tests.
22169 mptestsrc=t=dc_luma
22172 will generate a "dc_luma" test pattern.
22174 @section frei0r_src
22176 Provide a frei0r source.
22178 To enable compilation of this filter you need to install the frei0r
22179 header and configure FFmpeg with @code{--enable-frei0r}.
22181 This source accepts the following parameters:
22186 The size of the video to generate. For the syntax of this option, check the
22187 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22190 The framerate of the generated video. It may be a string of the form
22191 @var{num}/@var{den} or a frame rate abbreviation.
22194 The name to the frei0r source to load. For more information regarding frei0r and
22195 how to set the parameters, read the @ref{frei0r} section in the video filters
22198 @item filter_params
22199 A '|'-separated list of parameters to pass to the frei0r source.
22203 For example, to generate a frei0r partik0l source with size 200x200
22204 and frame rate 10 which is overlaid on the overlay filter main input:
22206 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
22211 Generate a life pattern.
22213 This source is based on a generalization of John Conway's life game.
22215 The sourced input represents a life grid, each pixel represents a cell
22216 which can be in one of two possible states, alive or dead. Every cell
22217 interacts with its eight neighbours, which are the cells that are
22218 horizontally, vertically, or diagonally adjacent.
22220 At each interaction the grid evolves according to the adopted rule,
22221 which specifies the number of neighbor alive cells which will make a
22222 cell stay alive or born. The @option{rule} option allows one to specify
22225 This source accepts the following options:
22229 Set the file from which to read the initial grid state. In the file,
22230 each non-whitespace character is considered an alive cell, and newline
22231 is used to delimit the end of each row.
22233 If this option is not specified, the initial grid is generated
22237 Set the video rate, that is the number of frames generated per second.
22240 @item random_fill_ratio, ratio
22241 Set the random fill ratio for the initial random grid. It is a
22242 floating point number value ranging from 0 to 1, defaults to 1/PHI.
22243 It is ignored when a file is specified.
22245 @item random_seed, seed
22246 Set the seed for filling the initial random grid, must be an integer
22247 included between 0 and UINT32_MAX. If not specified, or if explicitly
22248 set to -1, the filter will try to use a good random seed on a best
22254 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
22255 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
22256 @var{NS} specifies the number of alive neighbor cells which make a
22257 live cell stay alive, and @var{NB} the number of alive neighbor cells
22258 which make a dead cell to become alive (i.e. to "born").
22259 "s" and "b" can be used in place of "S" and "B", respectively.
22261 Alternatively a rule can be specified by an 18-bits integer. The 9
22262 high order bits are used to encode the next cell state if it is alive
22263 for each number of neighbor alive cells, the low order bits specify
22264 the rule for "borning" new cells. Higher order bits encode for an
22265 higher number of neighbor cells.
22266 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
22267 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
22269 Default value is "S23/B3", which is the original Conway's game of life
22270 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
22271 cells, and will born a new cell if there are three alive cells around
22275 Set the size of the output video. For the syntax of this option, check the
22276 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22278 If @option{filename} is specified, the size is set by default to the
22279 same size of the input file. If @option{size} is set, it must contain
22280 the size specified in the input file, and the initial grid defined in
22281 that file is centered in the larger resulting area.
22283 If a filename is not specified, the size value defaults to "320x240"
22284 (used for a randomly generated initial grid).
22287 If set to 1, stitch the left and right grid edges together, and the
22288 top and bottom edges also. Defaults to 1.
22291 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
22292 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
22293 value from 0 to 255.
22296 Set the color of living (or new born) cells.
22299 Set the color of dead cells. If @option{mold} is set, this is the first color
22300 used to represent a dead cell.
22303 Set mold color, for definitely dead and moldy cells.
22305 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
22306 ffmpeg-utils manual,ffmpeg-utils}.
22309 @subsection Examples
22313 Read a grid from @file{pattern}, and center it on a grid of size
22316 life=f=pattern:s=300x300
22320 Generate a random grid of size 200x200, with a fill ratio of 2/3:
22322 life=ratio=2/3:s=200x200
22326 Specify a custom rule for evolving a randomly generated grid:
22332 Full example with slow death effect (mold) using @command{ffplay}:
22334 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
22341 @anchor{haldclutsrc}
22344 @anchor{pal100bars}
22345 @anchor{rgbtestsrc}
22347 @anchor{smptehdbars}
22350 @anchor{yuvtestsrc}
22351 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
22353 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
22355 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
22357 The @code{color} source provides an uniformly colored input.
22359 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
22360 @ref{haldclut} filter.
22362 The @code{nullsrc} source returns unprocessed video frames. It is
22363 mainly useful to be employed in analysis / debugging tools, or as the
22364 source for filters which ignore the input data.
22366 The @code{pal75bars} source generates a color bars pattern, based on
22367 EBU PAL recommendations with 75% color levels.
22369 The @code{pal100bars} source generates a color bars pattern, based on
22370 EBU PAL recommendations with 100% color levels.
22372 The @code{rgbtestsrc} source generates an RGB test pattern useful for
22373 detecting RGB vs BGR issues. You should see a red, green and blue
22374 stripe from top to bottom.
22376 The @code{smptebars} source generates a color bars pattern, based on
22377 the SMPTE Engineering Guideline EG 1-1990.
22379 The @code{smptehdbars} source generates a color bars pattern, based on
22380 the SMPTE RP 219-2002.
22382 The @code{testsrc} source generates a test video pattern, showing a
22383 color pattern, a scrolling gradient and a timestamp. This is mainly
22384 intended for testing purposes.
22386 The @code{testsrc2} source is similar to testsrc, but supports more
22387 pixel formats instead of just @code{rgb24}. This allows using it as an
22388 input for other tests without requiring a format conversion.
22390 The @code{yuvtestsrc} source generates an YUV test pattern. You should
22391 see a y, cb and cr stripe from top to bottom.
22393 The sources accept the following parameters:
22398 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
22399 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
22400 pixels to be used as identity matrix for 3D lookup tables. Each component is
22401 coded on a @code{1/(N*N)} scale.
22404 Specify the color of the source, only available in the @code{color}
22405 source. For the syntax of this option, check the
22406 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
22409 Specify the size of the sourced video. For the syntax of this option, check the
22410 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22411 The default value is @code{320x240}.
22413 This option is not available with the @code{allrgb}, @code{allyuv}, and
22414 @code{haldclutsrc} filters.
22417 Specify the frame rate of the sourced video, as the number of frames
22418 generated per second. It has to be a string in the format
22419 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22420 number or a valid video frame rate abbreviation. The default value is
22424 Set the duration of the sourced video. See
22425 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22426 for the accepted syntax.
22428 If not specified, or the expressed duration is negative, the video is
22429 supposed to be generated forever.
22432 Set the sample aspect ratio of the sourced video.
22435 Specify the alpha (opacity) of the background, only available in the
22436 @code{testsrc2} source. The value must be between 0 (fully transparent) and
22437 255 (fully opaque, the default).
22440 Set the number of decimals to show in the timestamp, only available in the
22441 @code{testsrc} source.
22443 The displayed timestamp value will correspond to the original
22444 timestamp value multiplied by the power of 10 of the specified
22445 value. Default value is 0.
22448 @subsection Examples
22452 Generate a video with a duration of 5.3 seconds, with size
22453 176x144 and a frame rate of 10 frames per second:
22455 testsrc=duration=5.3:size=qcif:rate=10
22459 The following graph description will generate a red source
22460 with an opacity of 0.2, with size "qcif" and a frame rate of 10
22463 color=c=red@@0.2:s=qcif:r=10
22467 If the input content is to be ignored, @code{nullsrc} can be used. The
22468 following command generates noise in the luminance plane by employing
22469 the @code{geq} filter:
22471 nullsrc=s=256x256, geq=random(1)*255:128:128
22475 @subsection Commands
22477 The @code{color} source supports the following commands:
22481 Set the color of the created image. Accepts the same syntax of the
22482 corresponding @option{color} option.
22487 Generate video using an OpenCL program.
22492 OpenCL program source file.
22495 Kernel name in program.
22498 Size of frames to generate. This must be set.
22501 Pixel format to use for the generated frames. This must be set.
22504 Number of frames generated every second. Default value is '25'.
22508 For details of how the program loading works, see the @ref{program_opencl}
22515 Generate a colour ramp by setting pixel values from the position of the pixel
22516 in the output image. (Note that this will work with all pixel formats, but
22517 the generated output will not be the same.)
22519 __kernel void ramp(__write_only image2d_t dst,
22520 unsigned int index)
22522 int2 loc = (int2)(get_global_id(0), get_global_id(1));
22525 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
22527 write_imagef(dst, loc, val);
22532 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
22534 __kernel void sierpinski_carpet(__write_only image2d_t dst,
22535 unsigned int index)
22537 int2 loc = (int2)(get_global_id(0), get_global_id(1));
22539 float4 value = 0.0f;
22540 int x = loc.x + index;
22541 int y = loc.y + index;
22542 while (x > 0 || y > 0) {
22543 if (x % 3 == 1 && y % 3 == 1) {
22551 write_imagef(dst, loc, value);
22557 @section sierpinski
22559 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
22561 This source accepts the following options:
22565 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22566 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22569 Set frame rate, expressed as number of frames per second. Default
22573 Set seed which is used for random panning.
22576 Set max jump for single pan destination. Allowed range is from 1 to 10000.
22579 Set fractal type, can be default @code{carpet} or @code{triangle}.
22582 @c man end VIDEO SOURCES
22584 @chapter Video Sinks
22585 @c man begin VIDEO SINKS
22587 Below is a description of the currently available video sinks.
22589 @section buffersink
22591 Buffer video frames, and make them available to the end of the filter
22594 This sink is mainly intended for programmatic use, in particular
22595 through the interface defined in @file{libavfilter/buffersink.h}
22596 or the options system.
22598 It accepts a pointer to an AVBufferSinkContext structure, which
22599 defines the incoming buffers' formats, to be passed as the opaque
22600 parameter to @code{avfilter_init_filter} for initialization.
22604 Null video sink: do absolutely nothing with the input video. It is
22605 mainly useful as a template and for use in analysis / debugging
22608 @c man end VIDEO SINKS
22610 @chapter Multimedia Filters
22611 @c man begin MULTIMEDIA FILTERS
22613 Below is a description of the currently available multimedia filters.
22617 Convert input audio to a video output, displaying the audio bit scope.
22619 The filter accepts the following options:
22623 Set frame rate, expressed as number of frames per second. Default
22627 Specify the video size for the output. For the syntax of this option, check the
22628 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22629 Default value is @code{1024x256}.
22632 Specify list of colors separated by space or by '|' which will be used to
22633 draw channels. Unrecognized or missing colors will be replaced
22637 @section adrawgraph
22638 Draw a graph using input audio metadata.
22640 See @ref{drawgraph}
22642 @section agraphmonitor
22644 See @ref{graphmonitor}.
22646 @section ahistogram
22648 Convert input audio to a video output, displaying the volume histogram.
22650 The filter accepts the following options:
22654 Specify how histogram is calculated.
22656 It accepts the following values:
22659 Use single histogram for all channels.
22661 Use separate histogram for each channel.
22663 Default is @code{single}.
22666 Set frame rate, expressed as number of frames per second. Default
22670 Specify the video size for the output. For the syntax of this option, check the
22671 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22672 Default value is @code{hd720}.
22677 It accepts the following values:
22688 reverse logarithmic
22690 Default is @code{log}.
22693 Set amplitude scale.
22695 It accepts the following values:
22702 Default is @code{log}.
22705 Set how much frames to accumulate in histogram.
22706 Default is 1. Setting this to -1 accumulates all frames.
22709 Set histogram ratio of window height.
22712 Set sonogram sliding.
22714 It accepts the following values:
22717 replace old rows with new ones.
22719 scroll from top to bottom.
22721 Default is @code{replace}.
22724 @section aphasemeter
22726 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
22727 representing mean phase of current audio frame. A video output can also be produced and is
22728 enabled by default. The audio is passed through as first output.
22730 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
22731 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
22732 and @code{1} means channels are in phase.
22734 The filter accepts the following options, all related to its video output:
22738 Set the output frame rate. Default value is @code{25}.
22741 Set the video size for the output. For the syntax of this option, check the
22742 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22743 Default value is @code{800x400}.
22748 Specify the red, green, blue contrast. Default values are @code{2},
22749 @code{7} and @code{1}.
22750 Allowed range is @code{[0, 255]}.
22753 Set color which will be used for drawing median phase. If color is
22754 @code{none} which is default, no median phase value will be drawn.
22757 Enable video output. Default is enabled.
22760 @section avectorscope
22762 Convert input audio to a video output, representing the audio vector
22765 The filter is used to measure the difference between channels of stereo
22766 audio stream. A monaural signal, consisting of identical left and right
22767 signal, results in straight vertical line. Any stereo separation is visible
22768 as a deviation from this line, creating a Lissajous figure.
22769 If the straight (or deviation from it) but horizontal line appears this
22770 indicates that the left and right channels are out of phase.
22772 The filter accepts the following options:
22776 Set the vectorscope mode.
22778 Available values are:
22781 Lissajous rotated by 45 degrees.
22784 Same as above but not rotated.
22787 Shape resembling half of circle.
22790 Default value is @samp{lissajous}.
22793 Set the video size for the output. For the syntax of this option, check the
22794 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22795 Default value is @code{400x400}.
22798 Set the output frame rate. Default value is @code{25}.
22804 Specify the red, green, blue and alpha contrast. Default values are @code{40},
22805 @code{160}, @code{80} and @code{255}.
22806 Allowed range is @code{[0, 255]}.
22812 Specify the red, green, blue and alpha fade. Default values are @code{15},
22813 @code{10}, @code{5} and @code{5}.
22814 Allowed range is @code{[0, 255]}.
22817 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
22818 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
22821 Set the vectorscope drawing mode.
22823 Available values are:
22826 Draw dot for each sample.
22829 Draw line between previous and current sample.
22832 Default value is @samp{dot}.
22835 Specify amplitude scale of audio samples.
22837 Available values are:
22853 Swap left channel axis with right channel axis.
22863 Mirror only x axis.
22866 Mirror only y axis.
22874 @subsection Examples
22878 Complete example using @command{ffplay}:
22880 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
22881 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
22885 @section bench, abench
22887 Benchmark part of a filtergraph.
22889 The filter accepts the following options:
22893 Start or stop a timer.
22895 Available values are:
22898 Get the current time, set it as frame metadata (using the key
22899 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
22902 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
22903 the input frame metadata to get the time difference. Time difference, average,
22904 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
22905 @code{min}) are then printed. The timestamps are expressed in seconds.
22909 @subsection Examples
22913 Benchmark @ref{selectivecolor} filter:
22915 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
22921 Concatenate audio and video streams, joining them together one after the
22924 The filter works on segments of synchronized video and audio streams. All
22925 segments must have the same number of streams of each type, and that will
22926 also be the number of streams at output.
22928 The filter accepts the following options:
22933 Set the number of segments. Default is 2.
22936 Set the number of output video streams, that is also the number of video
22937 streams in each segment. Default is 1.
22940 Set the number of output audio streams, that is also the number of audio
22941 streams in each segment. Default is 0.
22944 Activate unsafe mode: do not fail if segments have a different format.
22948 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
22949 @var{a} audio outputs.
22951 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
22952 segment, in the same order as the outputs, then the inputs for the second
22955 Related streams do not always have exactly the same duration, for various
22956 reasons including codec frame size or sloppy authoring. For that reason,
22957 related synchronized streams (e.g. a video and its audio track) should be
22958 concatenated at once. The concat filter will use the duration of the longest
22959 stream in each segment (except the last one), and if necessary pad shorter
22960 audio streams with silence.
22962 For this filter to work correctly, all segments must start at timestamp 0.
22964 All corresponding streams must have the same parameters in all segments; the
22965 filtering system will automatically select a common pixel format for video
22966 streams, and a common sample format, sample rate and channel layout for
22967 audio streams, but other settings, such as resolution, must be converted
22968 explicitly by the user.
22970 Different frame rates are acceptable but will result in variable frame rate
22971 at output; be sure to configure the output file to handle it.
22973 @subsection Examples
22977 Concatenate an opening, an episode and an ending, all in bilingual version
22978 (video in stream 0, audio in streams 1 and 2):
22980 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
22981 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
22982 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
22983 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
22987 Concatenate two parts, handling audio and video separately, using the
22988 (a)movie sources, and adjusting the resolution:
22990 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
22991 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
22992 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
22994 Note that a desync will happen at the stitch if the audio and video streams
22995 do not have exactly the same duration in the first file.
22999 @subsection Commands
23001 This filter supports the following commands:
23004 Close the current segment and step to the next one
23010 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
23011 level. By default, it logs a message at a frequency of 10Hz with the
23012 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
23013 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
23015 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
23016 sample format is double-precision floating point. The input stream will be converted to
23017 this specification, if needed. Users may need to insert aformat and/or aresample filters
23018 after this filter to obtain the original parameters.
23020 The filter also has a video output (see the @var{video} option) with a real
23021 time graph to observe the loudness evolution. The graphic contains the logged
23022 message mentioned above, so it is not printed anymore when this option is set,
23023 unless the verbose logging is set. The main graphing area contains the
23024 short-term loudness (3 seconds of analysis), and the gauge on the right is for
23025 the momentary loudness (400 milliseconds), but can optionally be configured
23026 to instead display short-term loudness (see @var{gauge}).
23028 The green area marks a +/- 1LU target range around the target loudness
23029 (-23LUFS by default, unless modified through @var{target}).
23031 More information about the Loudness Recommendation EBU R128 on
23032 @url{http://tech.ebu.ch/loudness}.
23034 The filter accepts the following options:
23039 Activate the video output. The audio stream is passed unchanged whether this
23040 option is set or no. The video stream will be the first output stream if
23041 activated. Default is @code{0}.
23044 Set the video size. This option is for video only. For the syntax of this
23046 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23047 Default and minimum resolution is @code{640x480}.
23050 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
23051 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
23052 other integer value between this range is allowed.
23055 Set metadata injection. If set to @code{1}, the audio input will be segmented
23056 into 100ms output frames, each of them containing various loudness information
23057 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
23059 Default is @code{0}.
23062 Force the frame logging level.
23064 Available values are:
23067 information logging level
23069 verbose logging level
23072 By default, the logging level is set to @var{info}. If the @option{video} or
23073 the @option{metadata} options are set, it switches to @var{verbose}.
23078 Available modes can be cumulated (the option is a @code{flag} type). Possible
23082 Disable any peak mode (default).
23084 Enable sample-peak mode.
23086 Simple peak mode looking for the higher sample value. It logs a message
23087 for sample-peak (identified by @code{SPK}).
23089 Enable true-peak mode.
23091 If enabled, the peak lookup is done on an over-sampled version of the input
23092 stream for better peak accuracy. It logs a message for true-peak.
23093 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
23094 This mode requires a build with @code{libswresample}.
23098 Treat mono input files as "dual mono". If a mono file is intended for playback
23099 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
23100 If set to @code{true}, this option will compensate for this effect.
23101 Multi-channel input files are not affected by this option.
23104 Set a specific pan law to be used for the measurement of dual mono files.
23105 This parameter is optional, and has a default value of -3.01dB.
23108 Set a specific target level (in LUFS) used as relative zero in the visualization.
23109 This parameter is optional and has a default value of -23LUFS as specified
23110 by EBU R128. However, material published online may prefer a level of -16LUFS
23111 (e.g. for use with podcasts or video platforms).
23114 Set the value displayed by the gauge. Valid values are @code{momentary} and s
23115 @code{shortterm}. By default the momentary value will be used, but in certain
23116 scenarios it may be more useful to observe the short term value instead (e.g.
23120 Sets the display scale for the loudness. Valid parameters are @code{absolute}
23121 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
23122 video output, not the summary or continuous log output.
23125 @subsection Examples
23129 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
23131 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
23135 Run an analysis with @command{ffmpeg}:
23137 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
23141 @section interleave, ainterleave
23143 Temporally interleave frames from several inputs.
23145 @code{interleave} works with video inputs, @code{ainterleave} with audio.
23147 These filters read frames from several inputs and send the oldest
23148 queued frame to the output.
23150 Input streams must have well defined, monotonically increasing frame
23153 In order to submit one frame to output, these filters need to enqueue
23154 at least one frame for each input, so they cannot work in case one
23155 input is not yet terminated and will not receive incoming frames.
23157 For example consider the case when one input is a @code{select} filter
23158 which always drops input frames. The @code{interleave} filter will keep
23159 reading from that input, but it will never be able to send new frames
23160 to output until the input sends an end-of-stream signal.
23162 Also, depending on inputs synchronization, the filters will drop
23163 frames in case one input receives more frames than the other ones, and
23164 the queue is already filled.
23166 These filters accept the following options:
23170 Set the number of different inputs, it is 2 by default.
23173 @subsection Examples
23177 Interleave frames belonging to different streams using @command{ffmpeg}:
23179 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
23183 Add flickering blur effect:
23185 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
23189 @section metadata, ametadata
23191 Manipulate frame metadata.
23193 This filter accepts the following options:
23197 Set mode of operation of the filter.
23199 Can be one of the following:
23203 If both @code{value} and @code{key} is set, select frames
23204 which have such metadata. If only @code{key} is set, select
23205 every frame that has such key in metadata.
23208 Add new metadata @code{key} and @code{value}. If key is already available
23212 Modify value of already present key.
23215 If @code{value} is set, delete only keys that have such value.
23216 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
23220 Print key and its value if metadata was found. If @code{key} is not set print all
23221 metadata values available in frame.
23225 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
23228 Set metadata value which will be used. This option is mandatory for
23229 @code{modify} and @code{add} mode.
23232 Which function to use when comparing metadata value and @code{value}.
23234 Can be one of following:
23238 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
23241 Values are interpreted as strings, returns true if metadata value starts with
23242 the @code{value} option string.
23245 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
23248 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
23251 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
23254 Values are interpreted as floats, returns true if expression from option @code{expr}
23258 Values are interpreted as strings, returns true if metadata value ends with
23259 the @code{value} option string.
23263 Set expression which is used when @code{function} is set to @code{expr}.
23264 The expression is evaluated through the eval API and can contain the following
23269 Float representation of @code{value} from metadata key.
23272 Float representation of @code{value} as supplied by user in @code{value} option.
23276 If specified in @code{print} mode, output is written to the named file. Instead of
23277 plain filename any writable url can be specified. Filename ``-'' is a shorthand
23278 for standard output. If @code{file} option is not set, output is written to the log
23279 with AV_LOG_INFO loglevel.
23282 Reduces buffering in print mode when output is written to a URL set using @var{file}.
23286 @subsection Examples
23290 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
23293 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
23296 Print silencedetect output to file @file{metadata.txt}.
23298 silencedetect,ametadata=mode=print:file=metadata.txt
23301 Direct all metadata to a pipe with file descriptor 4.
23303 metadata=mode=print:file='pipe\:4'
23307 @section perms, aperms
23309 Set read/write permissions for the output frames.
23311 These filters are mainly aimed at developers to test direct path in the
23312 following filter in the filtergraph.
23314 The filters accept the following options:
23318 Select the permissions mode.
23320 It accepts the following values:
23323 Do nothing. This is the default.
23325 Set all the output frames read-only.
23327 Set all the output frames directly writable.
23329 Make the frame read-only if writable, and writable if read-only.
23331 Set each output frame read-only or writable randomly.
23335 Set the seed for the @var{random} mode, must be an integer included between
23336 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
23337 @code{-1}, the filter will try to use a good random seed on a best effort
23341 Note: in case of auto-inserted filter between the permission filter and the
23342 following one, the permission might not be received as expected in that
23343 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
23344 perms/aperms filter can avoid this problem.
23346 @section realtime, arealtime
23348 Slow down filtering to match real time approximately.
23350 These filters will pause the filtering for a variable amount of time to
23351 match the output rate with the input timestamps.
23352 They are similar to the @option{re} option to @code{ffmpeg}.
23354 They accept the following options:
23358 Time limit for the pauses. Any pause longer than that will be considered
23359 a timestamp discontinuity and reset the timer. Default is 2 seconds.
23361 Speed factor for processing. The value must be a float larger than zero.
23362 Values larger than 1.0 will result in faster than realtime processing,
23363 smaller will slow processing down. The @var{limit} is automatically adapted
23364 accordingly. Default is 1.0.
23366 A processing speed faster than what is possible without these filters cannot
23371 @section select, aselect
23373 Select frames to pass in output.
23375 This filter accepts the following options:
23380 Set expression, which is evaluated for each input frame.
23382 If the expression is evaluated to zero, the frame is discarded.
23384 If the evaluation result is negative or NaN, the frame is sent to the
23385 first output; otherwise it is sent to the output with index
23386 @code{ceil(val)-1}, assuming that the input index starts from 0.
23388 For example a value of @code{1.2} corresponds to the output with index
23389 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
23392 Set the number of outputs. The output to which to send the selected
23393 frame is based on the result of the evaluation. Default value is 1.
23396 The expression can contain the following constants:
23400 The (sequential) number of the filtered frame, starting from 0.
23403 The (sequential) number of the selected frame, starting from 0.
23405 @item prev_selected_n
23406 The sequential number of the last selected frame. It's NAN if undefined.
23409 The timebase of the input timestamps.
23412 The PTS (Presentation TimeStamp) of the filtered video frame,
23413 expressed in @var{TB} units. It's NAN if undefined.
23416 The PTS of the filtered video frame,
23417 expressed in seconds. It's NAN if undefined.
23420 The PTS of the previously filtered video frame. It's NAN if undefined.
23422 @item prev_selected_pts
23423 The PTS of the last previously filtered video frame. It's NAN if undefined.
23425 @item prev_selected_t
23426 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
23429 The PTS of the first video frame in the video. It's NAN if undefined.
23432 The time of the first video frame in the video. It's NAN if undefined.
23434 @item pict_type @emph{(video only)}
23435 The type of the filtered frame. It can assume one of the following
23447 @item interlace_type @emph{(video only)}
23448 The frame interlace type. It can assume one of the following values:
23451 The frame is progressive (not interlaced).
23453 The frame is top-field-first.
23455 The frame is bottom-field-first.
23458 @item consumed_sample_n @emph{(audio only)}
23459 the number of selected samples before the current frame
23461 @item samples_n @emph{(audio only)}
23462 the number of samples in the current frame
23464 @item sample_rate @emph{(audio only)}
23465 the input sample rate
23468 This is 1 if the filtered frame is a key-frame, 0 otherwise.
23471 the position in the file of the filtered frame, -1 if the information
23472 is not available (e.g. for synthetic video)
23474 @item scene @emph{(video only)}
23475 value between 0 and 1 to indicate a new scene; a low value reflects a low
23476 probability for the current frame to introduce a new scene, while a higher
23477 value means the current frame is more likely to be one (see the example below)
23479 @item concatdec_select
23480 The concat demuxer can select only part of a concat input file by setting an
23481 inpoint and an outpoint, but the output packets may not be entirely contained
23482 in the selected interval. By using this variable, it is possible to skip frames
23483 generated by the concat demuxer which are not exactly contained in the selected
23486 This works by comparing the frame pts against the @var{lavf.concat.start_time}
23487 and the @var{lavf.concat.duration} packet metadata values which are also
23488 present in the decoded frames.
23490 The @var{concatdec_select} variable is -1 if the frame pts is at least
23491 start_time and either the duration metadata is missing or the frame pts is less
23492 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
23495 That basically means that an input frame is selected if its pts is within the
23496 interval set by the concat demuxer.
23500 The default value of the select expression is "1".
23502 @subsection Examples
23506 Select all frames in input:
23511 The example above is the same as:
23523 Select only I-frames:
23525 select='eq(pict_type\,I)'
23529 Select one frame every 100:
23531 select='not(mod(n\,100))'
23535 Select only frames contained in the 10-20 time interval:
23537 select=between(t\,10\,20)
23541 Select only I-frames contained in the 10-20 time interval:
23543 select=between(t\,10\,20)*eq(pict_type\,I)
23547 Select frames with a minimum distance of 10 seconds:
23549 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
23553 Use aselect to select only audio frames with samples number > 100:
23555 aselect='gt(samples_n\,100)'
23559 Create a mosaic of the first scenes:
23561 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
23564 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
23568 Send even and odd frames to separate outputs, and compose them:
23570 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
23574 Select useful frames from an ffconcat file which is using inpoints and
23575 outpoints but where the source files are not intra frame only.
23577 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
23581 @section sendcmd, asendcmd
23583 Send commands to filters in the filtergraph.
23585 These filters read commands to be sent to other filters in the
23588 @code{sendcmd} must be inserted between two video filters,
23589 @code{asendcmd} must be inserted between two audio filters, but apart
23590 from that they act the same way.
23592 The specification of commands can be provided in the filter arguments
23593 with the @var{commands} option, or in a file specified by the
23594 @var{filename} option.
23596 These filters accept the following options:
23599 Set the commands to be read and sent to the other filters.
23601 Set the filename of the commands to be read and sent to the other
23605 @subsection Commands syntax
23607 A commands description consists of a sequence of interval
23608 specifications, comprising a list of commands to be executed when a
23609 particular event related to that interval occurs. The occurring event
23610 is typically the current frame time entering or leaving a given time
23613 An interval is specified by the following syntax:
23615 @var{START}[-@var{END}] @var{COMMANDS};
23618 The time interval is specified by the @var{START} and @var{END} times.
23619 @var{END} is optional and defaults to the maximum time.
23621 The current frame time is considered within the specified interval if
23622 it is included in the interval [@var{START}, @var{END}), that is when
23623 the time is greater or equal to @var{START} and is lesser than
23626 @var{COMMANDS} consists of a sequence of one or more command
23627 specifications, separated by ",", relating to that interval. The
23628 syntax of a command specification is given by:
23630 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
23633 @var{FLAGS} is optional and specifies the type of events relating to
23634 the time interval which enable sending the specified command, and must
23635 be a non-null sequence of identifier flags separated by "+" or "|" and
23636 enclosed between "[" and "]".
23638 The following flags are recognized:
23641 The command is sent when the current frame timestamp enters the
23642 specified interval. In other words, the command is sent when the
23643 previous frame timestamp was not in the given interval, and the
23647 The command is sent when the current frame timestamp leaves the
23648 specified interval. In other words, the command is sent when the
23649 previous frame timestamp was in the given interval, and the
23653 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
23656 @var{TARGET} specifies the target of the command, usually the name of
23657 the filter class or a specific filter instance name.
23659 @var{COMMAND} specifies the name of the command for the target filter.
23661 @var{ARG} is optional and specifies the optional list of argument for
23662 the given @var{COMMAND}.
23664 Between one interval specification and another, whitespaces, or
23665 sequences of characters starting with @code{#} until the end of line,
23666 are ignored and can be used to annotate comments.
23668 A simplified BNF description of the commands specification syntax
23671 @var{COMMAND_FLAG} ::= "enter" | "leave"
23672 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
23673 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
23674 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
23675 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
23676 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
23679 @subsection Examples
23683 Specify audio tempo change at second 4:
23685 asendcmd=c='4.0 atempo tempo 1.5',atempo
23689 Target a specific filter instance:
23691 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
23695 Specify a list of drawtext and hue commands in a file.
23697 # show text in the interval 5-10
23698 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
23699 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
23701 # desaturate the image in the interval 15-20
23702 15.0-20.0 [enter] hue s 0,
23703 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
23705 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
23707 # apply an exponential saturation fade-out effect, starting from time 25
23708 25 [enter] hue s exp(25-t)
23711 A filtergraph allowing to read and process the above command list
23712 stored in a file @file{test.cmd}, can be specified with:
23714 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
23719 @section setpts, asetpts
23721 Change the PTS (presentation timestamp) of the input frames.
23723 @code{setpts} works on video frames, @code{asetpts} on audio frames.
23725 This filter accepts the following options:
23730 The expression which is evaluated for each frame to construct its timestamp.
23734 The expression is evaluated through the eval API and can contain the following
23738 @item FRAME_RATE, FR
23739 frame rate, only defined for constant frame-rate video
23742 The presentation timestamp in input
23745 The count of the input frame for video or the number of consumed samples,
23746 not including the current frame for audio, starting from 0.
23748 @item NB_CONSUMED_SAMPLES
23749 The number of consumed samples, not including the current frame (only
23752 @item NB_SAMPLES, S
23753 The number of samples in the current frame (only audio)
23755 @item SAMPLE_RATE, SR
23756 The audio sample rate.
23759 The PTS of the first frame.
23762 the time in seconds of the first frame
23765 State whether the current frame is interlaced.
23768 the time in seconds of the current frame
23771 original position in the file of the frame, or undefined if undefined
23772 for the current frame
23775 The previous input PTS.
23778 previous input time in seconds
23781 The previous output PTS.
23784 previous output time in seconds
23787 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
23791 The wallclock (RTC) time at the start of the movie in microseconds.
23794 The timebase of the input timestamps.
23798 @subsection Examples
23802 Start counting PTS from zero
23804 setpts=PTS-STARTPTS
23808 Apply fast motion effect:
23814 Apply slow motion effect:
23820 Set fixed rate of 25 frames per second:
23826 Set fixed rate 25 fps with some jitter:
23828 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
23832 Apply an offset of 10 seconds to the input PTS:
23838 Generate timestamps from a "live source" and rebase onto the current timebase:
23840 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
23844 Generate timestamps by counting samples:
23853 Force color range for the output video frame.
23855 The @code{setrange} filter marks the color range property for the
23856 output frames. It does not change the input frame, but only sets the
23857 corresponding property, which affects how the frame is treated by
23860 The filter accepts the following options:
23865 Available values are:
23869 Keep the same color range property.
23871 @item unspecified, unknown
23872 Set the color range as unspecified.
23874 @item limited, tv, mpeg
23875 Set the color range as limited.
23877 @item full, pc, jpeg
23878 Set the color range as full.
23882 @section settb, asettb
23884 Set the timebase to use for the output frames timestamps.
23885 It is mainly useful for testing timebase configuration.
23887 It accepts the following parameters:
23892 The expression which is evaluated into the output timebase.
23896 The value for @option{tb} is an arithmetic expression representing a
23897 rational. The expression can contain the constants "AVTB" (the default
23898 timebase), "intb" (the input timebase) and "sr" (the sample rate,
23899 audio only). Default value is "intb".
23901 @subsection Examples
23905 Set the timebase to 1/25:
23911 Set the timebase to 1/10:
23917 Set the timebase to 1001/1000:
23923 Set the timebase to 2*intb:
23929 Set the default timebase value:
23936 Convert input audio to a video output representing frequency spectrum
23937 logarithmically using Brown-Puckette constant Q transform algorithm with
23938 direct frequency domain coefficient calculation (but the transform itself
23939 is not really constant Q, instead the Q factor is actually variable/clamped),
23940 with musical tone scale, from E0 to D#10.
23942 The filter accepts the following options:
23946 Specify the video size for the output. It must be even. For the syntax of this option,
23947 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23948 Default value is @code{1920x1080}.
23951 Set the output frame rate. Default value is @code{25}.
23954 Set the bargraph height. It must be even. Default value is @code{-1} which
23955 computes the bargraph height automatically.
23958 Set the axis height. It must be even. Default value is @code{-1} which computes
23959 the axis height automatically.
23962 Set the sonogram height. It must be even. Default value is @code{-1} which
23963 computes the sonogram height automatically.
23966 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
23967 instead. Default value is @code{1}.
23969 @item sono_v, volume
23970 Specify the sonogram volume expression. It can contain variables:
23973 the @var{bar_v} evaluated expression
23974 @item frequency, freq, f
23975 the frequency where it is evaluated
23976 @item timeclamp, tc
23977 the value of @var{timeclamp} option
23981 @item a_weighting(f)
23982 A-weighting of equal loudness
23983 @item b_weighting(f)
23984 B-weighting of equal loudness
23985 @item c_weighting(f)
23986 C-weighting of equal loudness.
23988 Default value is @code{16}.
23990 @item bar_v, volume2
23991 Specify the bargraph volume expression. It can contain variables:
23994 the @var{sono_v} evaluated expression
23995 @item frequency, freq, f
23996 the frequency where it is evaluated
23997 @item timeclamp, tc
23998 the value of @var{timeclamp} option
24002 @item a_weighting(f)
24003 A-weighting of equal loudness
24004 @item b_weighting(f)
24005 B-weighting of equal loudness
24006 @item c_weighting(f)
24007 C-weighting of equal loudness.
24009 Default value is @code{sono_v}.
24011 @item sono_g, gamma
24012 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
24013 higher gamma makes the spectrum having more range. Default value is @code{3}.
24014 Acceptable range is @code{[1, 7]}.
24016 @item bar_g, gamma2
24017 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
24021 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
24022 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
24024 @item timeclamp, tc
24025 Specify the transform timeclamp. At low frequency, there is trade-off between
24026 accuracy in time domain and frequency domain. If timeclamp is lower,
24027 event in time domain is represented more accurately (such as fast bass drum),
24028 otherwise event in frequency domain is represented more accurately
24029 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
24032 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
24033 limits future samples by applying asymmetric windowing in time domain, useful
24034 when low latency is required. Accepted range is @code{[0, 1]}.
24037 Specify the transform base frequency. Default value is @code{20.01523126408007475},
24038 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
24041 Specify the transform end frequency. Default value is @code{20495.59681441799654},
24042 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
24045 This option is deprecated and ignored.
24048 Specify the transform length in time domain. Use this option to control accuracy
24049 trade-off between time domain and frequency domain at every frequency sample.
24050 It can contain variables:
24052 @item frequency, freq, f
24053 the frequency where it is evaluated
24054 @item timeclamp, tc
24055 the value of @var{timeclamp} option.
24057 Default value is @code{384*tc/(384+tc*f)}.
24060 Specify the transform count for every video frame. Default value is @code{6}.
24061 Acceptable range is @code{[1, 30]}.
24064 Specify the transform count for every single pixel. Default value is @code{0},
24065 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
24068 Specify font file for use with freetype to draw the axis. If not specified,
24069 use embedded font. Note that drawing with font file or embedded font is not
24070 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
24074 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
24075 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
24079 Specify font color expression. This is arithmetic expression that should return
24080 integer value 0xRRGGBB. It can contain variables:
24082 @item frequency, freq, f
24083 the frequency where it is evaluated
24084 @item timeclamp, tc
24085 the value of @var{timeclamp} option
24090 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
24091 @item r(x), g(x), b(x)
24092 red, green, and blue value of intensity x.
24094 Default value is @code{st(0, (midi(f)-59.5)/12);
24095 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
24096 r(1-ld(1)) + b(ld(1))}.
24099 Specify image file to draw the axis. This option override @var{fontfile} and
24100 @var{fontcolor} option.
24103 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
24104 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
24105 Default value is @code{1}.
24108 Set colorspace. The accepted values are:
24111 Unspecified (default)
24120 BT.470BG or BT.601-6 625
24123 SMPTE-170M or BT.601-6 525
24129 BT.2020 with non-constant luminance
24134 Set spectrogram color scheme. This is list of floating point values with format
24135 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
24136 The default is @code{1|0.5|0|0|0.5|1}.
24140 @subsection Examples
24144 Playing audio while showing the spectrum:
24146 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
24150 Same as above, but with frame rate 30 fps:
24152 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
24156 Playing at 1280x720:
24158 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
24162 Disable sonogram display:
24168 A1 and its harmonics: A1, A2, (near)E3, A3:
24170 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),
24171 asplit[a][out1]; [a] showcqt [out0]'
24175 Same as above, but with more accuracy in frequency domain:
24177 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),
24178 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
24184 bar_v=10:sono_v=bar_v*a_weighting(f)
24188 Custom gamma, now spectrum is linear to the amplitude.
24194 Custom tlength equation:
24196 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)))'
24200 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
24202 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
24206 Custom font using fontconfig:
24208 font='Courier New,Monospace,mono|bold'
24212 Custom frequency range with custom axis using image file:
24214 axisfile=myaxis.png:basefreq=40:endfreq=10000
24220 Convert input audio to video output representing the audio power spectrum.
24221 Audio amplitude is on Y-axis while frequency is on X-axis.
24223 The filter accepts the following options:
24227 Specify size of video. For the syntax of this option, check the
24228 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24229 Default is @code{1024x512}.
24233 This set how each frequency bin will be represented.
24235 It accepts the following values:
24241 Default is @code{bar}.
24244 Set amplitude scale.
24246 It accepts the following values:
24260 Default is @code{log}.
24263 Set frequency scale.
24265 It accepts the following values:
24274 Reverse logarithmic scale.
24276 Default is @code{lin}.
24279 Set window size. Allowed range is from 16 to 65536.
24281 Default is @code{2048}
24284 Set windowing function.
24286 It accepts the following values:
24309 Default is @code{hanning}.
24312 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
24313 which means optimal overlap for selected window function will be picked.
24316 Set time averaging. Setting this to 0 will display current maximal peaks.
24317 Default is @code{1}, which means time averaging is disabled.
24320 Specify list of colors separated by space or by '|' which will be used to
24321 draw channel frequencies. Unrecognized or missing colors will be replaced
24325 Set channel display mode.
24327 It accepts the following values:
24332 Default is @code{combined}.
24335 Set minimum amplitude used in @code{log} amplitude scaler.
24339 @section showspatial
24341 Convert stereo input audio to a video output, representing the spatial relationship
24342 between two channels.
24344 The filter accepts the following options:
24348 Specify the video size for the output. For the syntax of this option, check the
24349 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24350 Default value is @code{512x512}.
24353 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
24356 Set window function.
24358 It accepts the following values:
24383 Default value is @code{hann}.
24386 Set ratio of overlap window. Default value is @code{0.5}.
24387 When value is @code{1} overlap is set to recommended size for specific
24388 window function currently used.
24391 @anchor{showspectrum}
24392 @section showspectrum
24394 Convert input audio to a video output, representing the audio frequency
24397 The filter accepts the following options:
24401 Specify the video size for the output. For the syntax of this option, check the
24402 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24403 Default value is @code{640x512}.
24406 Specify how the spectrum should slide along the window.
24408 It accepts the following values:
24411 the samples start again on the left when they reach the right
24413 the samples scroll from right to left
24415 frames are only produced when the samples reach the right
24417 the samples scroll from left to right
24420 Default value is @code{replace}.
24423 Specify display mode.
24425 It accepts the following values:
24428 all channels are displayed in the same row
24430 all channels are displayed in separate rows
24433 Default value is @samp{combined}.
24436 Specify display color mode.
24438 It accepts the following values:
24441 each channel is displayed in a separate color
24443 each channel is displayed using the same color scheme
24445 each channel is displayed using the rainbow color scheme
24447 each channel is displayed using the moreland color scheme
24449 each channel is displayed using the nebulae color scheme
24451 each channel is displayed using the fire color scheme
24453 each channel is displayed using the fiery color scheme
24455 each channel is displayed using the fruit color scheme
24457 each channel is displayed using the cool color scheme
24459 each channel is displayed using the magma color scheme
24461 each channel is displayed using the green color scheme
24463 each channel is displayed using the viridis color scheme
24465 each channel is displayed using the plasma color scheme
24467 each channel is displayed using the cividis color scheme
24469 each channel is displayed using the terrain color scheme
24472 Default value is @samp{channel}.
24475 Specify scale used for calculating intensity color values.
24477 It accepts the following values:
24482 square root, default
24493 Default value is @samp{sqrt}.
24496 Specify frequency scale.
24498 It accepts the following values:
24506 Default value is @samp{lin}.
24509 Set saturation modifier for displayed colors. Negative values provide
24510 alternative color scheme. @code{0} is no saturation at all.
24511 Saturation must be in [-10.0, 10.0] range.
24512 Default value is @code{1}.
24515 Set window function.
24517 It accepts the following values:
24542 Default value is @code{hann}.
24545 Set orientation of time vs frequency axis. Can be @code{vertical} or
24546 @code{horizontal}. Default is @code{vertical}.
24549 Set ratio of overlap window. Default value is @code{0}.
24550 When value is @code{1} overlap is set to recommended size for specific
24551 window function currently used.
24554 Set scale gain for calculating intensity color values.
24555 Default value is @code{1}.
24558 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
24561 Set color rotation, must be in [-1.0, 1.0] range.
24562 Default value is @code{0}.
24565 Set start frequency from which to display spectrogram. Default is @code{0}.
24568 Set stop frequency to which to display spectrogram. Default is @code{0}.
24571 Set upper frame rate limit. Default is @code{auto}, unlimited.
24574 Draw time and frequency axes and legends. Default is disabled.
24577 The usage is very similar to the showwaves filter; see the examples in that
24580 @subsection Examples
24584 Large window with logarithmic color scaling:
24586 showspectrum=s=1280x480:scale=log
24590 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
24592 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24593 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
24597 @section showspectrumpic
24599 Convert input audio to a single video frame, representing the audio frequency
24602 The filter accepts the following options:
24606 Specify the video size for the output. For the syntax of this option, check the
24607 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24608 Default value is @code{4096x2048}.
24611 Specify display mode.
24613 It accepts the following values:
24616 all channels are displayed in the same row
24618 all channels are displayed in separate rows
24620 Default value is @samp{combined}.
24623 Specify display color mode.
24625 It accepts the following values:
24628 each channel is displayed in a separate color
24630 each channel is displayed using the same color scheme
24632 each channel is displayed using the rainbow color scheme
24634 each channel is displayed using the moreland color scheme
24636 each channel is displayed using the nebulae color scheme
24638 each channel is displayed using the fire color scheme
24640 each channel is displayed using the fiery color scheme
24642 each channel is displayed using the fruit color scheme
24644 each channel is displayed using the cool color scheme
24646 each channel is displayed using the magma color scheme
24648 each channel is displayed using the green color scheme
24650 each channel is displayed using the viridis color scheme
24652 each channel is displayed using the plasma color scheme
24654 each channel is displayed using the cividis color scheme
24656 each channel is displayed using the terrain color scheme
24658 Default value is @samp{intensity}.
24661 Specify scale used for calculating intensity color values.
24663 It accepts the following values:
24668 square root, default
24678 Default value is @samp{log}.
24681 Specify frequency scale.
24683 It accepts the following values:
24691 Default value is @samp{lin}.
24694 Set saturation modifier for displayed colors. Negative values provide
24695 alternative color scheme. @code{0} is no saturation at all.
24696 Saturation must be in [-10.0, 10.0] range.
24697 Default value is @code{1}.
24700 Set window function.
24702 It accepts the following values:
24726 Default value is @code{hann}.
24729 Set orientation of time vs frequency axis. Can be @code{vertical} or
24730 @code{horizontal}. Default is @code{vertical}.
24733 Set scale gain for calculating intensity color values.
24734 Default value is @code{1}.
24737 Draw time and frequency axes and legends. Default is enabled.
24740 Set color rotation, must be in [-1.0, 1.0] range.
24741 Default value is @code{0}.
24744 Set start frequency from which to display spectrogram. Default is @code{0}.
24747 Set stop frequency to which to display spectrogram. Default is @code{0}.
24750 @subsection Examples
24754 Extract an audio spectrogram of a whole audio track
24755 in a 1024x1024 picture using @command{ffmpeg}:
24757 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
24761 @section showvolume
24763 Convert input audio volume to a video output.
24765 The filter accepts the following options:
24772 Set border width, allowed range is [0, 5]. Default is 1.
24775 Set channel width, allowed range is [80, 8192]. Default is 400.
24778 Set channel height, allowed range is [1, 900]. Default is 20.
24781 Set fade, allowed range is [0, 1]. Default is 0.95.
24784 Set volume color expression.
24786 The expression can use the following variables:
24790 Current max volume of channel in dB.
24796 Current channel number, starting from 0.
24800 If set, displays channel names. Default is enabled.
24803 If set, displays volume values. Default is enabled.
24806 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
24807 default is @code{h}.
24810 Set step size, allowed range is [0, 5]. Default is 0, which means
24814 Set background opacity, allowed range is [0, 1]. Default is 0.
24817 Set metering mode, can be peak: @code{p} or rms: @code{r},
24818 default is @code{p}.
24821 Set display scale, can be linear: @code{lin} or log: @code{log},
24822 default is @code{lin}.
24826 If set to > 0., display a line for the max level
24827 in the previous seconds.
24828 default is disabled: @code{0.}
24831 The color of the max line. Use when @code{dm} option is set to > 0.
24832 default is: @code{orange}
24837 Convert input audio to a video output, representing the samples waves.
24839 The filter accepts the following options:
24843 Specify the video size for the output. For the syntax of this option, check the
24844 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24845 Default value is @code{600x240}.
24850 Available values are:
24853 Draw a point for each sample.
24856 Draw a vertical line for each sample.
24859 Draw a point for each sample and a line between them.
24862 Draw a centered vertical line for each sample.
24865 Default value is @code{point}.
24868 Set the number of samples which are printed on the same column. A
24869 larger value will decrease the frame rate. Must be a positive
24870 integer. This option can be set only if the value for @var{rate}
24871 is not explicitly specified.
24874 Set the (approximate) output frame rate. This is done by setting the
24875 option @var{n}. Default value is "25".
24877 @item split_channels
24878 Set if channels should be drawn separately or overlap. Default value is 0.
24881 Set colors separated by '|' which are going to be used for drawing of each channel.
24884 Set amplitude scale.
24886 Available values are:
24904 Set the draw mode. This is mostly useful to set for high @var{n}.
24906 Available values are:
24909 Scale pixel values for each drawn sample.
24912 Draw every sample directly.
24915 Default value is @code{scale}.
24918 @subsection Examples
24922 Output the input file audio and the corresponding video representation
24925 amovie=a.mp3,asplit[out0],showwaves[out1]
24929 Create a synthetic signal and show it with showwaves, forcing a
24930 frame rate of 30 frames per second:
24932 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
24936 @section showwavespic
24938 Convert input audio to a single video frame, representing the samples waves.
24940 The filter accepts the following options:
24944 Specify the video size for the output. For the syntax of this option, check the
24945 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24946 Default value is @code{600x240}.
24948 @item split_channels
24949 Set if channels should be drawn separately or overlap. Default value is 0.
24952 Set colors separated by '|' which are going to be used for drawing of each channel.
24955 Set amplitude scale.
24957 Available values are:
24977 Available values are:
24980 Scale pixel values for each drawn sample.
24983 Draw every sample directly.
24986 Default value is @code{scale}.
24989 @subsection Examples
24993 Extract a channel split representation of the wave form of a whole audio track
24994 in a 1024x800 picture using @command{ffmpeg}:
24996 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
25000 @section sidedata, asidedata
25002 Delete frame side data, or select frames based on it.
25004 This filter accepts the following options:
25008 Set mode of operation of the filter.
25010 Can be one of the following:
25014 Select every frame with side data of @code{type}.
25017 Delete side data of @code{type}. If @code{type} is not set, delete all side
25023 Set side data type used with all modes. Must be set for @code{select} mode. For
25024 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
25025 in @file{libavutil/frame.h}. For example, to choose
25026 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
25030 @section spectrumsynth
25032 Synthesize audio from 2 input video spectrums, first input stream represents
25033 magnitude across time and second represents phase across time.
25034 The filter will transform from frequency domain as displayed in videos back
25035 to time domain as presented in audio output.
25037 This filter is primarily created for reversing processed @ref{showspectrum}
25038 filter outputs, but can synthesize sound from other spectrograms too.
25039 But in such case results are going to be poor if the phase data is not
25040 available, because in such cases phase data need to be recreated, usually
25041 it's just recreated from random noise.
25042 For best results use gray only output (@code{channel} color mode in
25043 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
25044 @code{lin} scale for phase video. To produce phase, for 2nd video, use
25045 @code{data} option. Inputs videos should generally use @code{fullframe}
25046 slide mode as that saves resources needed for decoding video.
25048 The filter accepts the following options:
25052 Specify sample rate of output audio, the sample rate of audio from which
25053 spectrum was generated may differ.
25056 Set number of channels represented in input video spectrums.
25059 Set scale which was used when generating magnitude input spectrum.
25060 Can be @code{lin} or @code{log}. Default is @code{log}.
25063 Set slide which was used when generating inputs spectrums.
25064 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
25065 Default is @code{fullframe}.
25068 Set window function used for resynthesis.
25071 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25072 which means optimal overlap for selected window function will be picked.
25075 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
25076 Default is @code{vertical}.
25079 @subsection Examples
25083 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
25084 then resynthesize videos back to audio with spectrumsynth:
25086 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
25087 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
25088 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
25092 @section split, asplit
25094 Split input into several identical outputs.
25096 @code{asplit} works with audio input, @code{split} with video.
25098 The filter accepts a single parameter which specifies the number of outputs. If
25099 unspecified, it defaults to 2.
25101 @subsection Examples
25105 Create two separate outputs from the same input:
25107 [in] split [out0][out1]
25111 To create 3 or more outputs, you need to specify the number of
25114 [in] asplit=3 [out0][out1][out2]
25118 Create two separate outputs from the same input, one cropped and
25121 [in] split [splitout1][splitout2];
25122 [splitout1] crop=100:100:0:0 [cropout];
25123 [splitout2] pad=200:200:100:100 [padout];
25127 Create 5 copies of the input audio with @command{ffmpeg}:
25129 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
25135 Receive commands sent through a libzmq client, and forward them to
25136 filters in the filtergraph.
25138 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
25139 must be inserted between two video filters, @code{azmq} between two
25140 audio filters. Both are capable to send messages to any filter type.
25142 To enable these filters you need to install the libzmq library and
25143 headers and configure FFmpeg with @code{--enable-libzmq}.
25145 For more information about libzmq see:
25146 @url{http://www.zeromq.org/}
25148 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
25149 receives messages sent through a network interface defined by the
25150 @option{bind_address} (or the abbreviation "@option{b}") option.
25151 Default value of this option is @file{tcp://localhost:5555}. You may
25152 want to alter this value to your needs, but do not forget to escape any
25153 ':' signs (see @ref{filtergraph escaping}).
25155 The received message must be in the form:
25157 @var{TARGET} @var{COMMAND} [@var{ARG}]
25160 @var{TARGET} specifies the target of the command, usually the name of
25161 the filter class or a specific filter instance name. The default
25162 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
25163 but you can override this by using the @samp{filter_name@@id} syntax
25164 (see @ref{Filtergraph syntax}).
25166 @var{COMMAND} specifies the name of the command for the target filter.
25168 @var{ARG} is optional and specifies the optional argument list for the
25169 given @var{COMMAND}.
25171 Upon reception, the message is processed and the corresponding command
25172 is injected into the filtergraph. Depending on the result, the filter
25173 will send a reply to the client, adopting the format:
25175 @var{ERROR_CODE} @var{ERROR_REASON}
25179 @var{MESSAGE} is optional.
25181 @subsection Examples
25183 Look at @file{tools/zmqsend} for an example of a zmq client which can
25184 be used to send commands processed by these filters.
25186 Consider the following filtergraph generated by @command{ffplay}.
25187 In this example the last overlay filter has an instance name. All other
25188 filters will have default instance names.
25191 ffplay -dumpgraph 1 -f lavfi "
25192 color=s=100x100:c=red [l];
25193 color=s=100x100:c=blue [r];
25194 nullsrc=s=200x100, zmq [bg];
25195 [bg][l] overlay [bg+l];
25196 [bg+l][r] overlay@@my=x=100 "
25199 To change the color of the left side of the video, the following
25200 command can be used:
25202 echo Parsed_color_0 c yellow | tools/zmqsend
25205 To change the right side:
25207 echo Parsed_color_1 c pink | tools/zmqsend
25210 To change the position of the right side:
25212 echo overlay@@my x 150 | tools/zmqsend
25216 @c man end MULTIMEDIA FILTERS
25218 @chapter Multimedia Sources
25219 @c man begin MULTIMEDIA SOURCES
25221 Below is a description of the currently available multimedia sources.
25225 This is the same as @ref{movie} source, except it selects an audio
25231 Read audio and/or video stream(s) from a movie container.
25233 It accepts the following parameters:
25237 The name of the resource to read (not necessarily a file; it can also be a
25238 device or a stream accessed through some protocol).
25240 @item format_name, f
25241 Specifies the format assumed for the movie to read, and can be either
25242 the name of a container or an input device. If not specified, the
25243 format is guessed from @var{movie_name} or by probing.
25245 @item seek_point, sp
25246 Specifies the seek point in seconds. The frames will be output
25247 starting from this seek point. The parameter is evaluated with
25248 @code{av_strtod}, so the numerical value may be suffixed by an IS
25249 postfix. The default value is "0".
25252 Specifies the streams to read. Several streams can be specified,
25253 separated by "+". The source will then have as many outputs, in the
25254 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
25255 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
25256 respectively the default (best suited) video and audio stream. Default
25257 is "dv", or "da" if the filter is called as "amovie".
25259 @item stream_index, si
25260 Specifies the index of the video stream to read. If the value is -1,
25261 the most suitable video stream will be automatically selected. The default
25262 value is "-1". Deprecated. If the filter is called "amovie", it will select
25263 audio instead of video.
25266 Specifies how many times to read the stream in sequence.
25267 If the value is 0, the stream will be looped infinitely.
25268 Default value is "1".
25270 Note that when the movie is looped the source timestamps are not
25271 changed, so it will generate non monotonically increasing timestamps.
25273 @item discontinuity
25274 Specifies the time difference between frames above which the point is
25275 considered a timestamp discontinuity which is removed by adjusting the later
25279 It allows overlaying a second video on top of the main input of
25280 a filtergraph, as shown in this graph:
25282 input -----------> deltapts0 --> overlay --> output
25285 movie --> scale--> deltapts1 -------+
25287 @subsection Examples
25291 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
25292 on top of the input labelled "in":
25294 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
25295 [in] setpts=PTS-STARTPTS [main];
25296 [main][over] overlay=16:16 [out]
25300 Read from a video4linux2 device, and overlay it on top of the input
25303 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
25304 [in] setpts=PTS-STARTPTS [main];
25305 [main][over] overlay=16:16 [out]
25309 Read the first video stream and the audio stream with id 0x81 from
25310 dvd.vob; the video is connected to the pad named "video" and the audio is
25311 connected to the pad named "audio":
25313 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
25317 @subsection Commands
25319 Both movie and amovie support the following commands:
25322 Perform seek using "av_seek_frame".
25323 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
25326 @var{stream_index}: If stream_index is -1, a default
25327 stream is selected, and @var{timestamp} is automatically converted
25328 from AV_TIME_BASE units to the stream specific time_base.
25330 @var{timestamp}: Timestamp in AVStream.time_base units
25331 or, if no stream is specified, in AV_TIME_BASE units.
25333 @var{flags}: Flags which select direction and seeking mode.
25337 Get movie duration in AV_TIME_BASE units.
25341 @c man end MULTIMEDIA SOURCES